Next Article in Journal
Roadmapping the Transition to Water Resource Recovery Facilities: The Two Demonstration Case Studies of Corleone and Marineo (Italy)
Next Article in Special Issue
Effect of Eutrophication Control Methods on the Generation of Greenhouse Carbon Gases in Sediment
Previous Article in Journal
Performance Evaluation of Watershed Environment Governance—A Case Study of Taihu Basin
Previous Article in Special Issue
Permanent Thermal and Chemical Stratification in a Restored Urban Meromictic Lake
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Increases in Picocyanobacteria Abundance in Agriculturally Eutrophic Pampean Lakes Inferred from Historical Records of Secchi Depth and Chlorophyll-a

1
Laboratorio de Ecología Acuática, Instituto Tecnológico de Chascomús (INTECH), Chascomús CC 164 (B7130IWA), Argentina
2
Instituto de Investigaciones Marinas y Costeras (IIMYC), Facultad de Ciencias Exactas y Naturales, Universidad Nacional de Mar del Plata, CONICET, Mar del Plata 7600, Argentina
3
Departamento de Ecología, Genética y Evolución, IEGEBA (UBA-CONICET), Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Buenos Aires 1428, Argentina
4
GESAP (Grupo de Ecología de Sistemas Acuáticos a Escala de Paisaje), INIBIOMA (UNComahue—CONICET), San Carlos de Bariloche 8400, Argentina
*
Author to whom correspondence should be addressed.
These authors contributed equally to this work.
Water 2022, 14(2), 159; https://doi.org/10.3390/w14020159
Submission received: 30 November 2021 / Revised: 28 December 2021 / Accepted: 30 December 2021 / Published: 7 January 2022
(This article belongs to the Special Issue Lake Eutrophication: Causes, Monitoring and Restoration)

Abstract

:
Phytoplankton size structure has profound consequences on food-web organization and energy transfer. Presently, picocyanobacteria (size < 2 µm) represent a major fraction of the autotrophic plankton of Pampean lakes. Glyphosate is known to stimulate the development of picocyanobacteria capable of degrading the herbicide. Due to the worldwide adoption of glyphosate-resistant crops, herbicide usage has increased sharply since the mid-1990s. Unfortunately, there are very few studies (none for the Pampa region) reporting picocyanobacteria abundance before 2000. The proliferation of µm sized particles should decrease Secchi disc depth (ZSD). Therefore ZSD, conditional to chlorophyll-a, may serve as an indicator of picocyanobacteria abundance. We use generalized additive models (GAMs) to analyze a “validation” dataset consisting of 82 records of ZSD, chlorophyll-a, and picocyanobacteria abundance from two Pampean lakes surveys (2009 and 2015). In support of the hypothesis, ZSD was negatively related to picocyanobacteria after accounting for the effect of chlorophyll-a. We then fitted a “historical” dataset using hierarchical GAMs to compare ZSD conditional to chlorophyll-a, before and after 2000. We estimated that ZSD levels during 2000–2021 were, on average, only about half as deep as those during 1980–1999. We conclude that the adoption of glyphosate-resistant crops has stimulated outbreaks of picocyanobacteria populations, resulting in lower water transparency.

1. Introduction

The size structure of phytoplankton has profound consequences on food-web organization and energy transfer within plankton communities [1]. The dominance of small-sized algae (picophytoplankton < 2 μm in diameter) would tend to increase the abundance of smaller grazers (i.e., heterotrophic flagellates and ciliates), promoting the heterotrophic pathway (picoplankton → heterotrophic protists → zooplankton) over the classical trophic chain (phytoplankton → zooplankton), therefore decreasing the energy transfer efficiency toward upper trophic levels [2,3,4]. Changes in land use might affect aquatic community composition and processes, altering food-web interaction and ultimately impacting carbon fluxes in lakes [5].
The fertile Pampa region of Argentina (33–39° S, 57–64° W) harbors large numbers of shallow, eutrophic lakes [6]. Over the last two centuries, the Pampean landscape changed progressively from natural grasslands into cultivated grasslands and croplands due to animal production and intensive crops (wheat, corn, sunflower, and, since the mid-1990s, soybean) [7]. The process of agricultural intensification aggravated the eutrophication of shallow Pampean lakes through increased nutrients (N, P) and agrochemical inputs into the aquatic environment [8,9]. Since the last decade of the 20th century, Pampean lakes have experienced remarkable increases in nutrients (N and P) and chlorophyll-a concentration (Chla) and decreases in water transparency, estimated as Secchi disk depth (ZSD) [10]. Within a historical framework, the eutrophication increase in shallow Pampean lakes was concurrent with the massive conversion to glyphosate-resistant crops (GRC) technologies, adopted by major grain producer countries [11] including Argentina. Glyphosate (N-(phosphonomethyl) glycine) is a broad-spectrum, post-emergent herbicide that is presently among the most widely used agricultural chemicals globally. Usage trends of this herbicide in Argentina have increased steadily, fueled by the widespread adoption of transgenic soybean monoculture and no-tillage practices since around the mid-1990s [12].
In aquatic environments, glyphosate tolerance is widespread among cyanobacteria [13,14], particularly within small-sized species [15]. Studies performed in mesocosms [16,17,18] demonstrated that additions of glyphosate to experimental vessels stimulated the development of picocyanobacteria (Pcy), which reached densities much higher than in control treatments. More recently, a survey of shallow, eutrophic lakes from Argentina [19] provided additional, non-manipulative evidence on the stimulatory effect of glyphosate on Pcy communities in their natural habitats (i.e., Pcy abundance was positively correlated with glyphosate indicators within the Pampa region). Bullerjahn and coworkers [20,21] investigated the mechanism of glyphosate tolerance in picocyanobacteria. They reported that Pcy, isolated from the Laurentian Great Lakes, were able to metabolize glyphosate and its main degradation product, aminomethylphosphonic acid (AMPA), which allowed them to grow in cultures with glyphosate or AMPA as the only sources of phosphorus. Glyphosate tolerance and its utilization by various freshwater Synechococcus strains is mediated by the phosphonate transporter gene phnD, suggesting that detection of phnD gene in plankton samples is an indication of the capacity of picoplankton to metabolize phosphonates, either natural or synthetic (e.g., glyphosate) [20]. We assessed the occurrence of the phnD gene using primers specific for Synechococcus in lake plankton from the Pampa (highly impacted by agricultural practices) and Patagonia (not impacted by agriculture activities) regions of Argentina. Detection of the phnD gene was more frequent and Pcy abundance much higher in lakes from the Pampean region (29% of lakes, 1.8 × 106 cells mL−1) than in similarly eutrophic lakes from Patagonia (5% of lakes, 2.0 × 105 cells mL−1) [19]. Along the same lines, Saxton and coworkers [14] concluded that glyphosate influences the phytoplankton community structure in Lake Erie, serving as a nutrient source to microbes (i.e., heterotrophic bacteria and Pcy) able to tolerate the herbicide effects of the compound, while killing those less tolerant algal species.
Presently, Pcy represent a major fraction of the autotrophic plankton of Pampean lakes. Current estimates often exceed 107 cells mL−1 and rank among the highest abundance reported for lakes worldwide [19,22,23]. Considering such values within the context of the previous arguments, we suspect that the continued usage of glyphosate-based herbicides during the last decades may have impacted natural phytoplankton assemblages of shallow Pampean lakes. More specifically, we hypothesize that by favoring glyphosate-resistant picocyanobacteria, modern agricultural practices have inadvertently contributed to the increase in Pcy abundance up to the very high values typical of present-day Pampean phytoplankton. Unfortunately, epifluorescence microscopy [24] and flow cytometry [25] techniques, which would later allow routine quantification of Pcy concentrations, were not readily implemented until around the 2000s. However, rather conveniently for our purposes, large numbers of small-sized particles, such as Pcy, should produce optical signals that, at least in principle, can be retrieved from historical records of ZSD measurements.
Although regression models of ZSD vs. Chla (often transformed as reciprocal-log or log-log) tend to be highly significant within the mesotrophic–eutrophic range [26,27], Chla usually accounts for only a modest fraction of the total variability in ZSD. The analysis of the residual variability has been used to explore the effect of additional variables, such as inorganic turbidity or water color [28]. ZSD is particularly affected by light scattering due to suspended particles, with small particles having a disproportionately large effect [29]. Stramski and Kiefer [30] estimated that in waters where picophytoplankton concentration approaches 106 cells mL−1, the microorganisms may dominate light scattering, accounting for >50% of its total magnitude. Gallegos et al. [31] reported long-term decreasing trends in ZSD in Chesapeake Bay associated with changes in phytoplankton size structure. In their study, increases in picoplankton abundance resulted in greater light scattering and shallower ZSD. Similarly, analysis of different phytoplankton assemblages differing in size structure showed that the dominance of small-sized algae resulted in remarkably shallower ZSD [32,33]. In other words, ZSD, conditional to Chla (i.e., after statistically removing the effect of Chla), can be expected to bear information on the size structure of the phytoplankton assemblage in general, and particularly, on the abundance of the small-sized plankton fraction.
Summarizing, the accumulated evidence consistently suggests that glyphosate-based herbicides stimulate Pcy populations, resulting in higher abundance of small-sized phytoplankton cells. On the other hand, based on previous theoretical and empirical studies, it is expected that larger numbers of small-sized (~1 µm) particles would decrease ZSD. The question that lingers is whether, under the complex optical climate of Pampean lakes [34], the decrease in ZSD would be sufficiently pronounced, i.e., not masked by confounding factors. Our first objective was to assess this hypothesis. For this purpose, we assembled a “validation” dataset combining two relatively recent (2009 and 2015) surveys of Pampean lakes, reporting data on ZSD, Chla, and Pcy abundance, among other variables. Using statistical tools, we asked whether, at comparable Chla, ZSD was negatively related to Pcy abundance—the practical implication being that ZSD, conditional to Chla, could then be used as an indicator of Pcy abundance. Our second objective was to compare ZSD measurements before and after the massive adoption of glyphosate-based herbicides in Argentina. If confirmed, this would provide a lacking piece of evidence supporting our main hypothesis that the continued use of glyphosate has impacted the phytoplankton size spectrum of shallow Pampean lakes by favoring pico-sized cyanobacteria populations, as previously observed in mesocosms experiments [16,17]. For achieving the second objective, we compiled a “historical” dataset (combining a bibliographic search with our own data).

2. Materials and Methods

The Argentine Pampa region is a large fertile plain in the center-east part of the country. The climate is temperate, with warm summers and mild winters. Precipitation ranges from 1000 mm year−1 in the northeast to 400 mm year−1 in the southwest. Shallow lakes are the most important freshwater ecosystems in the region. Only in Buenos Aires Province, there are roughly 14,000 large (>10 ha) and 146,000 small (0.05–10 ha) shallow lakes ([6], and references therein).
Our first objective was to evaluate whether, at comparable Chla, ZSD was negatively related to Pcy abundance. For testing this hypothesis, we assembled a “validation” dataset combining two relatively recent (2009 and 2015) surveys of Pampean lakes that report data on ZSD, Chla, and Pcy abundance, among other variables [19,23]. The combined dataset totaled 82 observations from 60 Pampean shallow lakes (22 lakes visited twice and 38 lakes visited once), ranging from mesotrophic to hypertrophic (see Table 1). For assessing whether Pcy abundance and Chla (i.e., the predictor variables) significantly contributed to explain the variability in ZSD (i.e., response variable), we used generalized additive models (GAMs) [35]. GAMs allow one to model flexible functions among the response and predictor variables without imposing an a priori specified functional form (e.g., linear, quadratic, etc.). Rather conversely, the shape of the relationship between the response and the predictors arises from the data analysis itself. GAMs were fitted to log10 transformed data to comply with the assumption of normality. We used REML (restricted maximum likelihood) to estimate model coefficients and smoothing parameters and thin plate regression splines (TPRS) as the smoother. The initial basis dimension was checked for adequacy and increased whenever necessary. The model was then refitted according to [36]. We evaluated the uncertainty of the fitted trends under a Bayesian formulation, using 10,000 simulations of the trend from the posterior distribution of the fitted GAM. Model selection was based on Akaike’s Information Criterion (AIC). We estimated GAMs using the mgcv package (version 1.8–36) [35] and uncertainty in the adjusted trends using the gratia package (version 0.6.0) [37]. Graphical outputs were obtained using the ggplot2 package [38] for R [39].
For our second objective (i.e., using ZSD, conditional to Chla, to infer changes in Pcy abundance after the adoption of glyphosate resistant crops in Argentina), we analyzed a “historical” data set consisting of 534 observations of shallow Pampean lakes, covering the period from 1984 to 2021. The database included published records, unpublished data collected by us, and unpublished data provided by colleagues. The full database, including its sources, has been published elsewhere [10]. For the present analyses, only data corresponding to the spring and summer seasons were included. In recent times, a few individual lakes were sampled on many occasions and they were therefore overrepresented. To avoid this potential bias, we randomly subsampled the set of data, leaving 5 observations per lake at most after the year 2000. The final working subset totaled 207 observations and included information from 1984 to 1999 (referred to as 20th century) and from 2000 to 2021 (21st century). We assessed the significance of the relationship between ZSD and Chla and evaluated whether such relationship differed between the two above mentioned periods. For these analyses, we used hierarchical GAM (HGAM) [40], which allows modeling nonlinear functional relationships between the predictor and response variables. Moreover, the model can handle different functional shapes for each grouping level (different centuries, in our case). Model selection involves the analysis of different degrees of intergroup (i.e., between centuries) variability in functional response. The model structure must be specified according to the hypothesis to be tested. Briefly, one must decide (i) whether each group should have its own smoother, or whether a common smoother would suffice; (ii) whether the group-specific smoothers should have the same wiggliness, or whether each group should have its own smoothing; and (iii) whether the smoothers for each group should have a similar shape to one another (i.e., a shared global smoother). Depending on the answers to the above questions, there are five possible model structures: a single common smoother for all observations (model G), a global smoother plus group-level smoothers with the same wiggliness (model GS), a global smoother plus group-level smoothers with different wiggliness (model GI), group-specific smoothers, but all of them having the same wiggliness (model S), and group-specific smoothers with different wiggliness (model I) [40]. Given that our interest was to assess the significance of inter-group (i.e., between centuries) variability, rather than investigating the functional form of the relationship between predictor and response variables, the suitable model structures are either S or I. The performance of these two models was compared based on AIC, which is a robust approach for comparing different model structures [40]. According to this criterion, the model with the lowest AIC should be preferred, and as a rule of thumb, the performance of models differing by less than 2 units are equivalent [41]. We used mgcv package to fit HGMs and estimate the confidence interval of the fitted trends, and ggplot2 to illustrate plots.

3. Results

The validation dataset included lakes in eutrophic–hypereutrophic status, which is the dominant condition in shallow Pampean lakes. In general, they are characterized by high nutrient and Chla concentrations, high Pcy abundance, and alkaline waters (Table 1). Despite the wide range in ZSD, most of them had shallow ZSD (median = 14 cm, Table 1). In order to assess whether Pcy abundance negatively affects ZSD at comparable Chla (objective 1), we fitted two models relating ZSD to either Chla or Pcy abundance separately and a model evaluating the combined effect of Chla and Pcy abundance (Table 2). The first two models significantly explained a percentage of the deviance (Pcy: 24.70%, Chla: 31.30%). However, the best fitting model included both predictor variables (explained deviance: Pcy + Chla: 38.70%). Moreover, the combined model had the lowest AIC, and the difference in AIC to the next best model (i.e., delta AIC) was greater than 4 units. The combined model indicates additive negative effects of Pcy and Chla on ZSD, i.e., increases in either Chla concentration and/or Pcy abundance result in decreased ZSD (Figure 1).
To assess whether ZSD (conditional to Chla) decreased after the adoption of glyphosate-resistant crops (objective 2), we compared measurements from 1984 to 1999 (20th century) vs. those from 2000 to 2021 (21st century). For this comparison, we used hierarchical GAM (HGAM) to assess whether the functional relationship between ZSD and Chla concentration varied between groups (here referred to as centuries for simplicity). Two HGAM, with slightly different model structures, were evaluated: S and I. Both models allowed for group-specific smooth terms, but model S imposed the same wiggliness to the smooth, while model I allowed different groups to differ in wiggliness. In both cases, model fits were significant, indicating not only a decreasing trend of ZSD with Chla but also significant intergroup differences (i.e., a significant “century” effect: ZSD20 > ZSD21). Both models explained a similar percentage of the deviance (46.2–46.4, Table 3), with model I having the lowest AIC. Model I, in addition to demonstrating significant trends for both centuries, showed a significant random effect for centuries to model century-specific intercepts. According to the rule of thumb, there was no ground to prefer one model over the other (i.e., they differed in less than AIC units) (Table 3). We prefer model I since it imposes fewer restrictions to the smoother function (i.e., each century has its own individual smoother) and shows in the best overall fit (higher explained deviance and higher estimated degrees of freedom for smooth terms) (Figure 2). To illustrate intergroup differences in ZSD over the last 40 years, we estimated ZSD for each century using the selected model I for three Chla levels (i.e., low, moderate, and high) (Table 4). This exercise showed that for the same Chla, ZSD corresponding to the 21st century was only about 50% of the corresponding levels of ZSD in the 20th century, regardless of the Chla level considered.

4. Discussion

Due to their small size, Pcy cells are expected to absorb and scatter light more strongly on a mass-specific basis than larger cells [29,31]. Therefore, after accounting for the effect of Chla, the depth of disappearance of the Secchi disc (ZSD) should decrease, in theory, with the concentration of Pcy. This effect, in turn, should be more apparent at high Pcy concentrations (>106 cells mL−1) [30]. The analysis of the “validation” dataset confirmed this prediction, proving that the theoretical expectation holds true, even for highly complex waters, such as those typical of shallow Pampean lakes [42]. Our findings reinforce the importance of water clarity as an ecosystem indicator of eutrophication. Although there are several measures of water transparency, ZSD has been the most consistent and frequent measure employed over time [43].
Previous experiments [16,17] and field surveys of shallow Pampean lakes [19] demonstrated that exposure to glyphosate stimulates the development of Pcy populations. On the other hand, we have reported that the ZSD levels, conditional to Chla, in glyphosate-impacted Pampean lakes are significantly shallower (~50–60%) than the values reported for shallow lakes worldwide, within the low temperate latitudinal range (defined as two belts from 23.6° to 44.5° North or South) [6]. Considering the steady increase in glyphosate use in Argentina and the very high Pcy concentrations reported for Pampean lakes in recent years [19,23,44], we suspect the shallower ZSD reported for Pampean lakes in recent times represents an unforeseen result of the adoption of glyphosate-resistant crops. The analysis of the “historical” dataset revealed that ZSD values, conditional to Chla, during the 21st century (i.e., after the adoption of GR-crops) are significantly lower than the values reported for the 20th century. In fact, current ZSD values are only about 50% of the values reported during the 1980s and 1990s, which were therefore similar to those reported for other lakes worldwide. All in all, the collected evidence suggests a scenario in which the massive use of glyphosate in the Pampean region has shifted the phytoplankton size structure (increased Pcy abundance), which in turn translated into lower water transparency (decreased ZSD).
As already mentioned, Pcy abundance in shallow Pampean lakes was not quantified by epifluorescence microscopy or cytometry before 2005. However, the increase in the abundance of tiny cells resembling Pcy over the last decades was apparent while quantifying nano- and micro-phytoplankton under the inverted microscope (I. Izaguirre, personal communication). On the other hand, Pcy cells often aggregate to form microcolonies in response to different stimuli [45,46], and there is typically a positive relationship between the abundance of microcolonies and that of single Pcy cells. Our historical records of microcolonies abundance for a few Pampean lakes from 1987 [47] to 2015 [48,49] show remarkable abundance increases. For example, in Lake Chascomús, microcolonies abundance increased from 710 to 16,379 ind. mL−1 in 1987–1989 to 7396 to 182,297 ind. mL−1 in 2005–2015. Moreover, since Pcy began to be quantified by epifluorescence and flow cytometry [19,23,44,48,50], Pcy abundances have increased by an order of magnitude in several lakes over the period 2005–2015. Presently, dominant Pcy in Pampean lakes belong to many different OTUs affiliated with the genera Synechococcus and Cyanobium, within the order Synechococcales [49,51].
Optical signals similar to those observed in the Pampean region have been reported in other agriculture-impacted areas. The use of glyphosate is widespread in the Chesapeake Bay watershed [52]. Long-term increases in light scattering in Chesapeake Bay have been inferred from ZSD, light attenuation, and remote sensing measurements [53]. Gallegos and co-workers [31] suggested that these long-term trends could be explained by decreases in the size structure of plankton populations. Independently, Wang et al. [54] reported high densities of picoplanktonic Synechococcus (106 cells mL−1) throughout the Bay, contributing 20% to 40% of total phytoplankton chlorophyll-a. In Lake Erie, Burns and coworkers [55] reported decreasing trends in ZSD during the period 1983–2002. According to these authors, the observed reduction in ZSD by 7 ± 3 cm year−1 was unrelated to variation in Chla. Moreover, Barbiero and Tuchman [56] observed decreases in Chla of 50% in the western basin of Lake Erie, but despite these decreases in phytoplankton abundance, they actually observed an increase in turbidity in the western basin [57]. Suggestively, the cyanobacteria community of Lake Erie is persistently dominated by small, single-celled species consistent with Synechococcus spp. And Cyanobium spp. [58].
Changes in plankton composition might alter the trophic interactions, and consequently the flow of matter and energy through the food web. For instance, certain species of zooplankton mediate the phenotypic plasticity of Pcy, promoting their aggregation into colonies [44,45,46]. The decrease in the average cell size of phytoplankton would enhance the relative importance of intraguild predators at intermediate trophic levels of the food web, favoring the heterotrophic microbial pathway through small protists with the increase in complexity and consequent reduction in trophic efficiency, as it occurs in large tropical lakes [4]. The increase in the relative importance of Pcy affects the underwater light climate, promoting more turbid conditions that might have a negative impact on primary production in light-limited turbid shallow lakes [59,60], reducing carbon fixation. Both processes—the increment of microbial heterotrophic activity and the reduction of photosynthesis—tend to occur in more heterotrophic environments.
Cultural eutrophication is a long-recognized phenomenon. Although at local scales it might have been similar or even more significant in the past than at present, it was only recently that it attained a true global dimension [61]. In its wider definition, anthropogenic eutrophication refers to the overproduction of organic material induced by anthropogenic inputs of phosphorus and nitrogen, and the concept of syndrome (i.e., a set of symptoms) has been introduced to refer to the myriad of biogeochemical and biological responses to increased nitrogen and phosphorus inputs [62]. Modern agricultural packages, however, involve the utilization of a large and diverse mixture of natural and synthetic chemicals whose effects go beyond those of traditional fertilizers. For example, the usage of glyphosate in the Pampa region not only resulted in increased abundance of small-sized Pcy, as argued in the present study, but also resulted in higher abundance of the smallest species within each major algal class [49], in clear contrast with the conventional pattern of increasing phytoplankton cell size with trophic state ([63], p. 275) [64,65]. Paleolimnological evidence also relates eutrophication with the shift in the phytoplankton community toward cyanobacteria dominance, along with a shift in zooplankton community towards small-bodied zooplankton, such as Bosmina and rotifers [66]. These findings suggest that agriculture, through non-point contamination of nutrients and glyphosate, impacts the composition of the phytoplankton and zooplankton communities and their size structure, with multiple ecological consequences. We therefore concur with the editors of this Special Issue (https://www.mdpi.com/si/water/lake_eutrophications#info, accessed on 16 June 2021) in that new types of water pollution continuously impose new challenges, requiring recurring reexamination of the eutrophication phenomenon. Our results highlight the relevance of adopting sustainable practices and restoring buffering areas in shallow lakes (e.g., riparian wetland zone at lake margins) to prevent contaminants from reaching the water bodies.

5. Conclusions

In this study, we first demonstrated that ZSD (conditional Chla) decreases with the abundance of Pcy. Next, we demonstrated a significant decrease in ZSD (conditional Chla) in Pampean lakes. Such a decrease, which roughly occurred at the turn of the century, was therefore contemporaneous with the sharp increase in glyphosate usage due to the adoption of GR-crops. Previous studies have experimentally demonstrated that the exposure to glyphosate stimulates the development of large Pcy populations, while recent lakes surveys revealed that the abundance of Pcy in Pampean lakes is, on average, remarkably higher than the world average. We conclude that this cumulus of evidence strongly suggests that the increase in glyphosate usage in recent decades has resulted in increased Pcy abundance in Pampean lakes and suggest that the same phenomenon may have also occurred in other agricultural areas of the World.

Author Contributions

Conceptualization, H.E.Z., M.d.l.Á.G.S. and N.R.D.; methodology, G.L.P., M.d.l.Á.G.S., M.C.B., P.F. and H.E.Z.; validation, M.C.B. and G.L.P.; formal analysis, M.d.l.Á.G.S.; investigation, I.I., N.R.D. and F.U.; data curation, M.C.B. and P.F.; writing—original draft preparation, H.E.Z., M.d.l.Á.G.S. and I.I.; writing—review and editing, H.E.Z., N.R.D., M.d.l.Á.G.S., I.I., P.F., F.U., M.C.B. and G.L.P.; visualization, H.E.Z. and M.d.l.Á.G.S.; supervision, H.E.Z.; funding acquisition, H.E.Z. and I.I. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by CONICET (red PAMPA2), ANCyPT (PICT RAICES 2017–2498); University of Buenos Aires (UBACyT2002013010045BA), ANCyPT (PICT-2016-1079); National Geographic Society (grant # 9736-15).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Acknowledgments

We thank all the colleagues who provided us their data sets of the Pampean shallow lakes: Sofía Baliña, Manuel Castro Berman, Darío Colautti, Victor Conzonno, Néstor Gabellone, Fabián Grosman, Gonzalo Pérez, M. Laura Sánchez, Pablo Sanzano, Ana Torremorell, and Alicia Vinocur.

Conflicts of Interest

The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

References

  1. Marañón, E. Cell Size as a Key Determinant of Phytoplankton Metabolism and Community Structure. Annu. Rev. Mar. Sci. 2015, 7, 241–264. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  2. Azam, F.; Fenchel, T.; Field, J.G.; Gray, J.S.; Meyer-Reil, L.A.; Thingstad, F. The Ecological Role of Water-Column Microbes in the Sea. Mar. Ecol. Prog. Ser. 1983, 10, 257. [Google Scholar] [CrossRef]
  3. Legendre, L.; LeFevre, J. Microbial Food Webs and the Export of Biogenic Carbon in Oceans. Aquat. Microb. Ecol. 1995, 09, 69–77. [Google Scholar] [CrossRef] [Green Version]
  4. Sarmento, H. New Paradigms in Tropical Limnology: The Importance of the Microbial Food Web. Hydrobiologia 2012, 686, 1–14. [Google Scholar] [CrossRef]
  5. Smith, V.H.; Joye, S.B.; Howarth, R.W. Eutrophication of Freshwater and Marine Ecosystems. Limnol. Oceanogr. 2006, 51, 351–355. [Google Scholar] [CrossRef] [Green Version]
  6. Diovisalvi, N.; Bohn, V.Y.; Piccolo, M.C.; Perillo, G.M.E.; Baigún, C.; Zagarese, H.E. Shallow Lakes from the Central Plains of Argentina: An Overview and Worldwide Comparative Analysis of Their Basic Limnological Features. Hydrobiologia 2015, 752, 5–20. [Google Scholar] [CrossRef]
  7. Castro Berman, M.; Marino, D.J.G.; Quiroga, M.V.; Zagarese, H. Occurrence and Levels of Glyphosate and AMPA in Shallow Lakes from the Pampean and Patagonian Regions of Argentina. Chemosphere 2018, 200, 513–522. [Google Scholar] [CrossRef]
  8. Rennella, A.M.; Quirós, R. The Effects of Hydrology on Plankton Biomass in Shallow Lakes of the Pampa Plain. Hydrobiologia 2006, 556, 181–191. [Google Scholar] [CrossRef]
  9. Zagarese, H.E.; González Sagrario, M.Á.; Wolf-Gladrow, D.; Nõges, P.; Nõges, T.; Kangur, K.; Matsuzaki, S.-I.S.; Kohzu, A.; Vanni, M.J.; Özkundakci, D.; et al. Patterns of CO2 Concentration and Inorganic Carbon Limitation of Phytoplankton Biomass in Agriculturally Eutrophic Lakes. Water Res. 2021, 190, 116715. [Google Scholar] [CrossRef]
  10. Izaguirre, I.; Zagarese, H.E.; O’Farrell, I. The Limnological Trace of Contemporaneous Anthropogenic Activities in the Pampa Region. Ecol. Austral. 2022, submitted.
  11. Duke, S.O. The History and Current Status of Glyphosate. Pest. Manag. Sci. 2018, 74, 1027–1034. [Google Scholar] [CrossRef]
  12. Binimelis, R.; Pengue, W.; Monterroso, I. “Transgenic Treadmill”: Responses to the Emergence and Spread of Glyphosate-Resistant Johnsongrass in Argentina. Geoforum 2009, 40, 623–633. [Google Scholar] [CrossRef]
  13. Forlani, G.; Pavan, M.; Gramek, M.; Kafarski, P.; Lipok, J. Biochemical Bases for a Widespread Tolerance of Cyanobacteria to the Phosphonate Herbicide Glyphosate. Plant Cell Physiol. 2008, 49, 443–456. [Google Scholar] [CrossRef] [PubMed]
  14. Saxton, M.A.; Morrow, E.A.; Bourbonniere, R.A.; Wilhelm, S.W. Glyphosate Influence on Phytoplankton Community Structure in Lake Erie. J. Great Lakes Res. 2011, 37, 683–690. [Google Scholar] [CrossRef]
  15. Lipok, J.; Studnik, H.; Gruyaert, S. The Toxicity of Roundup® 360 SL Formulation and Its Main Constituents: Glyphosate and Isopropylamine towards Non-Target Water Photoautotrophs. Ecotox. Environ. Safe 2010, 73, 1681–1688. [Google Scholar] [CrossRef]
  16. Pérez, G.L.; Torremorell, A.; Mugni, H.; Rodríguez, P.; Vera, M.S.; do Nascimento, M.; Allende, L.; Bustingorry, J.; Escaray, R.; Ferraro, M.; et al. Effects of the Herbicide Roundup on Freshwater Microbial Communities: A Mesocosm Study. Ecol. Appl. 2007, 17, 2310–2322. [Google Scholar] [CrossRef] [Green Version]
  17. Vera, M.S.; Lagomarsino, L.; Sylvester, M.; Pérez, G.L.; Rodríguez, P.; Mugni, H.; Sinistro, R.; Ferraro, M.; Bonetto, C.; Zagarese, H.; et al. New Evidences of Roundup (Glyphosate Formulation) Impact on the Periphyton Community and the Water Quality of Freshwater Ecosystems. Ecotoxicol. Lond. Engl 2010, 19, 710–721. [Google Scholar] [CrossRef]
  18. Vera, M.S.; Fiori, E.D.; Lagomarsino, L.; Sinistro, R.; Escaray, R.; Iummato, M.M.; Juárez, A.; Ríos de Molina, M.d.C.; Tell, G.; Pizarro, H. Direct and Indirect Effects of the Glyphosate Formulation Glifosato Atanor® on Freshwater Microbial Communities. Ecotoxicol. Lond. Engl. 2012, 21, 1805–1816. [Google Scholar] [CrossRef]
  19. Castro Berman, M.; Llames, M.E.; Minotti, P.; Fermani, P.; Quiroga, M.V.; Ferraro, M.A.; Metz, S.; Zagarese, H.E. Field Evidence Supports Former Experimental Claims on the Stimulatory Effect of Glyphosate on Picocyanobacteria Communities. Sci. Total Environ. 2020, 701, 134601. [Google Scholar] [CrossRef] [PubMed]
  20. Ilikchyan, I.N.; McKay, R.M.L.; Zehr, J.P.; Dyhrman, S.T.; Bullerjahn, G.S. Detection and Expression of the Phosphonate Transporter Gene PhnD in Marine and Freshwater Picocyanobacteria. Environ. Microbiol. 2009, 11, 1314–1324. [Google Scholar] [CrossRef]
  21. Kutovaya, O.A.; McKay, R.M.L.; Bullerjahn, G.S. Detection and Expression of Genes for Phosphorus Metabolism in Picocyanobacteria from the Laurentian Great Lakes. J. Great Lakes Res. 2013, 39, 612–621. [Google Scholar] [CrossRef]
  22. Fermani, P.; Diovisalvi, N.; Torremorell, A.; Lagomarsino, L.; Zagarese, H.E.; Unrein, F. The Microbial Food Web Structure of a Hypertrophic Warm-Temperate Shallow Lake, as Affected by Contrasting Zooplankton Assemblages. Hydrobiologia 2013, 714, 115–130. [Google Scholar] [CrossRef]
  23. Fermani, P.; Torremorell, A.; Lagomarsino, L.; Escaray, R.; Unrein, F.; Pérez, G. Microbial Abundance Patterns along a Transparency Gradient Suggest a Weak Coupling between Heterotrophic Bacteria and Flagellates in Eutrophic Shallow Pampean Lakes. Hydrobiologia 2015, 752, 103–123. [Google Scholar] [CrossRef]
  24. Daley, R.J.; Hobbie, J.E. Direct Counts of Aquatic Bacteria by a Modified Epifluorescence Technique. Limnol. Oceanogr. 1975, 20, 875–882. [Google Scholar] [CrossRef]
  25. Olson, R.J.; Vaulot, D.; Chisholm, S.W. Marine Phytoplankton Distributions Measured Using Shipboard Flow Cytometry. Deep Sea Res. Part Oceanogr. Res. Pap. 1985, 32, 1273–1280. [Google Scholar] [CrossRef]
  26. Bachmann, R.W.; Hoyer, M.V.; Croteau, A.C.; Canfield, D.E. Factors Related to Secchi Depths and Their Stability over Time as Determined from a Probability Sample of US Lakes. Environ. Monit. Assess. 2017, 189, 206. [Google Scholar] [CrossRef] [PubMed]
  27. Carlson, R.E. A Trophic State Index for Lakes. Limnol. Oceanogr. 1977, 22, 361–369. [Google Scholar] [CrossRef] [Green Version]
  28. Brezonik, P.L.; Bouchard, R.W.; Finlay, J.C.; Griffin, C.G.; Olmanson, L.G.; Anderson, J.P.; Arnold, W.A.; Hozalski, R. Color, Chlorophyll a, and Suspended Solids Effects on Secchi Depth in Lakes: Implications for Trophic State Assessment. Ecol. Appl. 2019, 29, e01871. [Google Scholar] [CrossRef]
  29. Kirk, J. Light and Photosynthesis in Aquatic Ecosystems; Cambridge University Press: Cambridge, UK, 1983. [Google Scholar] [CrossRef]
  30. Stramski, D.; Kiefer, D.A. Light Scattering by Microorganisms in the Open Ocean. Prog. Oceanogr. 1991, 28, 343–383. [Google Scholar] [CrossRef]
  31. Gallegos, C.L.; Werdell, P.J.; McClain, C.R. Long-term Changes in Light Scattering in Chesapeake Bay Inferred from Secchi Depth, Light Attenuation, and Remote Sensing Measurements. J. Geophys. Res. Ocean. 2011, 116, C00H08. [Google Scholar] [CrossRef]
  32. Mazumder, A.; Taylor, W.D.; McQueen, D.J.; Lean, D.R.S. Effects of Fish and Plankton on Lake Temperature and Mixing Depth. Science 1990, 247, 312–315. [Google Scholar] [CrossRef]
  33. Havens, K.E.; Heath, R.T. Increased Transparency Due to Changes in the Algal Size Spectrum during Experimental Acidification in Mesocosms. J. Plankton Res. 1991, 13, 673–679. [Google Scholar] [CrossRef]
  34. Pérez, G.L.; Lagomarsino, L.; Zagarese, H.E. Optical Properties of Highly Turbid Shallow Lakes with Contrasting Turbidity Origins: The Ecological and Water Management Implications. J. Environ. Manag. 2013, 130, 207–220. [Google Scholar] [CrossRef]
  35. Wood, N. Generalized Additive Models: An Introduction with R; Chapman and Hall/CRC: Boca Raton, FL, USA, 2017. [Google Scholar]
  36. Wood, S.N.; Pya, N.; Säfken, B. Smoothing Parameter and Model Selection for General Smooth Models. J. Am. Stat. Assoc. 2016, 111, 1548–1563. [Google Scholar] [CrossRef]
  37. Simpson, G. Gratia: Graceful Ggplot-Based Graphics and Other Useful Functions for Gams Fitted Using Mgcv. 2018. Available online: https://cran.r-project.org/web/packages/gratia/index.html (accessed on 29 December 2021).
  38. Wickham, H. Ggplot2: Elegant Graphics for Data Analysis; Springer: New York, NY, USA, 2016; ISBN 978-3-319-24277-4. [Google Scholar]
  39. R Core Team. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2018. [Google Scholar]
  40. Pedersen, E.J.; Miller, D.L.; Simpson, G.L.; Ross, N. Hierarchical Generalized Additive Models in Ecology: An Introduction with Mgcv. PeerJ 2019, 7, e6876. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  41. Burnham, K.P.; Anderson, D.R. Model Selection and Inference, A Practical Information-Theoretic Approach; Springer: New York, NY, USA, 1998; ISBN 9781475729191. [Google Scholar]
  42. Pérez, G.L.; Torremorell, A.; Bustingorry, J.; Escaray, R.; Pérez, P.; Diéguez, M.; Zagarese, H. Optical Characteristics of Shallow Lakes from the Pampa and Patagonia Regions of Argentina. Limnol. Ecol Manag. Inland Waters 2010, 40, 30–39. [Google Scholar] [CrossRef] [Green Version]
  43. Testa, J.M.; Lyubchich, V.; Zhang, Q. Patterns and Trends in Secchi Disk Depth over Three Decades in the Chesapeake Bay Estuarine Complex. Estuaries Coasts 2019, 42, 927–943. [Google Scholar] [CrossRef]
  44. Quiroga, M.V.; Huber, P.; Ospina-Serna, J.; Diovisalvi, N.; Odriozola, M.; Cueto, G.R.; Lagomarsino, L.; Fermani, P.; Bustingorry, J.; Escaray, R.; et al. The Dynamics of Picocyanobacteria from a Hypereutrophic Shallow Lake Is Affected by Light-Climate and Small-Bodied Zooplankton: A Ten-Year Cytometric Time-Series Analysis. FEMS Microbiol. Ecol. 2021, 97, fiab055. [Google Scholar] [CrossRef]
  45. Ospina-Serna, J.; Huber, P.; Odriozola, M.; Fermani, P.; Unrein, F. Picocyanobacteria Aggregation as a Response of Predation Pressure: The Direct Contact Is Not Necessary. FEMS Microbiol. Ecol. 2020, 96, fiaa153. [Google Scholar] [CrossRef]
  46. Huber, P.; Diovisalvi, N.; Ferraro, M.; Metz, S.; Lagomarsino, L.; Llames, M.E.; Royo-Llonch, M.; Bustingorry, J.; Escaray, R.; Acinas, S.G.; et al. Phenotypic Plasticity in Freshwater Picocyanobacteria. Environ. Microbiol. 2017, 19, 1120–1133. [Google Scholar] [CrossRef]
  47. Izaguirre, I.; Vinocur, A. Typology of Shallow Lakes of the Salado River Basin (Argentina), Based on Phytoplankton Communities. Hydrobiologia 1994, 277, 49–62. [Google Scholar] [CrossRef]
  48. Allende, L.; Tell, G.; Zagarese, H.; Torremorell, A.; Pérez, G.; Bustingorry, J.; Escaray, R.; Izaguirre, I. Phytoplankton and Primary Production in Clear-Vegetated, Inorganic-Turbid, and Algal-Turbid Shallow Lakes from the Pampa Plain (Argentina). Hydrobiologia 2009, 624, 45–60. [Google Scholar] [CrossRef]
  49. Castro Berman, M.; Farrell, I.O.; Huber, P.; Marino, D.; Zagarese, H. A Large-Scale Geographical Coverage Survey Reveals a Pervasive Impact of Agricultural Practices on Plankton Primary Producers. Agric. Ecosyst. Environ. 2022, 325, 107740. [Google Scholar] [CrossRef]
  50. Silvoso, J.; Izaguirre, I.; Allende, L. Picoplankton Structure in Clear and Turbid Eutrophic Shallow Lakes: A Seasonal Study. Limnol. Ecol. Manag. Inland Waters 2011, 41, 181–190. [Google Scholar] [CrossRef] [Green Version]
  51. Huber, P.; Cornejo-Castillo, F.M.; Ferrera, I.; Sánchez, P.; Logares, R.; Metz, S.; Balagué, V.; Acinas, S.G.; Gasol, J.M.; Unrein, F. Primer Design for an Accurate View of Picocyanobacterial Community Structure by Using High-Throughput Sequencing. Appl. Environ. Microb. 2019, 85, e02659-18. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  52. Hartwell, S.I. Chesapeake Bay Watershed Pesticide Use Declines but Toxicity Increases. Environ. Toxicol. Chem. 2011, 30, 1223–1231. [Google Scholar] [CrossRef]
  53. Harding, L.W.; Mallonee, M.E.; Perry, E.S.; Miller, W.D.; Adolf, J.E.; Gallegos, C.L.; Paerl, H.W. Long-Term Trends, Current Status, and Transitions of Water Quality in Chesapeake Bay. Sci. Rep. 2019, 9, 6709. [Google Scholar] [CrossRef] [PubMed]
  54. Wang, K.; Wommack, K.E.; Chen, F. Abundance and Distribution of Synechococcus Spp. and Cyanophages in the Chesapeake Bay. Appl. Environ. Microb. 2011, 77, 7459–7468. [Google Scholar] [CrossRef] [Green Version]
  55. Burns, N.M.; Rockwell, D.C.; Bertram, P.E.; Dolan, D.M.; Ciborowski, J.J.H. Trends in Temperature, Secchi Depth, and Dissolved Oxygen Depletion Rates in the Central Basin of Lake Erie, 1983–2002. J. Great Lakes Res. 2005, 31, 35–49. [Google Scholar] [CrossRef]
  56. Barbiero, R.P.; Tuchman, M.L. Long-Term Dreissenid Impacts on Water Clarity in Lake Erie. J. Great Lakes Res. 2004, 30, 557–565. [Google Scholar] [CrossRef]
  57. Binding, C.E.; Jerome, J.H.; Bukata, R.P.; Booty, W.G. Trends in Water Clarity of the Lower Great Lakes from Remotely Sensed Aquatic Color. J. Great Lakes Res. 2007, 33, 828–841. [Google Scholar] [CrossRef] [Green Version]
  58. Steffen, M.M.; Belisle, B.S.; Watson, S.B.; Boyer, G.L.; Wilhelm, S.W. Status, Causes and Controls of Cyanobacterial Blooms in Lake Erie. J. Great Lakes Res. 2014, 40, 215–225. [Google Scholar] [CrossRef]
  59. Torremorell, A.; Llames, M.E.; Pérez, G.L.; Escaray, R.; Bustingorry, J.; Zagarese, H. Annual Patterns of Phytoplankton Density and Primary Production in a Large, Shallow Lake: The Central Role of Light. Freshw. Biol. 2009, 54, 437–449. [Google Scholar] [CrossRef]
  60. Llames, M.E.; Lagomarsino, L.; Diovisalvi, N.; Fermani, P.; Torremorell, A.M.; Perez, G.; Unrein, F.; Bustingorry, J.; Escaray, R.; Ferraro, M.; et al. The Effects of Light Availability in Shallow, Turbid Waters: A Mesocosm Study. J. Plankton Res. 2009, 31, 1517–1529. [Google Scholar] [CrossRef] [Green Version]
  61. Wassmann, P. Cultural Eutrophication: Perspectives and Prospects. In Drainage Basin Nutrient Inputs and Eutrophication: An Integrated Approach; Wassmann, P., Olli, K., Eds.; University of Tromsø: Tromsø, Norway, 2004. [Google Scholar]
  62. Moal, M.L.; Gascuel-Odoux, C.; Ménesguen, A.; Souchon, Y.; Étrillard, C.; Levain, A.; Moatar, F.; Pannard, A.; Souchu, P.; Lefebvre, A.; et al. Eutrophication: A New Wine in an Old Bottle? Sci. Total Environ. 2019, 651, 1–11. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  63. Moss, B. Ecology of Freshwaters: A View for the Twenty-First Century, 4th ed.; Blackwell Scientific Publication: Hoboken, NJ, USA, 2010; ISBN 978-1-4051-1332-8. [Google Scholar]
  64. Schulhof, M.A.; Shurin, J.B.; Declerck, S.A.J.; de Waal, D.B.V. Phytoplankton Growth and Stoichiometric Responses to Warming, Nutrient Addition and Grazing Depend on Lake Productivity and Cell Size. Glob. Change Biol. 2019, 25, 2751–2762. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  65. Del Giorgio, P.A.; Peters, R.H. Balance between Phytoplankton Production and Plankton Respiration in Lakes. Can. J. Fish Aquat. Sci. 1993, 50, 282–289. [Google Scholar] [CrossRef]
  66. González Sagrario, M.A.; Musazzi, S.; Córdoba, F.; Mendiolar, M.; Lami, A. Inferring the Occurrence of Regime Shifts in a Shallow Lake during the Last 250 Years Based on Multiple Indicators. Ecol. Indic. 2020, 117, 106536. [Google Scholar] [CrossRef]
Figure 1. GAM results for spring–summer ZSD in shallow Pampean lakes, showing the fitted trend, confidence intervals, and observed values. Plots are partial plots of the smooth term in the model, and the y axis is the intercept plus the partial effect of the individual smooth. Data are log10 transformed, Chla was originally expressed in µg L−1 (a); picocyanobacterial abundance in cells mL−1; and ZSD in cm (b).
Figure 1. GAM results for spring–summer ZSD in shallow Pampean lakes, showing the fitted trend, confidence intervals, and observed values. Plots are partial plots of the smooth term in the model, and the y axis is the intercept plus the partial effect of the individual smooth. Data are log10 transformed, Chla was originally expressed in µg L−1 (a); picocyanobacterial abundance in cells mL−1; and ZSD in cm (b).
Water 14 00159 g001
Figure 2. HGAM results for model I, showing fitted trends, confidence intervals, and observed values for data corresponding to the period 1984–1999 (red) and 2000–2021 (blue). Plots are partial plots of the smooth term in the model, and the y axis is the intercept plus the effect of the individual smooth for each group (century). Data are log10 transformed; chlorophyll-a (Chla) was originally expressed in µg l−L and Secchi disc depth (ZSD) in cm.
Figure 2. HGAM results for model I, showing fitted trends, confidence intervals, and observed values for data corresponding to the period 1984–1999 (red) and 2000–2021 (blue). Plots are partial plots of the smooth term in the model, and the y axis is the intercept plus the effect of the individual smooth for each group (century). Data are log10 transformed; chlorophyll-a (Chla) was originally expressed in µg l−L and Secchi disc depth (ZSD) in cm.
Water 14 00159 g002
Table 1. Median and range of limnological variables corresponding to the “validation” dataset (i.e., 82 shallow Pampean lakes sampled during spring and summer surveys performed in 2009 and 2015).
Table 1. Median and range of limnological variables corresponding to the “validation” dataset (i.e., 82 shallow Pampean lakes sampled during spring and summer surveys performed in 2009 and 2015).
Limnological ParameterMedianRange
ZSD (cm)142–191
Chla (µg L−1)64.281.58–1549.79
Pcy (cells mL−1)9.60 × 1053.3 × 103–7.72 × 107
Total phosphorus (µg L−1)620124.2–4538
Total nitrogen (µg L−1)50511702–12,846
pH8.98–10.2
ZSD: Secchi disk depth; Chla: chlorophyll-a concentration; Pcy: picocyanobacteria abundance.
Table 2. Statistics from fitted GAMs to the “validation” dataset; grey shadow represents the best fitted model based on AIC and the percentage of deviance explained; edf, estimated degrees of freedom; p-value < 0.05 is statistically significant; AIC, Akaike Information Criterion; deltaAIC is the difference with respect to the lowest AIC value.
Table 2. Statistics from fitted GAMs to the “validation” dataset; grey shadow represents the best fitted model based on AIC and the percentage of deviance explained; edf, estimated degrees of freedom; p-value < 0.05 is statistically significant; AIC, Akaike Information Criterion; deltaAIC is the difference with respect to the lowest AIC value.
Modeledfp-ValueDevianceAICDeltaAIC
Secchi ~ Pcy1.42.75 × 10−524.70%47.6816.18
Secchi ~ Chla1<2 × 10−1631.30%38.87.3
Secchi ~ Pcy + Chla10.0028838.70%31.50
13.82 × 10−5
Table 3. HGAM results from the “historical” dataset exploring the functional relationship between ZSD and Chla. The two periods contrasted are: 1984–1999 (20th century) and 2000–2001 (21st century). Two slightly different model structures are presented (S and I, see Materials and Methods section). p-values associated with each term are listed in their order of appearance in the mathematical function; p-values < 0.05 are statistically significant; AIC, Akaike Information Criterion; delta AIC is the difference between a given model AIC with respect to the lowest AIC model; edf, estimated degrees of freedom for *, 20th century, 21st century.
Table 3. HGAM results from the “historical” dataset exploring the functional relationship between ZSD and Chla. The two periods contrasted are: 1984–1999 (20th century) and 2000–2001 (21st century). Two slightly different model structures are presented (S and I, see Materials and Methods section). p-values associated with each term are listed in their order of appearance in the mathematical function; p-values < 0.05 are statistically significant; AIC, Akaike Information Criterion; delta AIC is the difference between a given model AIC with respect to the lowest AIC model; edf, estimated degrees of freedom for *, 20th century, 21st century.
Modeledfp-Value% DevianceAICDelta AIC
S: logZSD = ƒcentury(logChla) +ε4.34<2 × 10−1646.271.521.4
I: logZSD = ƒ((logChla):century) +1 *5.22 × 10−646.470.120
ζ century +ε2.36·<2 × 10−16
0.975.51 × 10−7
Table 4. Estimated ZSD from shallow Pampean lakes according to model I, before (20th century) and after (21st century) the adoption of GR crops.
Table 4. Estimated ZSD from shallow Pampean lakes according to model I, before (20th century) and after (21st century) the adoption of GR crops.
Chla ConcentrationPredicted Secchi Disc Depth (cm)
(µg L−1)20th Century ZSD21st Century ZSD21st Century ZSD as % of 20th Century ZSD
779.440.751.3
5038.018.247.9
20023.412.955.0
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Zagarese, H.E.; Diovisalvi, N.R.; González Sagrario, M.d.l.Á.; Izaguirre, I.; Fermani, P.; Unrein, F.; Castro Berman, M.; Pérez, G.L. Increases in Picocyanobacteria Abundance in Agriculturally Eutrophic Pampean Lakes Inferred from Historical Records of Secchi Depth and Chlorophyll-a. Water 2022, 14, 159. https://doi.org/10.3390/w14020159

AMA Style

Zagarese HE, Diovisalvi NR, González Sagrario MdlÁ, Izaguirre I, Fermani P, Unrein F, Castro Berman M, Pérez GL. Increases in Picocyanobacteria Abundance in Agriculturally Eutrophic Pampean Lakes Inferred from Historical Records of Secchi Depth and Chlorophyll-a. Water. 2022; 14(2):159. https://doi.org/10.3390/w14020159

Chicago/Turabian Style

Zagarese, Horacio Ernesto, Nadia R. Diovisalvi, María de los Ángeles González Sagrario, Irina Izaguirre, Paulina Fermani, Fernando Unrein, Manuel Castro Berman, and Gonzalo Luis Pérez. 2022. "Increases in Picocyanobacteria Abundance in Agriculturally Eutrophic Pampean Lakes Inferred from Historical Records of Secchi Depth and Chlorophyll-a" Water 14, no. 2: 159. https://doi.org/10.3390/w14020159

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop