Lake water chemistry and population of origin interact to shape fecundity and growth in Daphnia ambigua

Abstract Freshwater environments vary widely in ion availability, owing to both natural and anthropogenic drivers. Field and laboratory work point to the importance of overall salinity, as well as cation depletion, in shaping the physiology, behavior, and ecology of freshwater taxa. Yet, we currently have a poor understanding of the degree to which populations may vary in response to ion availability. Using Daphnia collected from three lakes that differ greatly in salinity and calcium availability, we conducted a laboratory reciprocal transplant experiment to assess how animals representing these populations vary in fecundity, body size, and survival when reared in lake water from each environment. The lake water environment and population of origin strongly interacted to shape Daphnia growth and reproduction. Surprisingly, we found only modest evidence that lake water with abundant calcium (5.5 vs. 1.2–2.3 mg/L) increased Daphnia growth or reproduction. By contrast, water from a relatively ion‐rich lake (400 μS/cm specific conductance) strongly boosted Daphnia fecundity over lower‐ion lake water (20–50 μS/cm), especially for the population originating from the high‐ion environment. Our results suggest that ion‐poor conditions common in regions around the world may exert stress on freshwater organisms, even for populations inhabiting these environments. Meanwhile, moderate salt enrichment may not prove harmful but could even benefit freshwater taxa in these ion‐poor regions. The context dependence of how and when lake water chemistry affects Daphnia and other freshwater taxa deserves greater attention, in both ion‐depleted and ion‐rich conditions. Daphnia are key members of lake food webs and serve as an important model for ecology, evolution, and toxicology research. Consideration of how lake water chemistry may influence how Daphnia populations respond to abiotic and biotic stress may improve the ability to evaluate and predict ecological and evolutionary dynamics in lakes of varying chemical composition.


| INTRODUC TI ON
Salt availability varies widely across freshwater environments around the world and is shaped by both natural and anthropogenic drivers. Human activities including application of road salt during winter storms, changes in precipitation and evaporation patterns, and sea level rise have led to widespread salinization of freshwater bodies (Dugan et al., 2017;Estévez et al., 2019;Kaushal et al., 2018). Meanwhile, deforestation and acidic deposition resulting from industrial activity have resulted in the depletion of some essential cations (e.g., calcium and magnesium) from surface waters (Hessen et al., 2017;Jeziorski et al., 2008;Kahl et al., 1991;Norton et al., 1989;Stoddard et al., 1999). These anthropogenic influences overlay natural gradients in ion availability driven by geology, climate, and proximity to the coast (Estévez et al., 2019;Huser et al., 2020;Norton et al., 1989). Major ions (Na + , Ca 2+ , Mg 2+ , K + , and Cl − ) are essential micronutrients, needed for survival, growth, and reproduction (Kaspari, 2021). These ions play a key role in osmoregulation, as well as nervous system and motor system functioning, while calcium also serves as an important structural micronutrient (Bell et al., 1993;Cairns & Yan, 2009;Metz et al., 2014;Spence et al., 2012). At the same time, elevated concentrations of dissolved ions can be toxic to freshwater organisms (Brauner et al., 2012;Griffith, 2017). Determining the extent to which the ion landscape shapes ecological and evolutionary dynamics has emerged as a critical area of research (Kaspari, 2021).
Both ion-poor and extremely ion-rich freshwater environments are likely to impose physiological stress on freshwater biota.
All freshwater organisms contend with the osmo-ionoregulatory challenges of limiting water uptake and preventing ion loss to the surrounding environment. Most freshwater animals, including fish and crustaceans, are hyper-osmoregulators (Griffith, 2017)-they maintain ion concentrations higher than the surrounding waters by actively pumping ions (mainly Na + and Cl − ) across permeable membranes, minimizing ion leakage, and excreting excess water.
Osmoregulation in ion-poor environments is metabolically costly (Glazier & Sparks, 1997), and this stress may be further exacerbated by aqueous calcium depletion, which increases the permeability of gill epithelia (Brauner et al., 2012;Gundersen & Curtis, 1995).
Calcium also serves as a key structural micronutrient, important for building bone and the chitinous exoskeleton in crustaceans.
Calcium limitation leads to increased mortality, reduced body size, slower maturation, and reduced reproductive output in a variety of crustacean taxa (Ashforth & Yan, 2008;Cairns & Yan, 2009;Hessen et al., 2000). Low-calcium availability may also increase vulnerability to predation, for example, by limiting the formation of inducible defenses (Riessen et al., 2012) and reducing exoskeleton rigidity (Edwards et al., 2015). Exposure to elevated concentrations of dissolved ions can also prove stressful, with toxicity occurring when environmental concentrations approach or exceed that of intracellular fluids and hemolymph (Griffith, 2017). Thus, the cost of maintaining physiological homeostasis should shift across gradients of ion availability, with increased metabolic costs expected in freshwater environments at either extreme, and with calcium availability playing a particularly important role in supporting survival, growth, and reproduction.
Variation among taxa in the ability to tolerate ion availability likely plays a role in shaping community structure. Both ion-poor and extremely ion-rich freshwater environments tend to support lower diversity and altered species composition, with these trends observed across a range of taxa, including bacteria, zooplankton, mollusks, macroinvertebrates, and plants (Gutierrez et al., 2018;Piscart et al., 2005;Sowa et al., 2019). Field observations of shifting community structure along ion gradients may be difficult to disentangle from other abiotic or biotic variables that correlate with ion availability, such as pH, macronutrient availability, pollution, or changes in predation regime (Griffiths et al., 2019). Yet experimental studies support that ion availability per se can play an important role in shaping freshwater communities (Arnott et al., 2017;Lind et al., 2018;Sinclair & Arnott, 2018). In low-ion environments, calcium availability appears to be an important determinant of zooplankton population dynamics and community structure (Arnott et al., 2017;Durant et al., 2018); species with higher calcium needs tend to be absent or rare, while species with low calcium needs may become more competitive and dominant (Durant et al., 2018;Edwards et al., 2015;Jeziorski et al., 2008Jeziorski et al., , 2012Jeziorski et al., , 2014Waervågen et al., 2002). Daphniids are especially vulnerable to calcium depletion (Cairns & Yan, 2009), likely owing to the need to frequently replace their relatively high calcium exoskeleton following each molt during their life cycle (Alstad et al., 1999); they rely primarily on aqueous calcium to meet these demands (Tan & Wang, 2009). At the other end of the salinity spectrum, while many species are lost, some salt-tolerant species thrive in habitats naturally rich in ions (Michels et al., 2003;Ortega-Mayagoitia et al., 2000). Such salt-loving taxa have been observed colonizing environments experiencing anthropogenic salt inputs, even great distances from their typical geographic range (e.g., Hairston et al., 2005;Kipriyanova et al., 2007).
Given the significant physiological stress and ecological impacts associated with both ion-poor and high-salinity conditions, ion availability should exert selection pressure on freshwater taxa. Limited examination of intraspecific phenotypic variation among natural populations has provided some evidence of evolution in response to ion availability. Local adaptation to elevated salinity was observed in Daphnia pulex and Simocephalus vetulus populations inhabiting ponds impacted by sea spray or saltwater intrusion (Loureiro et al., 2012;Weider & Hebert, 1987). Experimental evolution work showed Daphia pulex can evolve tolerance for elevated sodium chloride (NaCl) in only 5-10 generations (Coldsnow et al., 2017). In addition, Ambystoma maculatum (spotted salamander) populations inhabiting roadside vernal pools showed evidence of local adaptation to road salt (Brady, 2012). Yet surprisingly Lithobates sylvaticus (wood frog) populations in the same wetlands displayed clear evidence of local maladaptation to road salt stress (Brady, 2013). Ion-poor freshwater environments also appear to drive a range of evolutionary responses. Low-calcium conditions may select for reduced investment in pelvic size in three-spined stickleback (Gasterosteus aculeatus) in the absence of strong predation pressure (Bell et al., 1993;Spence et al., 2012). In addition, Daphnia pulicaria from seven Canadian Shields lakes showed interclonal differences in response to calcium limitation but with little relation to the population of origin (Overhill, 2017). Meanwhile, Daphnia galeata from a low-calcium lake were more sensitive to calcium restriction relative to those from a higher calcium environment (Alstad Rukke, 2002). This evidence supports that both ion-poor and ion-enriched environments may shape the evolutionary trajectories of freshwater organisms; however, the variation in responses among taxa, including both adaptive and maladaptive responses to ion availability, points to the need for further study.
Here, we used a laboratory reciprocal transplant experiment to examine how Daphnia ambigua isolated from three natural populations differed in response to being reared in lake water with widely varying ion availability and composition: two inland lakes, one of which is especially ion-poor, and a coastal lake where sea salt influence has caused both ion enrichment and a shift in ion composition ( Figure 1). This design allowed us to determine the extent to which these varying lake water environments influenced Daphnia growth and reproduction and whether the population of origin interacted to shape these responses.

| Study lake selection
Both natural and anthropogenic factors influence the ion chemistry of lakes in our study region of Maine, USA. Bedrock, soil properties, hydrology, and watershed activities influence the availability of important cations like calcium (Ca 2+ ) and (Mg 2+ ), while sodium (Na + ) and chloride (Cl − ) availability increases with proximity to the coast, as well as anthropogenic inputs in the watershed (Norton et al., 1989).
Overall, ion availability in Maine lakes tends to be low relative to other lakes in the United States (National Lakes Assessment, 2007) ( Figure S1). In addition, of the 688 lakes regularly sampled by the Maine Department of Environmental Protection (DEP) in the mid-1990s-early 2010s (unpublished data), about 10% have concentrations of Ca 2+ below 1.5 mg/L (Figure 2), the minimum concentration thought to support Daphnia pulicaria in ideal laboratory conditions (Ashforth & Yan, 2008). Using this historic data collected by the Maine DEP, we selected three study lakes of similar size (17-29 hectares) and hydroperiod (permanent) that represent some of the breadth of this variation in ion availability and composition ( Figure 2) and support populations of our focal species, Daphnia ambigua. Hall Pond, with a specific conductance in the 10th percentile, is ion-poor even compared with other Maine lakes ( Figure S1). By contrast, Sewall Pond is a relatively ion-rich lake that receives sea salt from a brackish tidal creek (Figure 2a,c). Sewall Pond resembles other coastal Maine lakes (found within 10 km of the coast) in that Ca 2+ tends to contribute relatively less to overall ion availability, while NaCl provides a greater contribution (Figure 2a, Figure S2). Finally, Egypt Pond has intermediate ion availability but with calcium levels far exceeding those in the other two study lakes (Figure 2b). The three study lakes are separated from one another by a geographic distance of at least 45 km ( Figure S3).

| Establishment of Daphnia clonal lineages
Our focal species, Daphnia ambigua, is a small Daphnia species with a widespread geographic distribution, found in lakes throughout North America, South America, and Europe, across a wide range of environmental gradients (Hebert et al., 2003). We established clonal cultures of D. ambigua from each study lake in late-May 2019 from net collections (80μm mesh, 20-cm diameter tow net, Wildco) sampling the full water column at the deepest basin of the lake.
Zooplankton samples were kept in unfiltered lake water from the sampling site, in cool, dark conditions, and were processed within 2 h of collection. We examined the live zooplankton samples under a dissecting microscope at 10× magnification, isolating each female D. ambigua individual from random subsamples until at least 12 individuals were collected from each lake. Of the Daphnia that we collected that survived to reproduce, we randomly selected four clonal lineages from each lake to maintain for this study, using juveniles from the second to fifth brood to initiate each new generation.

| Study lake characterization
In May, June, and July 2019 we sampled each lake to characterize its chemical and physical properties (Table 1)  suppressed conductivity (Hautman et al., 1997). Total nitrogen and total phosphorus (TN and TP) were analyzed using alkaline persulfate digestion (Patton & Kryskalla, 2003). We measured chlorophyll a concentrations from an integrated water sample collected from the photic zone (the surface to twice the Secchi depth) in late-May and early-July 2019. Duplicate water subsamples were passed through glass fiber filters (Whatman GF/F, 0.7 μm pore size) and wet filters were frozen for 3 weeks before chlorophyll extraction (over 24 h in 10 mL 90% acetone at −20°C) and fluorometric analysis (Turner

| Daphnia culturing conditions
Water for Daphnia culture maintenance was collected from each study lake in 20-L polyethylene carboys, passed through glass fiber filters (Pall A/E, 1.0 μm pore size), and stored at 4°C before use. We maintained each Daphnia individual in a borosilicate beaker with 25 mL filtered water (warmed to room temperature) from their lake of origin. Daphnia cultures were maintained at 20°C with a 16-h light:8-h dark period. We changed the lake water and removed offspring twice weekly, and Daphnia were each fed 500,000 cells of Ankistrodesmus falcatus four times weekly. Ankistrodesmus was cultured in heat sterilized modified ASM-1 medium (Goulden & Hornig, 1980) at room temperature, harvested weekly during the logistic growth phase, and stored at 4°C. A vitamin mixture (Goulden et al., 1982) was added to the algal culture to support Daphnia nutritional needs. Algal media and vitamin supplement constituents were reagent grade (Fisher Scientific). Our Ankistrodesmus strain originated from the lab of C. E. Goulden at the Academy of Natural Sciences in Philadelphia, PA, USA and has been designated the AJT strain (Schomaker & Dudycha, 2021). To ensure consistent feeding levels throughout the experiment and to minimize any effects of feeding on water chemistry, algae were settled and resuspended with filtered lake water, maintaining a density of 1,000,000 cells/mL using a hemocytometer. Daphnia were fed resuspended algae mixed with the same lake water used for their culturing.

| Experimental design
To evaluate how Daphnia fitness is influenced by lake water chemistry and the extent to which Daphnia from the three populations varied in their response, we performed a laboratory reciprocal transplant experiment. Using four Daphnia clonal lineages collected from the three study lakes (low-ion Hall Pond, high-Ca 2+ Egypt Pond, and ion-rich Sewall Pond), we compared growth, reproduction, and survival over the course of 21 days, with genetically identical replicates reared in filtered lake water from one of each of the three study lakes.
Before conducting the experiment, Daphnia were reared in filtered lake water from their lake of origin for 3-5 generations in CelSens Entry imaging software) at 50× magnification. We measured body length from the center of the eye to the base of the tail-spine using ImageJ. We ended the experiment after 21 days and collected a second set of body size measurements for each surviving experimental animal. While body size was not measured for neonates, these measurements of body length at days 7 and 21 allowed us to examine the influence of the lake water environment on investment in growth, when accounting for variation among populations and clones.

| Statistical analyses
We evaluated whether the lake water environment, population of origin, or the interaction of these two variables predicted aspects of growth, reproduction, or survival using generalized linear mixed models (GLMMs). We selected the best fit model by comparing the Akaike information criterion (AIC) of the saturated model with nested simpler models containing fewer fixed effects. We used likelihood ratio tests to select which fixed effects to include in the final model, based on Zuur et al. (2009). We evaluated the normality of the residuals of the selected models using visual inspection of quantile plots and Kolmogorov-Smirnov tests.
Models examining variation in body size and growth included either body length at 7 days (around the timing of first reproduction for most animals), length at 21 days (the end of the trial), or the growth that occurred between days 7 and 21 ([size at 21 days − size at 7 days]/size at 7 days) as the response variable. We applied an arcsine square root transformation of the proportion growth to improve the normality of the residuals. To assess investment in reproduction, we examined whether reproduction had occurred by the 7th day of the trial (binomial response variable) and the total number offspring produced during the 21-day trial. Daphnia that died before day 7 of the trial (N = 10 of 329 replicates) were excluded from analyses of growth and reproduction. In a separate GLMM, we evaluated whether the likelihood of survival to maturity (as a binomial variable) was predicted by water source, population of origin, or their interaction.
In each GLMM, clonal lineage was included as a random intercept. All analyses were conducted using the statistical program R (v. 4.2.1) (R Core Team, 2016). Mixed effects modeling was conducted using the "glmmTMB" package (Brooks et al., 2017). Differences in growth, reproduction, and survival among all pairwise lake and treatment comparisons were evaluated using post hoc tests of Tukey's Honest Significant Difference (HSD) with the "emmeans" package.

| Body size
Seven days into the trial, Daphnia body length was strongly influenced by lake of origin (p < .001, χ 2 = 32.854 AIC = -958.70, Figure 3a). Daphnia from Egypt Pond (high-Ca 2+ lake) were 6.7% larger than those from Hall Pond (low-ion lake) (Tukey HSD, p < .001, t ratio = 5.340), which were 7.4% larger than animals from Sewall Pond (ion-rich lake) (Tukey HSD, p < .001, t statistic = 5.589). The lake water environment did not significantly affect body size meas- Ponds grew at a similar rate when comparing growth in their home lake water environment and the transplant treatments (Figure 3c).

| Reproduction
The proportion of Daphnia that reproduced by the seventh day ranged from 36% to 100% across clones and treatments. This ten-

| Survival to maturity
Early survival rates were high, with 97% of the 329 replicates surviving to maturity ( Figure 5). There was modest support for lake water treatment explaining this variation (p = .037, χ 2 = 6.5983, F I G U R E 3 Body size and growth rate varied depending on lake of origin, the lake water environment, and their interaction. (a) 7-day mean body size of Daphnia ambigua clonal replicates from the three study populations (Hall Pond = low-ion, Egypt Pond = high-Ca 2+ , Sewall Pond = ion-rich) reared in lake water collected from each study lake. (b) Final (21-day) mean body size of the same individuals. (c) Proportion increase in body size that occurred over the final 2 weeks of the experiment ([21-day size − 7-day size]/7-day size). Error bars represent ±1 standard error of the mean across individuals from all clones (N = 30-40 total replicates from four clones/lake). Lake water environment is shown on the x axis and colors represent the population of origin. 'a-d' designate significant pairwise differences (p < .05) among groups in post hoc Tukey HSD tests of GLMM results, where the same letter indicates no significant difference. F I G U R E 5 Rates of survival to maturity for Daphnia from the three study populations (Hall Pond = low-ion, Egypt Pond = high-Ca 2+ , Sewall Pond = ion-rich) reared in lake water collected from each study lake. Each data point represents the proportion of Daphnia that survived to reproduce for each clonal lineage × treatment combination. Data points are slightly offset to increase visibility. Cairns & Yan, 2009;Hessen et al., 2000;Jeziorski et al., 2012).

Calcium may serve as a limiting structural micronutrient for
Daphnia (Ashforth & Yan, 2008;Hessen et al., 2000), and low aqueous calcium is expected to increase the metabolic demands of osmoregulation in low-ion freshwaters (Brauner et al., 2012).
However, we observed little benefit to Daphnia ambigua growth or reproduction from Egypt Pond's high-Ca 2+ lake water environment. Laboratory studies have shown that Daphnia sp. tend to see a reduction in fitness between calcium levels of 0-2 mg/L, with levels below 1-0.5 mg/L lethal (Cairns & Yan, 2009). Above By contrast, lake water from ion-rich Sewall Pond provided a measurable fecundity benefit for Daphnia from all three lakes.
Daphnia from both Hall Pond and Egypt Pond produced more offspring in Sewall Pond water compared with their natal lower-ion lake water environments, and animals from Sewall Pond were much less fecund and smaller in size when reared in water from the two lower-ion lakes. In addition, Daphnia reared in Sewall Pond water were more likely to reproduce by the seventh day of the trial. This suggests that Sewall Pond's water provides a benefit that water from the two lower-ion lakes lacks. While calcium concentrations are moderate in Sewall Pond, magnesium, potassium, and sodium are orders of magnitude higher than levels in both lower-ion lakes.
These salts or some other unmeasured micronutrient may be limiting in the lower-ion lake waters. In addition, the metabolic costs of osmoregulation may be reduced in Sewall Pond's higher salinity water.
Salt toxicity may not be expected to occur until ion levels approach or exceed concentrations in the hemolymph (Griffith, 2017), and ion levels in Sewall Pond (Table 1) are below concentrations measured in hemolymph of Daphnia ambigua reared experimentally in low-ion conditions (Durant et al., 2018). In addition, dissolved organic carbon concentrations in Sewall Pond's water exceed levels found in the other two study lakes (Table 1). Experimental work suggests that lake browning may support increased survival and reproduction in Daphnia longispina (Minguez et al., 2020), though browning agents extracted from leonardite may poorly replicate the effects of naturally derived organic carbon on lake organisms (Scharnweber et al., 2021). Some combination of these factors, or another unmeasured difference in Sewall Pond's water may benefit Daphnia fitness over the low-ion lake water environments. Research on the impacts of salt enrichment in freshwater environments has often focused on the stress caused by very high salinity resulting from sea salt or road salt pollution (Brady, 2012(Brady, , 2013Weider & Hebert, 1987). Our results suggest that in regions where ion-poor conditions dominate, moderate introduction of sea salt may not prove harmful and could even benefit the resident freshwater taxa.
A central aim of this study was to explore not only the extent to which lake water chemistry influences Daphnia growth and reproduction but also the degree to which populations from divergent environments might vary in their responses. Overall, the environment that was most supportive of Daphnia growth and reproduction was Sewall Pond's ion-rich water, and Daphnia originating from Sewall Pond were most strongly impacted by the stress of the lower-ion environments. This evidence supports that Daphnia from Sewall Pond are adapted to their local lake water chemistry.
By the end of the 21-day trial, Sewall Pond Daphnia were larger and much more fecund in their natal lake water compared with low-ion lake waters from Hall or Egypt Ponds (27%-34% increase in total offspring). Daphnia from the two lower-ion lakes also showed increased fecundity in Sewall Pond's high-ion environment compared with their natal lake water, but the differences were more modest (17%-19% increase). Meanwhile, we see no evidence that Daphnia from Egypt and Hall Pond are adapted to differences in their natal lake water. Both populations showed a similar pattern of a modest (nonsignificant) increase in fecundity in Egypt Pond's high-Ca 2+ environment. It is possible that if we allowed for multiple generations of acclimation, the study animals would find their transplant conditions more stressful (low-Ca 2+ ) or beneficial (high-Ca 2+ ), increasing the differences observed in contrasting calcium environments (Giardini et al., 2015). However, findings from a subsequent trial also showed a modest fecundity benefit of Egypt Pond's higher-Ca 2+ environment after a generation of acclimation, with no difference seen between the two populations ( Figure S5). to individual stressors associated with lake water chemistry and pollution using artificial lake water as a test medium (e.g., COMBO [Kilham et al., 1998], ADaM [Klüttgen et al., 1994], Elendt M7 [Samel et al., 1999], FLAMES [Celis-Salgado et al., 2008]). The value in this approach lies in the ability of researchers to identify causal factors responsible for any observed differences among treatments.
However, if Daphnia are sensitive to differences (whether measured or unknown) between their natal lake water and the artificial medium, the findings generated from these assays may not provide a fair assessment of how organisms embedded in their natural environment may be impacted by the variable of interest. Lake water chemistry may also strongly influence responses to other stressors, including predation (Bell et al., 1993;Huang et al., 2021;Riessen et al., 2012;Spence et al., 2012), parasitism (Buss & Hua, 2018;Milotic et al., 2017;Stockwell et al., 2015), temperature (Ashforth & Yan, 2008) Overall, our study supports that variation in ion availability and composition may strongly influence Daphnia growth and reproduction and that populations may vary in the strength of this response.
The context dependence of how and when lake water chemistry affects Daphnia deserves greater attention, in both ion-depleted environments and ion-rich conditions. Individual ions may be important limiting micronutrients, but ion ratios and overall ion strength may also play a critical role in supporting Daphnia health or protecting Daphnia from toxic conditions (Celis-Salgado et al., 2016;Davies & Hall, 2007;Elphick et al., 2011). At the same time, the source of ion enrichment is likely to strongly influence the degree to which added salts might be beneficial, as activities such as road salt application, agriculture, mining, and wastewater disposal are likely to be accompanied by harmful toxicants and may involve complex biogeochemical interactions (Kaushal et al., 2022). Daphnia serve as an important model system for ecology, evolutionary biology, and toxicology research, while also playing a critical role in lake food webs as regulators of phytoplankton productivity and food for fish and larger invertebrates (Lampert, 2011;Miner et al., 2012). Our findings suggest that accounting for the effects of natural variation in lake water chemistry may help to uncover previously unrecognized patterns in Daphnia ecology and evolution.

ACK N OWLED G M ENTS
We wish to thank Elizabeth Baker, Susannah Lawhorn, and Martha Mixon for field and laboratory assistance, and Aaron Gilbreath for statistical consultation. We thank four anonymous reviewers for helpful recommendations that improved this manuscript. This work was supported by startup funds and a Faculty Research Award provided by Bowdoin College (Rogalski) and a Freedman Summer Research Fellowship in Coastal/Environmental Studies (Ferah).

This study was supported by Bowdoin College and the Freedman
Summer Research Fellowship in Coastal/Environmental Studies.

CO N FLI C T O F I NTER E S T S TATEM ENT
None.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are openly available