Wetland water quality patterns and anthropogenic pressure associations across the continental USA.

Anthropogenic impacts on lake and stream water quality are well established but have been much less studied in wetlands. Here we use data from the 2016 National Wetland Condition Assessment to characterize water quality and its relationship to anthropogenic pressure for inland wetlands across the conterminous USA. Water samples obtained from 525 inland wetlands spanned pH from <4 to >9 and 3 to 5 orders of magnitude in ionic strength (chloride, sulfate, conductivity), nutrients (total N and P), turbidity, planktonic chlorophyll, and dissolved organic carbon (DOC). Anthropogenic pressure levels were evaluated at two spatial scales - an adjacent scale scored from field checklists, and a catchment scale indicated by percent agricultural plus urban landcover. Pressure at the two spatial scales were uncorrelated and varied considerably across regions and wetland hydrogeomorphic types. Both adjacent- and catchment-scale pressure were associated with elevated ionic-strength metrics; chloride elevation was most evident in road-salt using states, and sulfate was strongly elevated in a few sites with coal mining nearby. Nutrients were elevated in association with catchment-scale pressure but concomitant changes were not seen in planktonic chlorophyll. Acidic pH and high DOC occurred primarily in upper Great Lakes and eastern seaboard sites having low anthropogenic pressure, suggesting natural organic acid sources. Ionic strength and nutrients increased with increasing catchment-scale pressure even in Flats and closed Depression and Lacustrine sites, which indicates connectivity to rather than isolation from upland anthropogenic landuse even for wetlands lacking inflowing streams.

Invasive Dreissena Mussel Coastal Transport From an Already Invaded Estuary to a Nearby Archipelago Detected in DNA and Zooplankton Surveys.

Coastal waters of Lake Superior are generally inhospitable to the establishment of invasive Dreissena spp. mussels (both Dreissena polymorpha and Dreissena bugensis). Dreissena have inhabited the Saint Louis River estuary (SLRE; largest commercial port in the Laurentian Great Lakes) for over three decades, but only in the last few years have small colonies been found in the Apostle Islands National Lakeshore (APIS, an archipelago situated 85 km to the east of SLRE) A 2017 survey determined a low abundance Dreissena spatial distribution in APIS, with the largest colonies on the north and west islands which suggested potential veliger transport from the SLRE via longshore currents. Our objective in this study was to determine if Dreissena veligers are transported by currents at low densities along the south shore of Lake Superior from the SLRE to APIS. To do so, we used both eDNA (water and passive substrate samples) and zooplankton collection methods at eight sites evenly spaced between the SLRE and APIS with three sampling times over five weeks. Dreissena veligers were consistently detected along the south shore, although at low abundances (veligers per m3 range = 0-690, median = 8), and for every 1 km increase in distance from the SLRE, both veliger counts and water eDNA copy numbers decreased on average by 5 and 7%, respectively. D. polymorpha (suited to estuary habitats) was detected two times more than D. bugensis (better suited to deep-lake habitats). There was not a trend in the veliger size distribution along the south shore, and temperature and calcium concentrations fluctuated around the threshold for Dreissena veliger and adult development, averaging 11.0°C and 14.8 ppm, respectively. Three zooplankton taxa representative of the estuary community-Daphnia retrocurva, Diaphanosoma birgei, and Mesocyclops copepodites-decreased as the distance from the SLRE increased mirroring Dreissena veliger abundance patterns. Findings represent multiple sources of evidence of a propagule "conveyor belt" for Dreissena along the south shore of Lake Superior. We conclude that veligers are functioning as a propagule, using coastal currents to spread from the point of invasion, thereby traversing coastal habitat previously reported as inhospitable to distant habitats suitable for colonization.

Evaluating the performance of DNA metabarcoding for assessment of zooplankton communities in Western Lake Superior using multiple markers.

For DNA metabarcoding to attain its potential as a community assessment tool, we need to better understand its performance versus traditional morphological identification and work to address any remaining performance gaps in incorporating DNA metabarcoding into community assessments. Using fragments of the 18S nuclear and 16S mitochondrial rRNA genes and two fragments of the mitochondrial COI marker, we examined the use of DNA metabarcoding and traditional morphological identification for understanding the diversity and composition of crustacean zooplankton at 42 sites across western Lake Superior. We identified 51 zooplankton taxa (genus or species, depending on the finest resolution of the taxon across all identification methods), of which 17 were identified using only morphological traits, 13 using only DNA and 21 using both methods. The taxa found using only DNA metabarcoding included four species and one genus-level identification not previously known to occur in Lake Superior, the presence of which still needs to be confirmed. A substantial portion of taxa that were identified to genus or species by morphological identification, but not identified using DNA metabarcoding, had zero ("no record") or ≤ 2 ("underrepresented records") reference barcodes in the BOLD or NCBI databases (63% for COI, 80% for 16S, 74% for 18S). The two COI marker fragments identified the most genus- and species-level taxa, whereas 18S was the only marker whose family-level percent sequence abundance patterns showed high correlation to composition patterns from morphological identification, based on a NMDS analysis of Bray-Curtis similarities. Multiple replicates were collected at a subset of sites and an occupancy analysis was performed, which indicated that rare taxa were more likely to be detected using DNA metabarcoding than traditional morphology. Our results support that DNA metabarcoding can augment morphological identification for estimating zooplankton diversity and composition of zooplankton over space and time, but may require use of multiple markers. Further addition of taxa to reference DNA databases will improve our ability to use DNA metabarcoding to identify zooplankton and other invertebrates in aquatic surveys.

Comparison of Larval Fish Detections Using Morphology-Based Taxonomy versus High-Throughput Sequencing for Invasive Species Early Detection.

When first introduced, invasive species typically evade detection; DNA barcoding coupled with high-throughput sequencing (HTS) may be more sensitive and accurate than morphology-based taxonomy, and thereby improve invasive (or rare) species detection. We quantified the relative error of species detection between morphology-based and HTS-based taxonomic identification of ichthyoplankton collections from the Port of Duluth, Minnesota, an aquatic non-native species introduction 'hot-spot' in the Laurentian Great Lakes. We found HTS-based taxonomy identified 28 species and morphology-based taxonomy 30 species, of which 27 were common to both. Among samples, 76% of family-level taxonomic assignments agreed; however, only 42% of species assignments agreed. Most errors were attributed to morphology-based taxonomy, whereas HTS-based taxonomy error was low. For this study system, for most non-native fishes, the detection probability by randomized survey for larvae was similar to that by a survey that is optimized for non-native species early detection of juveniles and adults. We conclude that classifying taxonomic errors by comparing HTS results against morphology-based taxonomy is an important step toward incorporating HTS-based taxonomy into biodiversity surveys.

Early detection monitoring for non-indigenous fishes; comparison of survey approaches during two species introductions in a Great Lakes port.

Assessing relative performance of different sampling methods used for early detection monitoring (EDM) is a critical step in understanding the likelihood of detecting new non-indigenous species (NIS) in an environment of interest. EDM performance metrics are typically based on the probability of detecting established NIS or rare indigenous species; however, detection probability estimates for these proxies may not accurately reflect survey effectiveness for newly introduced NIS. We used data from three different EDM survey approaches that varied by targeted life-stage (adult-juvenile versus ichthyoplankton), media (physical fish versus environmental DNA), and taxonomic method (morphology-based versus DNA-based taxonomy) to explore relative detection sensitivity for recently introduced white bass (Morone chrysops) and gizzard shad (Dorosoma cepedianum) in the Port of Duluth-Superior, a NIS introduction hot spot within the Laurentian Great Lakes. Detection efficiency, measured by the effort (number of samples) required to achieve 95% probability of detection, differed by EDM approach and species. Also, the relative sensitivity (detection rate) of each survey approach differed by species. For both species, detection in surveys using DNA-based taxonomy was generally as good or better than the adult-juvenile survey using morphology-based taxonomy. While both species appear to have been detected at early stages of invasion, white bass were likely present up to 5 years prior to initial detection, whereas gizzard shad may have been detected in the first year of introduction. We conclude that using complimentary sampling methods can help to balance the strengths and weaknesses of each approach and provide more reliable early detection of new invaders.

Spatial and temporal distributions of Dreissena spp. veligers in Lake Huron: does calcium limit settling success?

The larval stage of invasive Dreissena spp. mussels (i.e., veligers) are understudied despite their seasonal numerical dominance among plankton. We report the spring and summer veliger densities and size structure across the main basin, North Channel, and Georgian Bay of Lake Huron, and seek to explain spatiotemporal variation. Monthly sampling was conducted at 9 transects and up to 3 sites per transect from spring through summer 2017. Veliger densities peaked in June and July, and we found comparable densities and biomasses of veligers between basins, despite differences in density of juvenile and adult mussels across these regions. Using a generalized additive model to explain variations in veliger density, we found that temperature, chlorophyll a, and nitrates/nitrites were most important. We generated an index of veliger attrition based on size distributions that revealed a higher rate of attrition in the North Channel than the rest of the lake. A logistic model indicated a threshold calcium concentration of around 22 mg/L was necessary for veligers to survive to larger sizes and recruit to their juvenile and benthic adult life stages. Improved understanding of factors that regulate the production and survival of Dreissena veligers will improve the ability of managers to assess future invasion threats as well as explore potential control options.

Dreissena veligers in western Lake Superior - inference from new low-density detection.

The notion that Lake Superior proper is inhospitable to dreissenid mussel survival has been challenged by recent finds on shipwrecks and rocky reefs in the Apostle Islands region. Motivated by concerns surrounding these finds, we conducted an intensive sampling campaign of Apostle Islands waters in 2017 to understand Dreissena prevalence and distribution. The 100-site effort combined random and targeted sites and collected zooplankton, benthos, video, environmental DNA, and supporting water quality data. We did not find settled Dreissena in any video footage or benthos samples, and quantitative PCR applied to eDNA samples was negative for Dreissena. Dreissena veligers were found in almost half the zooplankton samples but at orders of magnitude lower densities than reported from other Laurentian Great Lakes. Veligers were most prevalent around the western islands and associated with shallower depths and slightly higher phosphorus and chlorophyll, but did not spatially match known (still very localized) settled Dreissena colonies. This is the first study to conduct veliger-targeted sampling in western Lake Superior and the first to report consistent detection of veligers there. We speculate that these Apostle Islands veligers are not a new locally-spawned component of the zooplankton community, but instead are transported from an established population in the St. Louis River estuary (~100 km away) by longshore currents; i.e., low-density propagule pressure that may have been present for years. Small-mesh zooplankton data collected along a gradient from the Apostle Islands to the St. Louis River estuary and enumerated with thorough veliger searching would help elucidate these alternatives.

Resolving taxonomic ambiguities: effects on rarity, projected richness, and indices in macroinvertebrate datasets.

Biodiversity information is an important basis for ecological research and environmental assessment, and can be impacted by choices made in the manipulation and analysis of taxonomic data. Such choices include methods for resolving multiple redundant levels of taxonomic resolution, as typically arise with morphological identification of damaged or immature aquatic macro-invertebrates. In particular, the effects of these processing choices on number of rare taxa are poorly understood yet potentially significant to the estimation of projected taxa richness and related evaluations such as biodiversity conservation value and survey sufficiency. Using aquatic macroinvertebrate data collected for two nearshore areas of Lake Superior, we determined how multiple methods of resolving taxonomic redundancies influence two commonly-used estimates of projected richness, Chao1 and Chao2, which hinge on the ratio of taxa that are singletons to doubletons (i.e., just one versus two individuals found) or uniques versus duplicates (i.e., just one versus two occurrences). We also determined how choice of ambiguous taxa method, including some modified specifically to retain rare taxa and others taken from the literature, influenced effort to reach 95% of projected richness, site-level richness and abundance, and representative invertebrate IBI scores. We found that Chao1 was more sensitive to method choice than Chao2, because singleton and doubleton status was more frequently affected when taxa were deleted, merged, or re-assigned in the process of resolving taxonomic redundancies than was unique and duplicate status. Methods that eliminated redundant taxa at the site scale but not the study-area scale tended to overinflate study area and projected richness, and resulted in a significant loss of abundance. The method that aggregated or deleted redundant taxa depending on abundance resulted in a decrease in site and study area richness, abundance, and an underestimation of projected richness. Methods which re-assigned parents to common children retained a majority of richness and abundance information and a more realistic estimate of projected taxa richness; however, the identity of poorly-identified parents was imputed. All methods resulted in little effect to typical IBI scores. Overall, no one method is fully capable of removing spurious richness at the study-area scale while preserving all taxa occurrence, abundance and rarity patterns. Therefore, the most appropriate method for making comparisons among sites may be different than the most appropriate method for comparing among surveys or among study areas, or if a goal is to estimate projected taxa richness or retain rare taxa information.

The response of wetland quality indicators to human disturbance indicators across the United States.

We analyzed data from 1138 wetland sites across the conterminous United States (US) as part of the 2011 National Wetland Condition Assessment (NWCA) to investigate the response of indicators of wetland quality to indicators of human disturbance at regional and continental scales. The strength and nature of these relationships in wetlands have rarely been examined over large regions, due to the paucity of large-scale datasets. Wetland response indicators were a multimetric index of vegetation condition (VMMI), percent relative cover of alien plant species, soil lead and phosphorus, and water column total nitrogen and total phosphorus. Site-level disturbance indices were generated from field observations of disturbance types within a circular 140-m radius area around the sample point. Summary indices were calculated representing disturbances for ditching, damming, filling/erosion, hardening, vegetation replacement, and vegetation removal. Landscape-level disturbance associated with agricultural and urban land cover, roads, and human population were based on GIS data layers quantified in 200, 500, and 1000-m circular buffers around each sample point. Among these three buffer sizes, the landscape disturbance indicators were highly correlated and had similar relationships with the response indictors. Consequently, only the 1000-m buffer data were used for subsequent analyses. Disturbance-response models built using only landscape- or only site-level disturbance variables generally explained a small portion of the variance in the response variables (R2 < 0.2), whereas models using both types of disturbance data were better at predicting wetland responses. The VMMI was the response variable with the strongest relationship to the disturbances assessed in the NWCA (national model R2 = 0.251). National multiple regression models for the soil and water chemistry and percent alien cover responses to disturbance indices were not significant. The generally low percentage of significant models and the wide variation in predictor variables suggests that stressor-response relationships vary considerably across the diversity of wetland types and landscape settings found across the conterminous US. Logistic regression modeling was more informative, resulting in significant national and regional models predicting site presence/absence of alien species and/or the concentration of lead in wetland soils above background.

USA-scale patterns in wetland water quality as determined from the 2011 National Wetland Condition Assessment.

Water quality is a central component of ecological assessments but less well characterized in wetlands than other waterbody types. The 2011 National Wetland Condition Assessment, spanning freshwater and brackish wetlands across the conterminous USA, provided an unprecedented opportunity to examine water quality patterns across broad wetland types and geographic scales. Surface water samples were obtained from 634 (56%) of sites visited. Total nitrogen (TN), total phosphorus (TP), planktonic chlorophyll (CHLA), and specific conductance (SPCOND) ranged 4 orders of magnitude across sites and were inter-correlated. Woody versus herbaceous vegetation type was an important classifier, with herbaceous sites having standing water more often and generally higher pH, nutrients, and CHLA. Nutrient ratios spanned a range from P-limited to N-limited in most biogeographic regions, and increasing TP was associated with decreasing TN:TP ratios. Compared to national-scale data for other waterbody types (lakes, streams, marine nearshore), wetlands had generally higher TN and TP but not higher CHLA. Differences among biogeographic regions in water quality were concordant between inland wetlands and lakes, and between marine-coast wetlands and the marine nearshore. Associations of TN, TP, and CHLA to percent agriculture or natural land were stronger for the watershed scale than for smaller concentric buffer scales, suggesting that wetlands are influenced by landuse some distance away. SPCOND was related to landuse in inland wetlands but reflected seawater influence in marine-coast wetlands. Water quality exhibits the same general patterns and responses across wetlands as across other waterbody types and thus can provide a basis for ecological classification and condition assessment.

A reference inventory for aquatic fauna of the Laurentian Great Lakes.

The Laurentian Great Lakes encompass an expansive and diverse set of freshwater ecosystems that contain a concordantly large and diverse vertebrate and invertebrate fauna. Although numerous publications exist concerning the composition and distribution of this fauna, there is at present no single readily available resource that brings all this information together. Here, we present and describe the compilation process for a comprehensive Great Lakes aquatic fauna inventory covering fishes, reptiles, amphibians, zooplankton, mollusks, annelids, insects, mites, and various other aquatic invertebrates. Inventory entries were developed via an extensive search of literature and internet sources and are attributed with detailed nomenclature information, general lake and habitat occurrences, and supporting citations and links to life history and genetic marker information. The inventory scope is the Laurentian Great Lakes proper and their connecting rivers, and their fringing coastal wetlands and lower tributaries. Over 2200 unique taxa are contained in the inventory -- 85% resolved to species and 14% to genus. The listing substantially expands previous richness estimates for invertebrates in the Great Lakes, but taxonomic resolution and spatial distribution information for them remains quite uneven. Example pattern analyses for fauna in this inventory show that aquatic vertebrates are generally more widely distributed than invertebrates, and that biodiversity is concentrated in the coastal margins. The inventory is being packaged into a public, searchable database that showcases the biodiversity of the Great Lakes aquatic fauna and can assist the research and management community in their biological investigations.

Early detection monitoring for aquatic non-indigenous species: Optimizing surveillance, incorporating advanced technologies, and identifying research needs.

Following decades of ecologic and economic impacts from a growing list of nonindigenous and invasive species, government and management entities are committing to systematic early- detection monitoring (EDM). This has reinvigorated investment in the science underpinning such monitoring, as well as the need to convey that science in practical terms to those tasked with EDM implementation. Using the context of nonindigenous species in the North American Great Lakes, this article summarizes the current scientific tools and knowledge - including limitations, research needs, and likely future developments - relevant to various aspects of planning and conducting comprehensive EDM. We begin with the scope of the effort, contrasting target-species with broad-spectrum monitoring, reviewing information to support prioritization based on species and locations, and exploring the challenge of moving beyond individual surveys towards a coordinated monitoring network. Next, we discuss survey design, including effort to expend and its allocation over space and time. A section on sample collection and analysis overviews the merits of collecting actual organisms versus shed DNA, reviews the capabilities and limitations of identification by morphology, DNA target markers, or DNA barcoding, and examines best practices for sample handling and data verification. We end with a section addressing the analysis of monitoring data, including methods to evaluate survey performance and characterize and communicate uncertainty. Although the body of science supporting EDM implementation is already substantial, research and information needs (many already actively being addressed) include: better data to support risk assessments that guide choice of taxa and locations to monitor; improved understanding of spatiotemporal scales for sample collection; further development of DNA target markers, reference barcodes, genomic workflows, and synergies between DNA-based and morphology-based taxonomy; and tools and information management systems for better evaluating and communicating survey outcomes and uncertainty.

A depth-adjusted ambient distribution approach for setting numeric removal targets for a Great Lakes Area of Concern beneficial use impairment: degraded benthos.

We compiled macroinvertebrate data collected from 1995 to 2014 from the St. Louis River Area of Concern (AOC) of Lake Superior. Our objective was to define depth-adjusted cutoff values for benthos condition classes to provide an analytical tool for quantifying progress toward achieving removal targets for the degraded benthos beneficial use impairment. We used quantile regression to model the limiting effect of depth on selected benthos metrics, including taxa richness, percent non-oligochaete individuals, combined percent Ephemeroptera, Trichoptera, and Odonata individuals, and density of ephemerid mayfly nymphs (Hexagenia). We created a scaled trimetric index from the first three metrics. Metric values above the 75th percentile quantile regression model prediction were defined as being in relatively excellent condition in the context of the degraded beneficial use impairment for that depth. We set the cutoff between good and fair condition as the 50th percentile model prediction, and we set the cutoff between fair and poor condition as the 25th percentile model prediction. We examined sampler type, geographic zone, and substrate type for confounding effects. Based on these analyses we combined data across sampler types and created separate models for each of three geographic zone. We used the resulting condition-class cutoff values to determine the relative benthic condition for three adjacent habitat restoration project areas. The depth-limited pattern of ephemerid abundance we observed in the St. Louis River AOC also occurred elsewhere in the Great Lakes. We provide tabulated model predictions for application of our depth-adjusted condition class cutoff values to new sample data.

Potential for DNA-based identification of Great Lakes fauna: match and mismatch between taxa inventories and DNA barcode libraries.

DNA-based identification of mixed-organism samples offers the potential to greatly reduce the need for resource-intensive morphological identification, which would be of value both to bioassessment and non-native species monitoring. The ability to assign species identities to DNA sequences found depends on the availability of comprehensive DNA reference libraries. Here, we compile inventories for aquatic metazoans extant in or threatening to invade the Laurentian Great Lakes and examine the availability of reference mitochondrial COI DNA sequences (barcodes) in the Barcode of Life Data System for them. We found barcode libraries largely complete for extant and threatening-to-invade vertebrates (100% of reptile, 99% of fish, and 92% of amphibian species had barcodes). In contrast, barcode libraries remain poorly developed for precisely those organisms where morphological identification is most challenging; 46% of extant invertebrates lacked reference barcodes with rates especially high among rotifers, oligochaetes, and mites. Lack of species-level identification for many aquatic invertebrates also is a barrier to matching DNA sequences with physical specimens. Attaining the potential for DNA-based identification of mixed-organism samples covering the breadth of aquatic fauna requires a concerted effort to build supporting barcode libraries and voucher collections.

Human influences on water quality in Great Lakes coastal wetlands.

A better understanding of relationships between human activities and water chemistry is needed to identify and manage sources of anthropogenic stress in Great Lakes coastal wetlands. The objective of the study described in this article was to characterize relationships between water chemistry and multiple classes of human activity (agriculture, population and development, point source pollution, and atmospheric deposition). We also evaluated the influence of geomorphology and biogeographic factors on stressor-water quality relationships. We collected water chemistry data from 98 coastal wetlands distributed along the United States shoreline of the Laurentian Great Lakes and GIS-based stressor data from the associated drainage basin to examine stressor-water quality relationships. The sampling captured broad ranges (1.5-2 orders of magnitude) in total phosphorus (TP), total nitrogen (TN), dissolved inorganic nitrogen (DIN), total suspended solids (TSS), chlorophyll a (Chl a), and chloride; concentrations were strongly correlated with stressor metrics. Hierarchical partitioning and all-subsets regression analyses were used to evaluate the independent influence of different stressor classes on water quality and to identify best predictive models. Results showed that all categories of stress influenced water quality and that the relative influence of different classes of disturbance varied among water quality parameters. Chloride exhibited the strongest relationships with stressors followed in order by TN, Chl a, TP, TSS, and DIN. In general, coarse scale classification of wetlands by morphology (three wetland classes: riverine, protected, open coastal) and biogeography (two ecoprovinces: Eastern Broadleaf Forest [EBF] and Laurentian Mixed Forest [LMF]) did not improve predictive models. This study provides strong evidence of the link between water chemistry and human stress in Great Lakes coastal wetlands and can be used to inform management efforts to improve water quality in Great Lakes coastal ecosystems.

Sensitivity of indices of biotic integrity to simulated fish assemblage changes.

Multimetric indices of biotic integrity (IBIs) are commonly used to assess condition of stream fish assemblages, but their ability to monitor trends within streams over time is largely unknown. We assessed the trend detection ability of two IBI formulations (one with traditional scoring and metrics, and one with nontraditional scoring and region-specific metrics) and of similarity and diversity indices using simulations that progressively altered the fish assemblages of 39 streams in the United States mid-Atlantic Highlands region. We also assessed responses to simulated 50% variability in fish abundances, as a measure of background "noise." Fish assemblage indices responded little to changes that affected all species proportionally despite substantial changes in total fish number. Assemblage indices responded better to scenarios that differentially affected fish species, either according to life history traits or by increasing dominance of already common species, but even these changes took some time to detect relative to background variability levels. Ordinations of stream fish assemblage data suggested that differences among sites were maintained even after substantial alterations of fish composition within sites. IBIs are designed to detect broad assemblage differences among sites while downplaying abundance changes and variability increases that were the first indications of within-site changes, and they appear more suited to detecting large departures from natural fish assemblages than for monitoring gradual changes such as those our simulations produced. Inferences about causes of assemblage changes should be made with caution because of correlations among species traits and interdependence among IBI component metrics. Site trend assessments should be made based on all available data rather than just by summary indices.