Next Article in Journal
Vegetation Structure and Invertebrate Food Availability for Birds in Intensively Used Arable Fields: Evaluation of Three Widespread Crops
Next Article in Special Issue
Salinity as a Key Factor in Structuring Macrophyte Assemblages in Transitional Water Bodies: The Case of the Apulian Coastal Lagoons (Southern Italy)
Previous Article in Journal
Patterns of Richness and Endemism in the Gypsicolous Flora of Mexico
Previous Article in Special Issue
The Foraging Window for Greater White-Fronted Geese (Anser albifrons) Is Consistent with the Growth Stage of Carex
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Diversity
Volume 15
Issue 4
10.3390/d15040523
diversity-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

22 Years of Aquatic Plant Spatiotemporal Dynamics in the Upper Mississippi River

by
Alicia M. Carhart
1,*,
Jason J. Rohweder
2 and
Danelle M. Larson
2
1
Wisconsin Department of Natural Resources, 2630 Fanta Reed Road, La Crosse, WI 54603, USA
2
U.S. Geological Survey, Upper Midwest Environmental Sciences Center, 2630 Fanta Reed Road, La Crosse, WI 54603, USA
*
Author to whom correspondence should be addressed.
Diversity 2023, 15(4), 523; https://doi.org/10.3390/d15040523
Submission received: 23 January 2023 / Revised: 28 February 2023 / Accepted: 2 March 2023 / Published: 4 April 2023
(This article belongs to the Special Issue Aquatic Plant Diversity, Conservation, and Restoration)
Browse Figures
Versions Notes

Abstract

:
Macrophyte (aquatic plant) recovery has occurred in rivers worldwide, but assemblage patterns and habitat requirements are generally not well understood. We examined patterns of species composition and macrophyte abundance in the Upper Mississippi River (UMR), spanning 22 years of monitoring and a period of vegetation recovery. Non-metric multidimensional scaling (NMDS) ordination revealed a gradient of macrophyte abundance and diversity for 25 species, which were associated with water velocity, depth, wind fetch, and water clarity. Three macrophyte genera of ecological and restoration interest (Zizania aquatica, Vallisneria americana, and Sagittaria spp.) occupied different ecological niches. Trends of NMDS values showed that Z. aquatica first co-occurred in shallow areas with Sagittaria spp. but then expanded into deeper, lotic habitats where V. americana often resided. Curve Fit regression analysis identified large areas of significant increases in the relative abundance of V. americana and percent cover of Z. aquatica in several reaches of the UMR from 1998–2019. Sagittaria spp. were more spatiotemporally dynamic, which may indicate specific habitat requirements and sensitivity to environmental gradients. Our analyses showed that these three ecologically important genera are spatiotemporally dynamic but have somewhat predictable habitat associations, which can guide macrophyte management and restoration in the UMR and other large, floodplain rivers.

1. Introduction

Aquatic vegetation (hereafter ‘macrophytes’) provides critical food, habitat, and refuge for river biota and has a strong influence on water quality in large river floodplains. The composition, abundance, and distribution of macrophytes is affected by complex interactions of physical and biological drivers, although, in many rivers, plants are primarily limited by light availability and water level fluctuations [1,2,3]. Many riverine systems lack sufficient data to assess long-term dynamics and the roles of environmental and spatial gradients that structure macrophyte communities [4].
Ecologists quantify species distributions and the environmental drivers (both biotic and abiotic factors) to understand limitations and restoration potential for species and communities. The conceptual development and quantification of the ‘environmental niche’ of species has progressed over the past century [5,6], and assessments today still aim to measure how environmental conditions and the interplay among species (e.g., niche overlap that induces competition) can affect species distributions and how those distributions change over time [7]. Methods for quantifying and defining the environmental niche typically include species distribution modeling or ordination techniques. Ordination quantifies the species–environment relationships (i.e., the environmental niches) by using maximization criteria to construct synthetic environmental gradients (i.e., axes) from the included environmental variables and placing species in ordination space according to their environmental associations [8]. Quantifying the multidimensional environmental niche space can guide restoration targets by revealing environmental drivers and the relative stability or vulnerability of the species niche [9].
There is growing interest in restoring the floodplain habitat and aquatic plants in regulated rivers, such as the Upper Mississippi River (UMR) [10]. Like many large floodplain rivers, the UMR is highly modified and degraded [11,12,13]. Prior to the construction of the lock and dam system in the 1930s, macrophytes were abundant and diverse [10]. The impoundment of the UMR and subsequent maintenance of a 2.7-m navigation channel changed the distribution of flow and depth throughout the system [12,14], resulting in a change in composition and an increase in the abundance of river vegetation [15]. Macrophytes flourished in these new habitats for decades following the construction. A widespread decline in macrophytes was observed in the UMR during the 1990s [16,17]. However, some areas within reaches (e.g., backwater lakes) maintained aquatic vegetation during this period, while other areas (e.g., impounded areas) experienced a substantial recovery from 2000–2016 [17,18,19]. Understanding where macrophytes were stable or underwent changes over this time period and identifying the environmental drivers would provide useful information about resilience of large river floodplains.
A considerable portion of the increase in the UMR macrophytes over this period is attributable to four native species: wild celery (Vallisneria americana Michx.), wild rice (Zizania aquatica L.), and arrowheads (Sagittaria latifolia Willd. and Sagittaria rigida Pursh.) [20]. Like other submersed species, V. americana provides food and shelter for a diverse assemblage of invertebrates, fish, amphibians, reptiles, and water birds [21,22,23]. Vallisneria americana is a predominantly clonal, perennial plant that produces overwintering buds with high energetic value for canvasback ducks (Aythya valisineria Wilson) and other migrating waterfowl [23]. Vallisneria americana is commonly found growing in open impounded areas with moderate flow [3]. Zizania aquatica, an annual aquatic grass, provides excellent cover for many species of water birds, including the federally listed black tern (Chlidonias niger), and provides food for waterfowl during the fall migration [24]. Zizania aquatica is also a valuable food source for humans, provides cultural connections to the river, and is found in many habitat types of this riverscape [20]. Broadleaf arrowhead (S. latifolia) and stiff arrowhead (S. rigida) are emergent species that provide critical wildlife food and habitat [15,25]. Sagittaria spp. are predominantly clonal, perennial plants that expand through the growth of rhizomes, and the edible tubers are commonly found in shallow, backwater habitats [25,26]. These three ecologically important genera appear to be different vegetation types [27] with differing traits and habitat preferences, yet niche differentiation and the annual to decadal spatiotemporal changes have not been well described.
In this study, we assessed the changes to three key genera within the aquatic vegetation assemblage at several spatiotemporal scales across the UMR. We focused on the dynamics among Z. aquatica, V. americana, and Sagittaria spp. due the recent expansion of Z. aquatica in some areas and a concern from river managers that Z. aquatica may be outcompeting and replacing other species. Our specific objectives were to: (1) determine if Sagittaria spp., Z. aquatica, and V. americana occupy different ecological niches or overlap, (2) identify the associated environmental gradients for each species, and (3) quantify how these species have changed in abundance and niche space over 22 years. We expected significant changes in abundance of Sagittaria spp., Z. aquatica, and V. americana to have occurred over the 22-year recovery period and that these changes would be associated with geomorphic characteristics, water quality, and hydrology. We hypothesized that the three genera would separate in ordination space, indicating distinct ecological niches. We also hypothesized that nine environmental factors (such as depth, clarity, velocity, and wind fetch) would separate these species into different and somewhat predictable niches across the riverscape. Based on our field experiences, we suspected that large river sections transitioned from the dominance of either V. americana or Sagittaria spp. into predominantly Z. aquatica, and that Z. aquatica had shifted its ecological niche to deeper, flowing water.

2. Methods

We used multiple, long-term monitoring datasets of macrophytes, hydrology, and water quality metrics to analyze macrophyte assemblage dynamics and habitat associations across space and time in the UMR. We first used non-metric multidimensional scaling (NMDS) to describe species compositional gradients (n = 13,802 sites) and to observe the niche space of the three primary genera. We then used the Envfit function in R (vegan package, version 2.5-7) to evaluate if species composition was associated with hypothesized environmental factors. Lastly, we mapped changes in relative species abundance for V. americana, Z. aquatica, and Sagittaria spp. from 1998–2019 (at annual and decadal time steps) to identify areas that have shifted or remained stable for macrophyte abundance or composition.

2.1. Study Area

The UMR has 29 locks and dams that separate the river into navigation pools. Each pool contains a mosaic of aquatic habitats including a main channel, channel border, side channels, contiguous and isolated floodplain lakes, and impounded areas upstream from the dams. Each of these riverine habitats span large ecological gradients and have varying degrees of flow velocity, water exchange rates, and substrate type [14].
Our study used macrophyte and habitat data from the Long Term Resource Monitoring (LTRM) element of the U.S. Army Corps of Engineers Upper Mississippi River Restoration Program. The entire study area spans approximately 400 km of the UMR, from Lock and Dam 3 near Red Wing, Minnesota to Lock and Dam 13 near Fulton, Illinois (Figure 1). We focused on Pools 4, 8, and 13, because LTRM has monitored aquatic vegetation and water quality using standardized protocols since 1998 in these pools. Pool 4 (~70 km long and ~15,557 ha aquatic area) is sometimes analyzed as “upper Pool 4” and “lower Pool 4” because Lake Pepin, a natural tributary delta lake (~34 km long) between upper- and lower- Pool 4, acts as a major sediment sink and greatly affects water clarity downstream from the lake in Lower Pool 4 [3]. In Pool 8 (~37 km long and ~9424 ha aquatic area) and Pool 13 (~45 km long and ~13,526 ha aquatic area), the hydrologic effects of the navigation system are most pronounced in open impounded areas upstream from the lock and dams (i.e., mean annual water level fluctuation is generally <0.5 m in these areas). Much of the open impounded areas are shallow with high wind fetch and, therefore, are subject to island erosion and sediment resuspension.

2.2. Macrophyte and Habitat Data

Macrophyte data were collected in Pools 4, 8, and 13 using consistent, standardized sampling protocols over the 22-year record (1998–2019) [28]. A minimum of 450 sites in each pool were sampled annually between 15 June and 15 August using a stratified random sampling design, where strata included main channel borders, side channels, contiguous and isolated floodplain lakes, and impounded areas. In each pool, a 50 × 50 m grid was generated and overlaid onto the stratified aquatic areas, and sites were reselected annually using a random number generator. Five aquatic vegetation lifeforms were measured at each site: submersed, emersed, rooted floating-leaved, free-floating (e.g., duckweeds), and filamentous algae. Submersed species and filamentous algae were sampled at six subsites (0.3 m × 1.5 m) around the boat using a double-sided rake. Relative abundance was estimated for each submersed species and algae present on the rake with scores ranging from 0–5 (0 = no plants retrieved, and 1 = 1–20% rake teeth filled to 5 = 81–100% rake teeth filled). Emersed, rooted floating-leaved, and free-floating lifeforms were recorded as percent cover ranging from 0–5 (0 = no plants retrieved, 1–5 in 20% increments) within a 2 m buffer around the boat (total area = ~44 m2).
We standardized all macrophyte lifeforms to the same scale using similar methods published by Bouska [9]. Specifically, for submersed species and filamentous algae, we first summed the rake score of each species across subsites and divided by the maximum possible rake score (30) at each site, and then multiplied by 100 to convert the result into a percentage (0–100%). For emersed, rooted floating-leaved and free-floating lifeforms, we converted % cover to the category’s maximum % cover (20% increments from 0–100%).
We held an “aquatic plant vulnerability workshop” in 2021 to gather professional knowledge from scientists and resource managers regarding the environmental predictors of macrophytes in the UMR. Using workshop knowledge and a priori hypotheses, we selected nine environmental predictors expected to be important for influencing macrophytes (Table S1). Selection criteria also included data availability and low multicollinearity (Pearson’s r < 0.35). Average water depth and substrate type were recorded in the field at each vegetation site. Substrate type was classified as silt/clay, mostly silt with sand, mostly sand with silt, hard clay, gravel/rock, or sand. We recoded substrate to represent a fine- to coarse-scale and velocity gradient (hard clay < silt/clay < sand < gravel/rock). Additional long-term water quality data were collected in these study pools using standardized protocols [29]; however, the location and time of measurement often deviated from that of the vegetation sites. We selected five water quality variables that we expected to be important for structuring plant communities, namely flow velocity (m/s), total nitrogen (mg/L), total phosphorus (mg/L), total suspended solids (mg/L), and chlorophyll a (µg/L). We used ArcGIS (version 10.8.1) to extract water quality values recorded within 60 days and 100 m of each vegetation sample and used the cost distance function to ensure that the nearest water quality site was hydrologically connected to the vegetation site. The selected water quality variables were then merged with the vegetation site data. Wind fetch values (m), modeled for the years 2000 and 2010 [30], were extracted from rasters (10-m × 10-m cell size) and merged with the vegetation site data. Wind fetch values from the year 2000 were applied to vegetation sites sampled in 1998–2009, and wind fetch values from the year 2010 were applied to vegetation sites sampled in 2010–2019. Lastly, we summarized mean discharge during the growing season (m3/s; May–September) at the pool-scale for each year using public data from the U.S. Army Corps of Engineers [31] at Lock and Dams 3, 7, and 12.

2.3. Data Analyses

2.3.1. Using Ordination to Define Niches and Environmental Gradients

We used community ordination and boxplots to address Objectives 1 and 2 for defining the ecological niches and gradients of the three genera. There were 13,802 macrophyte sites across the period of 1998–2019. The final dataset used in the NMDS ordination included 25 species that were detected at >1% of sites (Table 1). This species list included 15 submersed species, 6 emersed species, 2 rooted floating-leaved species, duckweeds (Lemna, Spirodela, and Wolfia spp. combined), and filamentous algae (not identified further). Many of the species that were considered rare were not truly aquatic, so exclusion did not affect the results for the aquatic community analysis. To further improve computation, we removed unvegetated sites and included a dummy variable for each remaining site. We applied a Bray–Curtis dissimilarity measure [32] to the relative abundance dataset, which uses a ranking system to quantify the dissimilarity of species composition among sampling sites. We then ordinated the relationships between sites using NMDS. Sampling sites that were similar to each other with respect to species composition were ordinated closer together on the NMDS biplot. The final NMDS solution had the lowest stress after 20 random starts and 400 iterations. The appropriate number of dimensions for the NMDS was assessed using a scree plot. We then fitted the nine environmental variables to the NMDS ordination and assessed the significance of each factor using the EnvFit function in R [33]. All calculations, including the NMDS ordination, were performed in R version 4.0.3 [33] with the vegan package (version 2.5-7) [34]. Next, we created boxplots for the nine environmental variables to further determine habitat preferences and variation among the three genera.

2.3.2. Mapping Spatial and Temporal Changes in Sagittaria spp., Z. aquatica, and V. americana

To address Objective 3, the spatiotemporal dynamics, we examined NMDS scores through time using kernel density estimation and mapped changes in relative abundance and percent cover across the three pools using curve fit regression. We hypothesized that Z. aquatica had changed its environmental niche space over time, so we plotted annual average NMDS scores for all sites that had ≥20% Z. aquatica coverage at any time within the 22 years. Plotting Z. aquatica sites with ≥20% coverage at any time point allows detection of temporal change in the abundance and colonization of new sites not previously occupied by Z. aquatica.
We interpolated the percent cover of Sagittaria spp. and Z. aquatica, as well as the relative abundance of V. americana for each sampling year (1998–2019) and each pool. Interpolation used the inverse distance weighting (IDW) tool in ArcMap 10.8.1 [35]. The IDW predicts values for cells (10-m × 10-m cell size) in a raster using the average of the eight nearest sampling points within a 2000 m maximum distance search radius. The river’s main channel was used as a barrier in the interpolations. This resulted in annual maps for each species for 21 years (1998–2019, excluding 2003 due to limited sampling) in Pools 4, 8, and 13 (n = 168 annual maps; Figures S1–S8). We used raster calculator in ArcMap to aggregate Sagittaria spp. maps for each year (e.g., the sum of S. rigida and S. latifolia, maximum cell value of 100). The strata that were modeled in the interpolations include channel borders, contiguous floodplain lakes, side channels, and tertiary channels. We excluded the main navigation channel, isolated floodplain lakes, tributaries, and tributary delta lake strata (i.e., Lake Pepin) due to the small number of sampling points. We also excluded all data in Upper Pool 4 (i.e., above Lake Pepin) due to few observations of the three genera of interest.
We used Curve Fit (Version 10.1) [36], an extension developed for use within ArcMap, to run regression analyses on the interpolated community raster datasets from 1998–2019. The Curve Fit output showed areas of significant change in macrophyte abundance over the 22-year period for the three genera of interest. A similar method was used in Pool 8 for submersed vegetation [18]; we expand on this with the addition of Pool 4 and Pool 13, the inclusion of emergents, and a longer temporal scale due to additional sampling years. Each explanatory variable (i.e., year) was paired with the corresponding raster (one raster for the sum of S. rigida and S. latifolia, one raster for Z. aquatica, one raster for V. americana). Curve Fit estimated the linear relationship between year and pixel value (one relative abundance or percent cover value per year) and created an output raster containing parameter estimates, model error, and r2. The r2 was converted to a correlation coefficient (r). The y-intercepts (starting conditions) were then compared to the correlation, or degree of change across the years from 1998–2019. We used a two-tailed t test on the Pearson’s correlation coefficient (t(19) = 0.433, p = 0.05) to define pixels with significant decreases or increases in the relative abundance of V. americana and percent cover of Z. aquatica and Sagittaria spp. The critical value was used to scale and color code the maps to define areas that underwent significant change. A significant positive correlation indicated an increase in relative abundance over time, and a significant negative correlation indicated a decrease in relative abundance over time. Each of these scenarios indicated areas that underwent temporary or perhaps long-term shifts in composition and abundance. Non-significant correlations identified areas that did not undergo significant changes in relative abundance or percent cover over time, and, therefore, may indicate stability or the absence of that species. Standard error estimates were mapped for each Curve Fit analysis to estimate uncertainty and further show relative stability or variability at various spatiotemporal scales.
We repeated the Curve Fit analyses to examine changes between the years 2010 and 2019 when Z. aquatica expanded substantially in Pools 4 and 8. We used a two-tailed t test on the Pearson’s correlation coefficient (t(8) = 0.632, p = 0.05) to define pixels with significant decreases or increases in the percent cover of Z. aquatica and Sagittaria spp. and relative abundance of V. americana. Examining change across several temporal scales allowed us to determine whether species were highly dynamic, possibly in response to environmental drivers, or more stable and predictable from year to year.

3. Results

3.1. Defining Ecological Niches and Associated Environmental Gradients

The NMDS of the entire macrophyte assemblage across all three pools (2D Stress = 0.20) showed that the three genera of interest were separated distinctly in ordination space across two main axes (Figure 2, Table S2). The first NMDS axis (NMDS 1) identified a strong gradient of plant abundance and diversity. The second NMDS axis (NMDS 2) identified a gradient of lotic and lentic macrophytes associated with depth, velocity, and wind fetch (Table S3); specifically, positive NMDS 2 values were characterized by lotic habitats that exhibited deeper depths and greater velocity and wind fetch compared to negative NMDS 2 values. Vallisneria americana was in the upper left quadrant with negative NMDS 1 values and positive NMDS 2 values; Z. aquatica was in the upper right quadrant with positive NMDS 1 and NMDS 2 values; and NMDS 1 values varied by Sagittaria spp. but both S. rigida and S. latifolia had negative NMDS 2 values. The species in the upper portion of the ordination plot with positive NMDS 2 values, such as V. americana and Z. aquatica, were generally found in greater velocities and deeper water depths, whereas other species, such as Sagittaria spp., which were in the lower portion of the ordination plot, were associated with lower velocities, shallower depths, and minimal wind fetch.
Similar associations between environmental variables and macrophytes were presented in boxplots (Figure 3). Sagittaria spp. were generally found in shallow, protected areas with minimal wind fetch, low velocity, and a silty substrate. Sagittaria spp. were also associated with lower discharge during the growing season compared to Z. aquatica and V. americana. Vallisneria americana had the greatest environmental variability (i.e., a large range of environmental conditions at surveyed sites), although this species was commonly found at sites with the deepest water depths (up to 3 m), largest wind fetch (>2 km), highest water velocity, and silt-sand substrate (Figure 3). Zizania aquatica was generally found in silty substrates, areas with high water clarity, and sites with moderate water depths, wind fetch, and water velocity.

3.2. Ordination-Quantifying Patterns of Macrophyte Abundance and Niche Space over Time

Ordination scores at the pool-scale showed changes in species assemblages and the dominance of the three macrophyte genera from 1998–2019 (Figure 4). Zizania aquatica abundance substantially increased in Lower Pool 4 by 2019, as indicated by more NMDS values in the upper right quadrant in later years (Figure 4). Pool 8 became more vegetated and speciose in later years and had especially high V. americana and Z. aquatica abundances. Furthermore, Pool 8 had the greatest abundance of Sagittaria spp. in the mid-2000s and less Sagittaria spp. towards 2019. Pool 13 sites were dominated by either V. americana or Sagittaria spp. in later years. Zizania aquatica was absent in Pool 13 throughout the study period and the ordination space in the upper right quadrant was relatively empty, meaning there may be an unoccupied habitat in Pool 13.
Zizania aquatica expanded its ecological niche over the 22 years, as indicated by changes in annual mean NMDS values at sites with ≥20% coverage of Z. aquatica (Figure 5). Initially, Z. aquatica overlapped spatially with several lentic submersed species and emergent species (e.g., stiff arrowhead) and was present at high diversity sites (+ NMDS 1, − NMDS 2). Over time, mean centroid values decreased for NMDS 1 and increased for NMDS 2, showing that Z. aquatica initially occupied high-diversity, shallow backwater habitats in 1998 but subsequently expanded to deeper, lotic areas by 2019.

3.3. Curve Fit- Spatial and Temporal Trends of Z. aquatica, V. americana, and Sagittaria spp.

Curve Fit analysis identified large areas with significant increases in percent cover of Z. aquatica in Lower Pool 4 and Pool 8 from 1998–2019. Specifically, there was a significant increase in the percent cover of Z. aquatica at 1128 hectares in Lower Pool 4 (Figure 6E) and 3657 hectares in Pool 8 (Figure 7E), and no areas had significant decreases from 1998–2019 (Table S4). Zizania aquatica prevalence also substantially increased in Pools 4 and 8, particularly in the impounded areas upstream of locks and dams (Figure S9).
Sagittaria spp. increased in percent cover from 1998–2019, but the total area was much smaller than that of Z. aquatica (Table S4); specifically, Sagittaria spp. coverage increased by ~46 hectares in Lower Pool 4 (Figure 6D), ~109 hectares in Pool 8 (Figure 7D), and ~130 hectares in Pool 13 (Figure 8C). Small areas in all study pools decreased in percent cover from 1998–2019 (Table S4). The prevalence of Sagittaria spp. in the contiguous backwaters of Pools 4 and 8 declined substantially starting in 2010, nearly simultaneously with the expansion of Z. aquatica (Figure S9). Over time, the dominant emergent species in Pools 4 and 8 shifted from Sagittaria spp. in 1998 to Z. aquatica in 2019; however, both S. rigida and S. latifolia remained the dominant species in Pool 13. Curve Fit results for 2010–2019 (Figures S10–S12) did not identify large areas of significant change in any pool. Furthermore, standard error estimates (Figure S13) were greater for Sagittaria spp. compared to Z. aquatica and V. americana, especially in Lower Pool 4 and Pool 8, showing that Sagittaria spp. were more variable at an annual-scale.
Similar to Z. aquatica, Curve Fit analysis identified large areas with significant increases in V. americana across all study pools, especially Lower Pool 4 and the impounded area of Pool 8 from 1998–2019. There was a significant increase in the relative abundance of V. americana of 539 hectares in Lower Pool 4 (Figure 6F), 2514 hectares in Pool 8 (Figure 7F), and 526 hectares in Pool 13 (Figure 8D). There were small, localized areas of decline in V. americana from 1998–2019 scattered throughout Lower Pool 4 and the impounded area of Pool 13 ((Figure 6F and Figure 8D). The prevalence of V. americana increased over time in all pools; however, V. americana declined in Pool 13 from 2014–2019 (Figure S9). Curve Fit analysis for 2010–2019 revealed more areas that experienced a significant decline in the abundance of V. americana in Pool 13′s impounded area (434 hectares; Figure S12).
In summary, in terms of the interplay among the three macrophyte genera, Curve Fit maps identified specific locations within each pool that experienced significant changes in V. americana, Z. aquatica, and Sagittaria spp. (Figure 6, Figure 7 and Figure 8). Zizania aquatica increased in Lower Pool 4 and Pool 8, and often overlapped spatially with Sagittaria spp. and V. americana; however, V. americana and Sagittaria spp. rarely grew in the same locations. In Pools 4 and 8, increases in Z. aquatica and V. americana often co-occurred throughout the pool, but V. americana often dominated large areas in the impounded area. In Pool 13, one small area showed a decrease in V. americana and increase in Sagittaria spp.; however, V. americana mostly increased in the absence of Sagittaria spp. and Z. aquatica (Figure 8C,D). Similarly, the substantial decline in V. americana in Pool 13 from 2010–2019 did not overlap with the spatial distribution of Sagittaria spp. or Z. aquatica (Figure S12).

4. Discussion

Our results confirmed that V. americana, Z. aquatica, and Sagittaria spp. occupied different ecological niches in the UMR. Sagittaria spp. were generally found in shallow, protected backwaters, co-occurring with other lentic aquatic macrophytes, whereas the lotic habitat tends to be dominated by V. americana. Notably, Z. aquatica separated from all other emergent species in the NMDS ordination. Zizania aquatica was found at high densities in shallow backwaters of the UMR, although trends in NMDS values showed that Z. aquatica later expanded into deeper, flowing water and suggests that it may be more tolerant of lotic conditions than previously understood [37,38,39]. Vallisneria americana generally increased across the study area, but Pool 13 showed a decline in V. americana since 2014. These results can aid restoration and management of the UMR and other large, regulated rivers by providing an understanding of the spatiotemporal dynamics of common species and the likely environmental drivers of species changes and stability.
We found that water depth, water velocity, water clarity, and wind fetch were important drivers for explaining differences in species composition and abundance of the entire macrophyte assemblage and these three genera of interest. This finding was consistent with environmental drivers governing macrophyte growth in other large rivers [2,40]. Many interactions exist between macrophytes, water quality, and the physical environment [41]. For example, wind fetch and water velocity can have a substantial effect on macrophyte growth due to physical stress caused by wind-generated waves and high current velocity. Furthermore, increased erosion and sediment re-suspension reduce water clarity and light availability, which can affect submersed plants, such as V. americana [3]. However, once plants become established, depth and flow may be altered enough to support increased coverage and density of macrophytes which, in turn, can feedback to reduce the effects of wind-generated waves and sediment re-suspension [41]. Our field experiences lead us to suggest that feedback among Z. aquatica, sediment deposition, water depth, and water clarity occurred in the UMR and possibly facilitated Z. aquatica expansion, although more experimental research would be needed to confirm this.
Our hypothesis that Z. aquatica has expanded into deeper water and a new niche space was supported (Figure 4, Figure 5, Figure 6 and Figure 7). The expansion of Z. aquatica into deeper water may be a result of the substantial improvement in water clarity that occurred throughout the UMR over the last two decades [42,43], which increased light penetration and may have allowed seedlings to emerge from deeper water [44]. This is supported by the strong negative association with suspended solids (SS) (Figure 2) and low SS measured near Z. aquatica sites (Figure 3F). In addition, dense stands of Z. aquatica may be effectively reducing water velocity [39], capturing sediment [38], and aiding in natural island formation (A. Carhart, pers. observations, 2022), although the stability of these habitats is unknown. Asaeda [39] found that Zizania adapted to greater water velocity by increasing shoot density within patches, thereby reducing overall drag force and altering the flow pattern. Although Z. aquatica was not strongly correlated with mean discharge during the growing season, annual plants often benefit from occasional disturbances and water level fluctuations that set back perennial competition and create habitat for re-colonization during low water [38].
Curve Fit results confirmed significant changes in the total area and relative abundance of the three macrophyte genera (Figure 6, Figure 7 and Figure 8), but there was little support for the hypothesis that Z. aquatica had overlapped its niche space and displaced V. americana and Sagittaria spp. (Figure 2, Figure 3, Figure 4, Figure 6, Figure 7 and Figure 8). Zizania aquatica expanded rapidly in Lower Pool 4 and Pool 8 circa 2010 and became the most prevalent emergent species in these reaches [20]. However, as of 2019, Z. aquatica had not been detected in Pool 13 during LTRM surveys. Sagittaria spp. remains the dominant emergent species in Pool 13, although the recent invasion of flowering rush (Butomus umbellatus L.) may compete for Sagittaria spp.’s niche space. Furthermore, trends in percent cover of Sagittaria spp. were variable across all pools, which may indicate sensitivity to environmental gradients, such as the lotic and lentic hydrologic gradient in Figure 2. The relative abundance of V. americana increased substantially in all study pools over the 22-year study; however, large areas within the Pool 13 impounded area experienced a reduction in the abundance of V. americana since 2014. This indicates that restoration and management inventions for V. americana in Pool 13 may be helpful to reduce loss and increase resilience.
To prevent vulnerability to macrophyte loss and increase ecosystem resilience, river managers must understand the environmental drivers and community dynamics of the past two decades spanning a period of widespread vegetation recovery. Our analyses showed that Z. aquatica, V. americana, and Sagittaria spp. are spatiotemporally dynamic in the UMR, but have some predictable habitat associations that can help guide future management and restoration efforts. Given the ecological importance of these three macrophyte genera as defined vegetation types [27] and specified as restoration targets in the UMR [13,45,46], modeling the hypothesized ecological drivers and dynamics of each species individually would further aid ecological understanding and management. Although we focused on patterns in relative abundance and defining ecological niches in the UMR, Zizania, Vallisneria, and Sagittaria are common genera in rivers worldwide [39,47,48,49], and, therefore, the methods and results provided here may be transferrable to other riverine systems where data may be sparse.

5. Conclusions

The patterns documented here illustrated the importance of analyzing species compositional changes at multiple spatial and temporal scales. As one example from Pool 13, a few large V. americana patches increased in total area and abundance since 1998 (Table S4, Figure 8D), yet many small patches had significant decreases since 2014 (Figure S8 and Figure S12) which may signal vulnerability and V. americana bed fragmentation. We found that the extent and abundance of Z. aquatica and V. americana patches significantly increased over time and that Sagittaria spp. were more dynamic and experienced very little long-term change. Further research can quantify threshold responses of the determined environmental drivers (Figure 2, Table S3) and associated feedback mechanisms that affected change and stability for each species individually. Our findings contribute to macrophyte restoration and management efforts in the UMR and other large, floodplain rivers worldwide.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/d15040523/s1, Table S1: A summary of the 9 environmental variables included in the EnvFit model that were important in explaining site-level differences in species composition and relative abundance of aquatic vegetation in the Upper Mississippi River (UMR), USA These predictors were selected following our “aquatic plant vulnerability workshop” in 2021 that gathered professional knowledge from 65 scientists and resource managers regarding the environmental predictors of aquatic vegetation in the UMR. Selection criteria also included data availability and low multicollinearity (Pearson’s r < 0.35); Table S2: Species centroid values (n = 25) for the first 2 non-metric multidimensional scaling (NMDS) axes across Pools 4, 8, and 13 of the Upper Mississippi River, USA; Table S3: Summary of environmental variable (n = 9) centroid values for the first two non-metric multidimensional scaling (NMDS) axes across Pools 4, 8, and 13 of the Upper Mississippi River, USA; Table S4: Area (hectares) of significant change in percent cover of arrowheads (Sagittaria rigida and S. latifolia) and wild rice (Zizania americana) and relative abundance of wild celery (Vallisneria americana) in Pools 4, 8, and 13 of the Upper Mississippi River, USA, from 1998 to 2019; Figure S1: Spatial changes in percent cover of wild rice (Zizania aquatica) in Lower Pool 4, Upper Mississippi River, from 1998 to 2019. Areas not sampled are shown in gray; Figure S2: Spatial changes in percent cover of arrowheads (Sagittaria spp.) in Lower Pool 4, Upper Mississippi River, from 1998 to 2019. Areas not sampled are shown in gray; Figure S3: Spatial changes in relative abundance of wild celery (Vallisneria americana) in Lower Pool 4, Upper Mississippi River, from 1998 to 2019. Areas not sampled are shown in gray; Figure S4: Spatial changes in percent cover of wild rice (Zizania aquatica) in Pool 8, Upper Mississippi River, from 1998 to 2019. Areas not sampled are shown in gray; Figure S5: Spatial changes in percent cover of arrowheads (Sagittaria spp.) in Pool 8, Upper Mississippi River, from 1998 to 2019. Areas not sampled are shown in gray; Figure S6: Spatial changes in relative abundance of wild celery (Vallisneria americana) in Pool 8, Upper Mississippi River, from 1998 to 2019. Areas not sampled are shown in gray; Figure S7: Spatial changes in percent cover of arrowheads (Sagittaria spp.) in Pool 13, Upper Mississippi River, from 1998 to 2019. Areas not sampled are shown in gray; Figure S8: Spatial changes in relative abundance of wild celery (Vallisneria americana) in Pool 13, Upper Mississippi River, from 1998 to 2019. Areas not sampled are shown in gray; Figure S9: Annual prevalence (±1 SE) of (A) wild rice (Zizania aquatica), (B) arrowheads (Sagittaria spp.), and (C) wild celery (Vallisneria americana) in Pool 4, Pool 8, and Pool 13 of the Upper Mississippi River, USA Locally weighted regression (LOESS model) trendlines (solids lines) and 95% confidence intervals (shaded areas) depict temporal patterns from 1998 to 2019. Panel A modified from Larson et al. (2022); Figure S10: Curve Fit regression output for arrowheads (Sagittaria latifolia and S. rigida, combined), wild rice (Zizania aquatica), and wild celery (Vallisneria americana) in Lower Pool 4, Upper Mississippi River, USA Y-intercept indicates the estimated percent cover in 2010 (A–C), correlation indicates the slope of percent cover from 2010–2019 (D–F), and (G–I) indicates the estimated percent cover in 2019. Correlation values less than −0.632 are significant negative slopes (blue) and correlation values greater than 0.632 are significant positive slopes (red). Intermediate values indicate no significant change. Areas that are not sampled are shown in gray.; Figure S11: Curve Fit regression output for arrowheads (Sagittaria latifolia and S. rigida, combined), wild rice (Zizania aquatica), and wild celery (Vallisneria americana) in Pool 8, Upper Mississippi River, USA Y-intercept indicates the estimated percent cover in 2010 (A–C), correlation indicates the slope of percent cover from 2010 to 2019 (D–F), and (G–I) indicates the estimated percent cover in 2019. Correlation values less than −0.632 are significant negative slopes (blue) and correlation values greater than 0.632 are significant positive slopes (red). Intermediate values indicate no significant change. Areas that are not sampled are shown in gray; Figure S12: Curve Fit regression output for arrowheads (Sagittaria latifolia and S. rigida, combined) and wild celery (Vallisneria americana) in Pool 13, Upper Mississippi River, USA Y-intercept indicates the estimated percent cover in 2010 (A,B), correlation indicates the slope of percent cover from 2010 to 2019 (C,D), and (E,F) indicates the estimated percent cover in 2019. Correlation values less than −0.632 are significant negative slopes (blue) and correlation values greater than 0.632 are significant positive slopes (red). Intermediate values indicate no significant change. Areas that are not sampled are shown in gray; Figure S13: Curve fit standard error estimates for arrowheads (Sagittaria latifolia and S. rigida, combined), wild celery (Vallisneria americana), and wild rice (Zizania aquatica) in Lower Pool 4 (A–C), Pool 8 (D–F), and Pool 13 (G–H) of the Upper Mississippi River, USA Areas that are not sampled are shown in gray. References [50,51] are cited in the supplementary materials.

Author Contributions

Conceptualization, A.M.C., J.J.R. and D.M.L.; methodology, A.M.C., J.J.R. and D.M.L.; hosted workshops, D.M.L.; formal analysis, A.M.C., J.J.R. and D.M.L.; data curation, J.J.R.; writing—original draft preparation, A.M.C.; writing—review and editing, A.M.C., D.M.L. and J.J.R.; visualization, A.M.C., D.M.L. and J.J.R.; funding acquisition, A.M.C., J.J.R. and D.M.L. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the U.S. Army Corps of Engineers Upper Mississippi River Restoration Program. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

Institutional Review Board Statement

Not applicable.

Data Availability Statement

The aquatic vegetation and water quality datasets are publicly available through the Upper Mississippi River Restoration Program website (https://www.umesc.usgs.gov/ltrm-home.html). Daily discharge is publicly provided by the U.S. Army Corps of Engineers for the Mississippi River at Red Wing, Minnesota (Lock and Dam 3), La Crescent, Minnesota (Lock and Dam 7), and Bellevue, Iowa (Lock and Dam 12) and obtained from https://rivergages.mvr.usace.army.mil/WaterControl/new/layout.cfm. The wind fetch data were publicly accessed from https://doi.org/10.5066/P99OVGZB. All spatial datasets (NMDS rasters, Curve Fit outputs) are publicly available at https://doi.org/10.5066/P96A7PC5 [52].

Acknowledgments

This study was funded as part of the U.S. Army Corps of Engineers Upper Mississippi River Restoration (UMRR) Program, Long Term Resource Monitoring (LTRM) element. We first acknowledge that this work was only made possible through the efforts of the many thoughtful people who designed and implemented the LTRM sampling program with rigorous attention to detail and quality assurance, as well as the field and laboratory staff who have collected and produced these data. We are thankful for the included discharge data from the U.S. Army Corps of Engineers from the following gages: Red Wing, Minnesota (Lock and Dam 3), La Crescent, Minnesota (Lock and Dam 7), and Bellevue, Iowa (Lock and Dam 12). We thank the “2021 aquatic vegetation vulnerability workshop” participants for engaging in adaptive management and inspiration. The authors thank Kristen Bouska, U.S. Geological Survey, for initial discussions and R-code pertaining to ordination plots, Eric Lund and Nicole Ward, MN DNR, for initial discussions, and Tim Fox, U.S. Geological Survey, for assisting with Curve Fit analyses. We also thank Jeff Houser and Nathan De Jager, U.S. Geological Survey, Benjamin Patschull and Stephanie Szura, WI DNR, and two other anonymous reviewers for providing helpful reviews.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Van Geest, G.J.; Wolters, H.; Roozen, F.C.J.M.; Coops, H.; Roijackers, R.M.M.; Buijse, A.D.; Scheffer, M. Water-Level Fluctuations Affect Macrophyte Richness in Floodplain Lakes. Hydrobiologia 2005, 539, 239–248. [Google Scholar] [CrossRef]
  2. Schneider, B.; Cunha, E.R.; Marchese, M.; Thomaz, S.M. Explanatory Variables Associated with Diversity and Composition of Aquatic Macrophytes in a Large Subtropical River Floodplain. Aquat. Bot. 2015, 121, 67–75. [Google Scholar] [CrossRef]
  3. Carhart, A.M.; Kalas, J.E.; Rogala, J.T.; Rohweder, J.J.; Drake, D.C.; Houser, J.N. Understanding Constraints on Submersed Vegetation Distribution in a Large, Floodplain River: The Role of Water Level Fluctuations, Water Clarity and River Geomorphology. Wetlands 2021, 41, 57. [Google Scholar] [CrossRef]
  4. Alahuhta, J.; Lindholm, M.; Baastrup-Spohr, L.; García-Girón, J.; Toivanen, M.; Heino, J.; Murphy, K. Macroecology of Macrophytes in the Freshwater Realm: Patterns, Mechanisms and Implications. Aquat. Bot. 2021, 168, 103325. [Google Scholar] [CrossRef]
  5. Grinnell, J. The niche-relationships of the California thrasher. The Auk 1917, 34, 427–433. [Google Scholar] [CrossRef]
  6. Chase, J.M.; Leibold, M.A. Ecological Niche: Linking Classical and Contemporary Approaches; The University of Chicago Press: Chicago, IL, USA, 2003. [Google Scholar]
  7. Devictor, V.; Clavel, J.; Julliard, R.; Lavergne, S.; Mouillot, D.; Thuiller, W.; Venail, P.; Villéger, S.; Mouquet, N. Defining and measuring ecological specialization. J. Appl. Ecol. 2010, 47, 15–25. [Google Scholar] [CrossRef] [Green Version]
  8. Jongman, R.H.G.; ter Braak, C.J.F.; Van Tongeren, O.F.R. Data Analysis in Community and Landscape Ecology; Cambridge University Press: Cambridge, UK, 1995. [Google Scholar]
  9. Bouska, K.L.; Larson, D.M.; Drake, D.C.; Lund, E.M.; Carhart, A.M.; Bales, K.R. Aquatic Vegetation Dynamics in the Upper Mississippi River over 2 Decades Spanning Vegetation Recovery. Freshw. Sci. 2022, 41, 33–44. [Google Scholar] [CrossRef]
  10. USACE (U.S. Army Corps of Engineers). Upper Mississippi River System Ecosystem Restoration Objectives—2009. U.S. Army Corps of Engineers, Rock Island District, Rock Island, Illinois. 2011. Available online: https://www.mvr.usace.army.mil/Portals/48/docs/Environmental/UMRR/UMRR_Ecosystem_Restoration_Objectives_2009.pdf (accessed on 1 October 2022).
  11. Fremling, C.R. Immortal River—The Mississippi in Ancient and Modern Times; University of Wisconsin Press: Madison, WI, USA, 2005; p. 472. [Google Scholar]
  12. Houser, J.N.; Bouska, K.L.; De Jager, N.R.; Ickes, B.S.; Jankowski, K.J.; Larson, D.M.; Van Appledorn, M.; Rohweder, J.J. Introduction, Chap. In Ecological Status and Trends of the Upper Mississippi and Illinois Rivers; U.S. Geological Survey Open-File Report 2022-1039; Houser, J.N., Ed.; Upper Midwest Environmental Sciences Center: La Crosse, WI, USA, 2022; Volume 1, pp. 162–177. [Google Scholar] [CrossRef]
  13. McCain, K.N.S.; Schmuecker, S.; De Jager, N.R. Habitat Needs Assessment-II for the Upper Mississippi River Restoration Program: Linking Science to Management Perspectives; U.S. Army Corps of Engineers: Rock Island, IL, USA, 2018. Available online: https://pubs.er.usgs.gov/publication/70202203 (accessed on 9 January 2023).
  14. De Jager, N.R.; Rogala, J.T.; Rohweder, J.J.; Van Appledorn, M.; Bouska, K.L.; Houser, J.N.; Jankowski, K.J. Indicators of Ecosystem Structure and Function for the Upper Mississippi River System; Open-file report 2018-1143; U.S. Geological Survey, upper Midwest Environmental Sciences Center: La Crosse, WI, USA, 2018. [Google Scholar] [CrossRef]
  15. Peck, J.H.; Smart, M.M. An Assessment of the Aquatic and Wetland Vegetation of the Upper Mississippi River. Hydrobiologia 1986, 136, 57–75. [Google Scholar] [CrossRef]
  16. Fischer, J.R.; Claflin, T.O. Declines in Aquatic Vegetation in Navigation Pool No. 8, Upper Mississippi River between 1975 and 1991. Regul. Rivers Res. Manag. 1995, 11, 157–165. [Google Scholar] [CrossRef]
  17. De Jager, N.R.; Rohweder, J.J. Changes in Aquatic Vegetation and Floodplain Land Cover in the Upper Mississippi and Illinois Rivers (1989–2000–2010). Environ. Monit. Assess. 2017, 189, 77. [Google Scholar] [CrossRef]
  18. Carhart, A.M.; De Jager, N.R. Spatial and Temporal Changes in Species Composition of Submersed Aquatic Vegetation Reveal Effects of River Restoration. Restor. Ecol. 2018, 27, 672–682. [Google Scholar] [CrossRef]
  19. Burdis, R.M.; DeLain, S.A.; Lund, E.M.; Moore, M.J.C.; Popp, W.A. Decadal Trends and Ecological Shifts in Backwater Lakes of a Large Floodplain River: Upper Mississippi River. Aquat. Sci. 2020, 82, 27. [Google Scholar] [CrossRef]
  20. Larson, D.M.; Lund, E.M.; Carhart, A.M.; Drake, D.C.; Houser, J.N.; De Jager, N.R.; Bouska, K.L.; Bales, K.R.; Giblin, S.M. Aquatic Vegetation Indicators. In Ecological Status and Trends of the Upper Mississippi and Illinois Rivers; U.S. Geological Survey Open-File Report, 2022-1039; Houser, J.N., Ed.; USGS: La Crosse, WI, USA, 2022; Volume 1, pp. 112–133. [Google Scholar] [CrossRef]
  21. DeLain, S.A.; Popp, W.A. Relationship of Weed Shiner and Young-of Year Bluegill and Largemouth Bass Abundance to Submersed Aquatic Vegetation in Navigation Pools 4, 8, and 13 of the Upper Mississippi River, 1998–2012; LTRM technical report 2014–T001; U.S. Geological Survey—USGS: La Crosse, WI, USA, 2014. Available online: https://pubs.usgs.gov/mis/ltrmp2014-t001/pdf/ltrmp2014-t001.pdf (accessed on 1 January 2022).
  22. Weeks, A.M.; De Jager, N.R.; Haro, R.J.; Sandland, G.J. Spatial and Temporal Relationships Between the Invasive Snail Bithynia tentaculata and Submersed Aquatic Vegetation in Pool 8 of the Upper Mississippi River: Spatial and Temporal Patterns of Bithynia tentaculata. River Res. Appl. 2017, 33, 729–739. [Google Scholar] [CrossRef] [Green Version]
  23. Korschgen, C.E.; Green, W.L. American Wildcelery (Vallisneria americana): Ecological Considerations for Restoration; U.S. Department of the Interior, Fish and Wildlife Service: Bailey’s Crossroads, VA, USA, 1988; p. 27. [Google Scholar]
  24. U.S. Fish and Wildlife Service. Upper Mississippi River National Wildlife and Fish Refuge Comprehensive Conservation Plan and Habitat Management Plan: U.S. Fish and Wildlife Service. 2019. Available online: https://www.fws.gov/midwest/planning/uppermiss/ (accessed on 22 November 2022).
  25. Kenow, K.P.; Gray, B.R.; Lyon, J.E. Flooding Tolerance of Sagittaria latifolia and Sagittaria rigida under Controlled Laboratory Conditions: Flooding Tolerance of Sagittaria under Laboratory Conditions. River Res. Appl. 2018, 34, 1024–1031. [Google Scholar] [CrossRef]
  26. Marburger, J.E. Biology and Management of Sagittaria latifolia Willd (Broad-Leaf Arrow-Head) for Wetland Restoration and Creation. Restor. Ecol. 1993, 1, 248–255. [Google Scholar] [CrossRef]
  27. Larson, D.M.; Carhart, A.; Lund, E. Aquatic vegetation types identified during early and late phases of vegetation recovery in the upper Mississippi River. Ecosphere 2023, 14, e4468. [Google Scholar] [CrossRef]
  28. Yin, Y.; Winkelman, J.S.; Langrehr, H.A. Long Term Resource Monitoring Program Procedures: Aquatic Vegetation Monitoring; Program Report 95-P002-7; United States Geological Survey, Upper Midwest Environmental Science Center: La Crosse, WI, USA, 2000. Available online: https://umesc.usgs.gov/documents/reports/1995/95p00207.pdf (accessed on 1 January 2023)95p.
  29. Soballe, D.M.; Fischer, J.R. Long Term Resource Monitoring Program Procedures: Water Quality Monitoring; Technical Report 2004-T002-01; United States Geological Survey, Upper Midwest Environmental Science Center: La Crosse, WI, USA, 2004. Available online: https://www.umesc.usgs.gov/documents/reports/2004/04t00201.pdf (accessed on 1 January 2023).
  30. Rohweder, J.; Rogala, J. Upper Mississippi River System Weighted Wind Fetch Analysis (1989, 2000, 2010/2011); Contract report prepared for the U.S. Army Corps of Engineers’ Upper Mississippi River Restoration—Long Term Resource Monitoring element; U.S. Geological Survey: La Crosse, WI, USA, 2020; p. 26. [Google Scholar] [CrossRef]
  31. U.S. Army Corps of Engineers—Water Levels of Rivers and Lakes. Available online: https://rivergages.mvr.usace.army.mil/WaterControl/new/layout.cfm (accessed on 1 March 2021).
  32. Bray, J.R.; Curtis, J.T. An ordination of the upland forest communities of southern Wisconsin. Ecol. Monogr. 1957, 27, 326–349. [Google Scholar] [CrossRef]
  33. R Core Team. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2020; Available online: https://www.R-project.org/ (accessed on 1 January 2023).
  34. Oksanen, J.F.; Blanchet, G.; Friendly, M.; Kindt, R.; Legendre, P.; McGlinn, D.; Minchin, P.R.; O’Hara, R.B.; Simpson, G.L.; Solymos, P.; et al. Vegan: Community Ecology Package. 2019. Available online: https://cran.r-project.org/web/packages/vegan/index.html (accessed on 1 January 2022).
  35. ESRI. ArcGIS Desktop, Version 10.8.1; Environmental Systems Research Institute: Redlands, CA, USA, 2022.
  36. De Jager, N.R.; Fox, T.J. Curve Fit: A Pixel-Level Raster Regression Tool for Mapping Spatial Patterns. Methods Ecol. Evol. 2013, 4, 789–792. [Google Scholar] [CrossRef]
  37. Pillsbury, R.W.; McGuire, M.A. Factors Affecting the Distribution of Wild Rice (Zizania palustris) and the Associated Macrophyte Community. Wetlands 2009, 29, 724–734. [Google Scholar] [CrossRef]
  38. Meeker, J.E. The Ecology of “Wild” Wild-Rice (Zizania palustris var. palustris) in the Kakagon Sloughs, a Riverine Wetland on Lake Superior;  The University of Wisconsin–Madison ProQuest Dissertations Publishing: Madison, WI, USA, 1993. [Google Scholar]
  39. Asaeda, T.; Fujino, T.; Manatunge, J. Morphological Adaptations of Emergent Plants to Water Flow: A Case Study with Typha angustifolia, Zizania latifolia and Phragmites australis. Freshw. Biol. 2005, 50, 1991–2001. [Google Scholar] [CrossRef]
  40. Franklin, P.; Dunbar, M.; Whitehead, P. Flow controls on lowland river macrophytes: A review. Sci. Total Environ. 2008, 400, 369–378. [Google Scholar] [CrossRef] [PubMed]
  41. Madsen, J.D.; Chambers, P.A.; James, W.F.; Koch, E.W.; Westlake, D.F. The interaction between water movement, sediment dynamics and submersed macrophytes. Hydrobiologia 2001, 444, 71–84. [Google Scholar] [CrossRef]
  42. Jankowski, K.J. Water Quality Indicators. In Ecological Status and Trends of the Upper Mississippi and Illinois Rivers; U.S. Geological Survey Open-File Report, 2022-1039; Houser, J.N., Ed.; USGS: La Crosse, WI, USA, 2022; Volume 1, pp. 84–110. [Google Scholar] [CrossRef]
  43. Houser, J.N.; Bales, K.R.; Bouska, K.L.; Burdis, R.M.; Carhart, A.M.; Drake, D.C.; Jankowski, K.J.; Kalas, J.E.; Larson, D.M.; Lund, E.M.; et al. Using Long-Term Data to Understand the Causes and Consequences of Changes in Water Clarity and Aquatic Vegetation in the Upper Impounded Reach of the Upper Mississippi River. In Ecological Status and Trends of the Upper Mississippi and Illinois Rivers; U.S. Geological Survey Open-File Report, 2022-1039; Houser, J.N., Ed.; Upper Midwest Environmental Sciences Center: La Crosse, WI, USA, 2022; Volume 1, pp. 167–177. [Google Scholar] [CrossRef]
  44. Sims, L.; Pastor, J.; Lee, T.; Dewey, B. Nitrogen, Phosphorus and Light Effects on Growth and Allocation of Biomass and Nutrients in Wild Rice. Oecologia 2012, 170, 65–76. [Google Scholar] [CrossRef]
  45. Devendorf, R.D. A Dabbling Duck Migration Model for the Upper Mississippi River: St. Paul, Minn., U.S. Army Corps of Engineers, St. Paul District. 2013. Available online: https://cw-environment.erdc.dren.mil/model-library.cfm?CoP=Restore&Option=View&Id=622 (accessed on 9 January 2023).
  46. Devendorf, R.D. A Migratory Habitat Model for Diving Ducks Using the Upper Mississippi River: St. Paul, Minn., U.S. Army Corps of Engineers, St. Paul District. 2013. Available online: https://cw-environment.erdc.dren.mil/model-library.cfm?CoP=Restore&Option=View&Id=200 (accessed on 9 January 2023).
  47. Wang, H.; Xiao, K.; Wu, Z.; Chen, J.; Xiong, W.; Wang, Z.; Wang, Q.; Zhu, H.; Bowler, P. Delta Arrowhead (Sagittaria platyphylla) in the Yangtze River: An Invasive Aquatic Plant and the Potential Ecological Consequences. BioInvasions Rec 2020, 9, 618–626. [Google Scholar] [CrossRef]
  48. Liao, Y.Y.; Gichira, A.W.; Wang, Q.-F.; Chen, J.-M. Molecular Phylogeography of Four Endemic S. Agittaria Species (Alismataceae) in the Sino-Japanese Floristic Region of East Asia: Phylogeography of East Asian Sagittaria. Bot. J. Linn. Soc. 2016, 180, 6–20. [Google Scholar] [CrossRef] [Green Version]
  49. Jarvis, J.C.; Moore, K.A. Influence of Environmental Factors on Vallisneria americana Seed Germination. Aquat. Bot. 2008, 88, 283–294. [Google Scholar] [CrossRef]
  50. Long Term Resource Monitoring—Aquatic Vegetation. Available online: https://umesc.usgs.gov/data_library/vegetation/vegetation_page.html (accessed on 22 September 2021).
  51. Long Term Resource Monitoring—Water Quality. Available online: https://umesc.usgs.gov/data_library/water_quality/water_quality_page.html (accessed on 22 September 2021).
  52. Rohweder, J.J.; Carhart, A.M.; Larson, D.M. 22 Years of Aquatic Plant Spatiotemporal Dynamics in the Upper Mississippi River—Derived Spatial Data (Pools 4, 8, and 13); U.S. Geological Survey Data Release; USGS: La Crosse, WI, USA, 2023. [Google Scholar] [CrossRef]
Diversity 15 00523 g001 550
Figure 1. Map of the Upper Mississippi River, USA Left panel shows a map of study reach locations and extent of the three floodplain pools. Right panels show locations of arrowheads (Sagittaria spp.; purple), wild rice (Zizania aquatica; green), and wild celery (Vallisneria americana; blue) in Pools 4, 8, and 13 during 1998–2019. Zizania aquatica was absent from Upper Pool 4 and Pool 13.
Figure 1. Map of the Upper Mississippi River, USA Left panel shows a map of study reach locations and extent of the three floodplain pools. Right panels show locations of arrowheads (Sagittaria spp.; purple), wild rice (Zizania aquatica; green), and wild celery (Vallisneria americana; blue) in Pools 4, 8, and 13 during 1998–2019. Zizania aquatica was absent from Upper Pool 4 and Pool 13.
Diversity 15 00523 g001
Diversity 15 00523 g002 550
Figure 2. Non-metric multidimensional scaling (NMDS) ordination in two dimensions (NMDS 1, NMDS 2) of the relative abundance of 25 aquatic vegetation species and associated environmental drivers in Pools 4, 8, and 13 in the Upper Mississippi River, USA, from 1998–2019. Drivers occurred along vectors and included TN = total nitrogen, Q = discharge, SS = suspended solids, Chl a = chlorophyll a, and TP = total phosphorus. Substrate represents a fine- to coarse-scale gradient (hard clay < silt/clay < sand < gravel/rock). Aquatic vegetation species are color-coded based on lifeform, and the common names are cross referenced in Table 1. Bray–Curtis similarity; 2D stress = 0.20.
Figure 2. Non-metric multidimensional scaling (NMDS) ordination in two dimensions (NMDS 1, NMDS 2) of the relative abundance of 25 aquatic vegetation species and associated environmental drivers in Pools 4, 8, and 13 in the Upper Mississippi River, USA, from 1998–2019. Drivers occurred along vectors and included TN = total nitrogen, Q = discharge, SS = suspended solids, Chl a = chlorophyll a, and TP = total phosphorus. Substrate represents a fine- to coarse-scale gradient (hard clay < silt/clay < sand < gravel/rock). Aquatic vegetation species are color-coded based on lifeform, and the common names are cross referenced in Table 1. Bray–Curtis similarity; 2D stress = 0.20.
Diversity 15 00523 g002
Diversity 15 00523 g003 550
Figure 3. Boxplots of arrowheads (Sagittaria latifolia and S. rigida), wild celery (Vallisneria americana), and wild rice (Zizania aquatica) in Pools 4, 8, and 13 in the Upper Mississippi River, USA, in relation to 9 hypothesized environmental predictors (A–I). The boxes encapsulate the interquartile range where the middle line is the median, the whiskers are 1.5 times the interquartile range, and the dots are outliers. Extreme outliers for concentrations of total suspended solids (1500 mg/L) and chlorophyll a (500 µg/L) at V. americana sites were not plotted. Substrate categories (E) included hard clay (1), silt/clay (2), mostly silt with sand (3), mostly sand with silt (4), sand (5), and gravel/rock (6).
Figure 3. Boxplots of arrowheads (Sagittaria latifolia and S. rigida), wild celery (Vallisneria americana), and wild rice (Zizania aquatica) in Pools 4, 8, and 13 in the Upper Mississippi River, USA, in relation to 9 hypothesized environmental predictors (A–I). The boxes encapsulate the interquartile range where the middle line is the median, the whiskers are 1.5 times the interquartile range, and the dots are outliers. Extreme outliers for concentrations of total suspended solids (1500 mg/L) and chlorophyll a (500 µg/L) at V. americana sites were not plotted. Substrate categories (E) included hard clay (1), silt/clay (2), mostly silt with sand (3), mostly sand with silt (4), sand (5), and gravel/rock (6).
Diversity 15 00523 g003
Diversity 15 00523 g004 550
Figure 4. Two-dimensional kernel density estimation of non-metric multidimensional scaling (NMDS) site values in Lower Pool 4, Pool 8, and Pool 13 in the Upper Mississippi River, USA, based on the entire macrophyte assemblage (25 species) included in Figure 2. The color gradient represented annual changes from 1998–2019. Upper Pool 4 was not shown due to the small sample sizes and low macrophyte abundance and diversity.
Figure 4. Two-dimensional kernel density estimation of non-metric multidimensional scaling (NMDS) site values in Lower Pool 4, Pool 8, and Pool 13 in the Upper Mississippi River, USA, based on the entire macrophyte assemblage (25 species) included in Figure 2. The color gradient represented annual changes from 1998–2019. Upper Pool 4 was not shown due to the small sample sizes and low macrophyte abundance and diversity.
Diversity 15 00523 g004
Diversity 15 00523 g005 550
Figure 5. Biplot of annual, mean non-metric multidimensional scaling (NMDS) values at sites with ≥20 percent cover of wild rice (Zizania aquatica) detected at any time within the 22 years (1998–2019) in Pools 4 and 8 of the Upper Mississippi River, USA Times series are represented with a purple (1998) to yellow (2019) gradient.
Figure 5. Biplot of annual, mean non-metric multidimensional scaling (NMDS) values at sites with ≥20 percent cover of wild rice (Zizania aquatica) detected at any time within the 22 years (1998–2019) in Pools 4 and 8 of the Upper Mississippi River, USA Times series are represented with a purple (1998) to yellow (2019) gradient.
Diversity 15 00523 g005
Diversity 15 00523 g006 550
Figure 6. Curve fit regression output for arrowheads (Sagittaria latifolia and S. rigida, combined), wild rice (Zizania aquatica), and wild celery (Vallisneria americana) in Lower Pool 4, Upper Mississippi River, USA Y-intercept indicates the estimated percent cover or relative abundance in 1998 (A–C), correlation indicates the slope of percent cover or relative abundance from 1998–2019 (D–F), and (G–I) indicates the estimated percent cover or relative abundance in 2019. Correlation values less than −0.434 are significant negative slopes (blue) and correlation values greater than 0.434 are significant positive slopes (red). Intermediate values indicate no significant change. Areas that are not sampled are shown in gray.
Figure 6. Curve fit regression output for arrowheads (Sagittaria latifolia and S. rigida, combined), wild rice (Zizania aquatica), and wild celery (Vallisneria americana) in Lower Pool 4, Upper Mississippi River, USA Y-intercept indicates the estimated percent cover or relative abundance in 1998 (A–C), correlation indicates the slope of percent cover or relative abundance from 1998–2019 (D–F), and (G–I) indicates the estimated percent cover or relative abundance in 2019. Correlation values less than −0.434 are significant negative slopes (blue) and correlation values greater than 0.434 are significant positive slopes (red). Intermediate values indicate no significant change. Areas that are not sampled are shown in gray.
Diversity 15 00523 g006
Diversity 15 00523 g007 550
Figure 7. Curve fit regression output for arrowheads (Sagittaria latifolia and S. rigida, combined), wild rice (Zizania aquatica), and wild celery (Vallisneria americana) in Pool 8, Upper Mississippi River, USA Y-intercept indicates the estimated percent cover or relative abundance in 1998 (A–C), correlation indicates the slope of percent cover or relative abundance from 1998–2019 (D–F), and (G–I) indicates the estimated percent cover or relative abundance in 2019. Correlation values less than −0.434 are significant negative slopes (blue) and correlation values greater than 0.434 are significant positive slopes (red). Intermediate values indicate no significant change. Areas that are not sampled are shown in gray.
Figure 7. Curve fit regression output for arrowheads (Sagittaria latifolia and S. rigida, combined), wild rice (Zizania aquatica), and wild celery (Vallisneria americana) in Pool 8, Upper Mississippi River, USA Y-intercept indicates the estimated percent cover or relative abundance in 1998 (A–C), correlation indicates the slope of percent cover or relative abundance from 1998–2019 (D–F), and (G–I) indicates the estimated percent cover or relative abundance in 2019. Correlation values less than −0.434 are significant negative slopes (blue) and correlation values greater than 0.434 are significant positive slopes (red). Intermediate values indicate no significant change. Areas that are not sampled are shown in gray.
Diversity 15 00523 g007
Diversity 15 00523 g008 550
Figure 8. Curve fit regression output for arrowheads (Sagittaria latifolia and S. rigida, combined) and wild celery (Vallisneria americana) in Pool 13, Upper Mississippi River, USA Wild rice (Zizania aquatica) was not detected during LTRM sampling in Pool 13. Y-intercept indicates the estimated percent cover or relative abundance in 1998 (A,B), correlation indicates the slope of percent cover or relative abundance from 1998–2019 (C,D), and (E,F) indicates the estimated percent cover or relative abundance in 2019. Correlation values less than −0.434 are significant negative slopes (blue) and correlation values greater than 0.434 are significant positive slopes (red). Intermediate values indicate no significant change. Areas that are not sampled are shown in gray.
Figure 8. Curve fit regression output for arrowheads (Sagittaria latifolia and S. rigida, combined) and wild celery (Vallisneria americana) in Pool 13, Upper Mississippi River, USA Wild rice (Zizania aquatica) was not detected during LTRM sampling in Pool 13. Y-intercept indicates the estimated percent cover or relative abundance in 1998 (A,B), correlation indicates the slope of percent cover or relative abundance from 1998–2019 (C,D), and (E,F) indicates the estimated percent cover or relative abundance in 2019. Correlation values less than −0.434 are significant negative slopes (blue) and correlation values greater than 0.434 are significant positive slopes (red). Intermediate values indicate no significant change. Areas that are not sampled are shown in gray.
Diversity 15 00523 g008
Table
Table 1. Aquatic vegetation species (n = 25) found at >1% of sites and used in the non-metric multidimensional scaling (NMDS) ordination. Life forms include submersed (S), emersed (E), rooted floating leaved (RFL), free floating (FF), and filamentous algae (A). Species list is ordered by combined number of occurrences in Pools 4, 8, and 13 of the Upper Mississippi River, USA.
Table 1. Aquatic vegetation species (n = 25) found at >1% of sites and used in the non-metric multidimensional scaling (NMDS) ordination. Life forms include submersed (S), emersed (E), rooted floating leaved (RFL), free floating (FF), and filamentous algae (A). Species list is ordered by combined number of occurrences in Pools 4, 8, and 13 of the Upper Mississippi River, USA.
Common NameScientific NameSpecies CodeLife FormNumber of OccurrencesMean Relative Abundance
CoontailCeratophyllum demersum L.CEDE4S79754.477
Canadian waterweedElodea canadensis Michx.ELCA7S68023.967
American wild celeryVallisneria americana Michx.VAAM3S57344.177
Water stargrassHeteranthera dubia (Jacq.) MacMill.ZODUS51261.857
Narrow-leaved pondweedPotamogeton foliosus Raf., P. pusillus L.NLPWS44861.891
Filamentous algaenot identified furtherALGAA42121.928
DuckweedsLemna, Spirodela, Wolfia spp., combinedNRFLFF41766.811
Sago pondweedStuckenia pectinata (L.) BörnerPOPE6S40781.403
White waterlilyNymphaea odorata Aiton NYTURFL36864.322
Eurasian watermilfoil *Myriophyllum spicatum L.MYSP2S32790.942
Curly pondweed *Potamogeton crispus L.POCR3S30600.945
American lotusNelumbo lutea Willd.NELURFL26794.117
Flatstem pondweedPotamogeton zosteriformis Fern.POZOS21510.704
Broadleaf arrowheadSagittaria latifolia Willd.SALA2E17432.855
Stiff arrowhead Sagittaria rigida Pursh.SARIE12571.593
Longleaf pondweedPotamogeton nodosus PoirPONO2S12180.306
Wild riceZizania aquatica L.ZIAQE10731.120
Nodding waternymphNajas flexilis (Willd.) Rostk. and SchmidtNAFLS7440.267
Reed canarygrass *Phalaris arundinacea L.PHAR3E4640.556
Giant burreedSparganium eurycarpum Engelm. SPEUE4340.520
Southern waternymphNajas guadalupensis (Spreng.) MagnusNAGUS4330.202
River bulrushBolboschoenus fluviatilis Torr. SojákSCFLE4230.576
Longbeak buttercupRanunculus longirostris Godr.RALO2S3410.113
Brittle waternymphNajas minor All.NAMIS2570.105
Northern watermilfoilMyriophyllum sibiricum KomarovMYSIS2470.071
* Indicates not native/invasive to the Upper Mississippi River Basin.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Carhart, A.M.; Rohweder, J.J.; Larson, D.M. 22 Years of Aquatic Plant Spatiotemporal Dynamics in the Upper Mississippi River. Diversity 2023, 15, 523. https://doi.org/10.3390/d15040523

AMA Style

Carhart AM, Rohweder JJ, Larson DM. 22 Years of Aquatic Plant Spatiotemporal Dynamics in the Upper Mississippi River. Diversity. 2023; 15(4):523. https://doi.org/10.3390/d15040523

Chicago/Turabian Style

Carhart, Alicia M., Jason J. Rohweder, and Danelle M. Larson. 2023. "22 Years of Aquatic Plant Spatiotemporal Dynamics in the Upper Mississippi River" Diversity 15, no. 4: 523. https://doi.org/10.3390/d15040523

APA Style

Carhart, A. M., Rohweder, J. J., & Larson, D. M. (2023). 22 Years of Aquatic Plant Spatiotemporal Dynamics in the Upper Mississippi River. Diversity, 15(4), 523. https://doi.org/10.3390/d15040523

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