Research findings

Research Findings by Discipline:

Applied mathematics (Pasour)
Aquatic macroinvertebrates (Bain)
Aquatic macrophytes (Johnson)
Fish (Arend, Bain, Singkran)
Food web dynamics (Arend)
Human dimensions (Abrams, Buck, Morad, Pendall, Wei)
Hydrodynamics (Cowen, King, Pasour, Tanaka)
Land use change (Meixler)
Paleoecology (Distler, Leopold)
Synthetic project-wide analysis (All project members)
Zooplankton & Phytoplankton (Doyle-Morin, Hairston)
Water chemistry (Chen, Driscoll)
Watershed modeling (Loucks, Hamade)
Wetland ecology (Distler, Hajek, Hotaling, Kelsall, Leopold)

 

back to top

 

---

Synthetic project-wide analysis: patterns among 2002 averages

A preliminary look at each group’s 2002 data, calculated both as an average from the growing season and as a July 2002 average, suggested some interesting characteristics for each of the bays. When combined in a table in trophic order, we see that the embayments do indeed exhibit very different trophic dynamics relative to one another (Figure 1). Blind Sodus communities exhibit classic top-down control behavior, with very high piscivorous fish biomass (piscivores plus piscivorous omnivores), which presumably drive the low amounts of invertivorous fish (planktivores plus planktivorous omnivores). Few invertivores release the next trophic level, leading to high zooplankton abundance, which graze down the phytoplankton community to the second-lowest average biovolume of the eight communities. Sterling Pond also exhibits a modified topdown system, with a large amount of piscivores, and medium amount of invertivores. These invertivorous fish appear be at high enough total biomass to control the zooplankton population, which is the smallest of the eight systems. The phytoplankton community is released, in turn, and hosts the highest total average biovolume of edible phytoplankton. It is interesting that this bay supports such a high phytoplankton biovolume because there it also hosts a dense macrophyte community. This high vegetation abundance could be due to the high nutrient load to this system (P and N), and correlated high ambient nutrient conditions, fueling the growth of the phytoplankton and macrophyte communities.

Comparatively, Floodwood has even higher nutrient loading and ambient nutrient conditions, yet it is not as clearly represented in the plankton or fish communities. The plankton communities are both at low abundance, which is likely a result of the high-flow environment. The fish were at medium biomass values relative to the other bays, but it is unclear as to whether or not they are subsidized from nearby low-flow environments, as the benthos does contain a relatively high abundance of insects. Little Sodus, South Colwell, and North Sandy plankton and fish communities also compare in relative biomass to the Floodwood communities, but these three communities do not receive the nutrient inputs or have the ambient nutrient levels that Floodwood receives. Of the three embayments, North Sandy receives the greatest nutrient load, and also houses the highest phytoplankton biovolume, but this is as far as the excess nutrients are expressed up the food-web. North Sandy actually contains the least amount of fish of the three embayments.

The adjacent South Sandy pond, on the other hand, has high fish biomass, including the highest amounts of intertivores, yet also has a relatively high amount of zooplankton, including the highest predatory zooplankton biomass level (which are typically targeted by invertivores). South Sandy also supports a relatively large phytoplankton community, although the large zooplankton community may be driving the fact that it is predominantly composed of inedible species. And yet, this embayment does not receive a high nutrient load through the tributaries and is not high in either ambient N or P (although the wetland N values are high). This suggests that these communities are very efficiently using the nutrients available to them, and in fact, plankton capable of fixing their own nitrogen are most abundant in this system. Finally, Juniper is unique to the eight embayments for a number of reasons. It does not receive a great deal of nutrients from the watershed, and has relatively low ambient nutrient levels. The phytoplankton are able to take advantage of what is there, however, and as a result, Juniper supports the largest phytoplankton biovolume of the eight systems. This pattern does not transfer up the food chain, however. The zooplankton abundance falls in the middle of the pack relative to the other embayments, and may not be able to take advantage of the excessive phytoplankton abundance because the community is composed primarily of inedibles. The zooplankton community is not being substantially reduced from above, however, as Juniper is home to the smallest fish communities, likely because the small embayment could freeze solid in the winter.

With these very different trophic patterns, it seems unlikely that we would find strong generalizable patterns among all eight embayments. Preliminary comparisons of each of the food web’s components against one another suggest that this might be the case. However, there are some notable exceptions. One of the most common patterns found in freshwater systems is a direct linear correlation between phytoplankton abundance and phosphorus levels—as phosphorus levels increase in these typically P limited systems, phytoplankton also increase in abundance. An initial look at this pattern in our embayments using 2002 values that have been averaged over the growing season clearly does not support that pattern (Figure 2). This scattered spread of points does not suggest a strong linear correlation, much less a positive linear correlation, and is representative of the seemingly random spread of points seen in many of these comparisons.

However, when you compare the average phosphorus and phytoplankton values averaged only over July (Figure 3), the height of the growing season, the positive linear correlation appears. This change in pattern is driven in part by a dramatic increase in both phosphorus and phytoplankton in July in Sterling Pond (the point in the upper righthand corner of the plot). This is not a surprising shift when you consider the different flow regimes of the eight systems. Average May discharge values vary from 0.02 to 45.8 CMS, while they become more similar among the embayments in July, varying from 0.002 to 1.6 CMS. Nutrient loading rates are thus more similar in July, as are phytoplankton flushing rates, which bring the impacts of external forcing to similar levels for the two components.

Similar correlations can be found between the chemical and biological constituents of the systems when we look at the entire fish community and phosphorus load levels in 2002 (Figure 4).

back to top

---

Fish, macroinvertebrates and land cover change through time

Patterns and Processes Shaping Habitats and Pelagic Communities

The quest for the relationship between patterns and processes has a cornerstone in the field of spatial ecology. In particular, it is the aim of many ecologists to discern the dynamic functionality of ecosystems. Life history and dispersal attributes of species have been coupled with landscape models to infer plausible mechanisms shaping patterns of species distributions and abundances. It is the interplay of space and ecological traits what seems to affect the scale of populations and the nature of communities. The use of fractals in ecology is currently pervasive over many areas. However, very few studies have linked fractal properties of landscapes to generating ecological mechanisms and dynamics. Our work showed that lacunarity (a measure of the landscape texture) is a well suited ecologically scaled landscape index that can be explicitly incorporated in metapopulation models. The average lacunarity of a landscape is linearly correlated to the habitat that a species may perceive. This advancement provides a general approach to multifractal landscapes, and the specific effect of landscape heterogeneity. It also introduces a novel dimension to population modeling and conservation planning.

Our project is aimed at the role water residence time has on the structure of aquatic ecosystems. However, we are finding that water flow and nutrient loads are intertwined to explain biotic patterns such as the loss or gain of phytoplankton biomass. Flowthrough chemostat models can include the effects of external (i.e. flowthrough rate and nutrient concentration) and internal (e.g. population growth rate) variables. Aquatic ecosystems, such as estuaries or embayments, have often been modeled as simple variants of chemostats. Flow rate and nutrient input into the system can largely delimit extinction thresholds for plankton. However, real ecosystems are often split into spatial compartments where the flow of water and particles is not homogeneous, but follows recirculation patterns (See Figure 5 for one of our sites). Using former approaches for spatially homogeneous populations, we derive analytical solutions for critical dilution rates (aka extinction thresholds for phytoplankton) under different flow network configurations. We showed that eigenvalues for connectivity matrices can be explicitly incorporated into the classical critical dilution rates for simple chemostats. More importantly, we prove an inverse relationship between water residence time and nutrient input defining phytoplankton persistence in any flow network configuration. Thus, our research amplifies classical results from simple chemostats to multipatch networks and shows the role of recirculation patterns in the increase of water residence times, and henceforth, the persistence of plankton in lentic systems.

Going further, we hypothesize that the effect flow rate and nutrient load have on higher trophic levels will be similar. Future work will incorporate this simple formulation in the parameterization of phytoplankton washout rates and water residence times across embayments with different discharge conditions and sizes. We are also using artificial substrates in the pelagic zone of all embayments to detect changes in colonizing communities in different hydrologic settings. As expected from past applications of this technique, diverse (16 to 32 taxa) and abundant (ca. 600 to 12,000 organisms per set) assemblages of organisms colonized the substrates during the set period (1 month). An analysis of these assemblages consistently revealed characteristics of well-developed communities: a few dominant taxa with decreasing numbers of subdominant and sparse taxa. The composition of the colonizing assemblages were similar in dominant taxa but varied sharply in taxa collected in moderate numbers. Organism abundances varied among months and embayment, with a significant relation with embayment watershed area, total P, and embayment depth. These variables suggest that embayment productivity is influencing organism abundances. Taxa diversity also varied among months and embayment with significantly relations to bay properties of water depth and SO4 concentration (indicator of water mixing). This finding suggests that water residence time influences taxonomic composition of the colonizers. We are now seeking ways to link taxonomic structure to plankton and hydrology to infer mechanisms producing variation among embayments.

Biotic Patterns Along Aquatic Gradients

The influence of changing landscapes and environments on organism distribution along gradients and across ecotones (transition zones) of contrasting habitats has not been well studied in aquatic ecosystems. The ecotone concept is almost as old as the field of ecology and broadly useful because ecotones change biodiversity, biotic interactions, and material and nutrient flows. Ecotones were detected and anlyzed on embayment-stream gradients associated with Lake Ontario using abrupt changes in habitat variables and peak turnover rates in fish species compositions. The results revealed that the ecotones were static in their orientations on both gradients (Figure 6). Abrupt changes in the habitat variables and peak species turnover rate showed strong congruence at the same location on one gradient. On another gradient, only the peak species turnover rate could define an ecotone, whereas the abrupt changes in the habitat variables occurred at multiple locations. Four ecotone functions (biotic aggregator, mediator, soft barrier, and hard barrier) were inferred from a comparison of relative abundance of fish species among downstream, ecotonal, and upstream habitats. In general, the ecotone acted as a hard barrier for most of the fish species. The distinct physical discontinuities below and above the ecotonal habitat appear to be an important factor shaping and impeding fish distributions along complex habitat gradients. An expansion of this research to landscape scale properties revealed additional spatial factors explaining fish distributions along baystream gradients. Regression analyses was used to select the best habitat and land cover variables for explaining variations in distributions of fish. The distribution patterns of most fish species were explained by either a set of the selected habitat or land cover predictor variables or some combination. Key habitat factors were water depth, current velocity, aquatic plants, algae, woody debris, sand, and rock-bedrock while the most important land cover types were wetlands, forest plantations, shrub swamps, roads, and urban areas.

Although distribution of organisms is a key mechanism underlying the population dynamics, few empirical studies have quantitatively determined organism distribution in association with both changes in habitat and population characteristics (birth, death, and migration) over space and time. A spatially explicit abundance exchange model (AEM) was developed to predict distribution patterns of fish species in relation to their population characteristics and habitat preferences along the bay-stream gradients. Preference indexes of each target fish species for water depth, water temperature, current velocity, cover types, and bottom substrates were estimated from the field observations, and these were used to compute habitat preference of the associated fish species. Fish habitat preference was a key variable in the AEM to quantify abundance exchange of an associated fish species among habitats on each study gradient. AEM can efficiently determined local distribution ranges of the fish species. Model validation supported the AEM for quantifying most species distributions. Overall, With its flexible structure that is applicable for array functions and differential equations from both static and dynamic components, the AEM can be modified to determine patterns of organism distribution in complex systems with different environments and geography.

Biotic Patterns at the Upper Trophic Levels

We explored the relationship between embayment morphometry and hydrology and fish community structure among eight project embayments. Embayments ranged in surface area and depth, varied in their connections to Lake Ontario and their watersheds, and drained watersheds representing a gradient of agricultural to forested. Most embayment fish communities were dominated numerically by yellow perch and centrarchids (e.g., pumpkinseed and largemouth bass; Figure 7), whereas biomass was dominated by piscivorous fishes including brown bullhead, bowfin, and northern pike. We related various physicochemical factors, including total phosphorus load, embayment area, and vegetation, to fish community relative abundance and biomass. Abundance was positively related to percent vegetation within embayments and fish biomass was positively related to total phophorus loading and embayment area. The importance of total phosphorus loading and vegetation in structuring fish communities indicates a strong role by internal mechanisms shaping embayment biota.

Four embayments received extra attention to explore differences in habitat and connectivity with adjacent systems can have on energy flow and trophic structure. Two embayments were shallow and dominated by littoral habitat: one received direct tributary inputs from an agricultural watershed and was permanently connected to Lake Ontario, the other received wetland inputs from a forested watershed and was intermittently connected to Lake Ontario. The other two embayments were deep and dominated by pelagic habitat but differed in watershed area and connection to Lake Ontario. In each embayment, we measured delta 13C and delta 15N values of seston and epilithon during the summer of 2004 to use as isotope baselines to estimate the relative contribution of pelagic (seston) and littoral (epilithon) energy pathways to the fish community and the relative trophic positions for each fish species. Baseline stable isotope values varied among and within embayments, reflecting differences in the dominant water sources to each embayment and in watershed land use. The shallow and littoral dominated embayments had fish communities that relied on a combination of littoral and pelagic energy sources. In contrast, fish in the deep and large embayments depended primarily on phtyoplankton-derived carbon.

Many Great Lakes embayment fish communities are dominated numerically by yellow perch populations. We hypothesized that variation among embayment yellow perch population size structure results from differences in embayment morphometry that influence prey availability to and consumption by yellow perch. To test this hypothesis, we developed an energy budget model for yellow perch based on bioenergetics equations and literature estimates of energy storage and allocation to somatic and gonadal tissue. We used the model to predict yellow perch growth in four southeastern Lake Ontario embayments; the two shallow embayments dominated by littoral habitat and two relatively deep embayments dominated by pelagic habitat. Estimates of ration and the proportion of different prey types consumed were generated from diet contents of yellow perch collected in each embayment. Growth was predicted by the model to be fastest in the two shallow embayments, largely due to greater consumption of amphipods and non-Dipteran aquatic insect larvae. Model growth estimates were compared to growth curves generated from the otoliths of over 60 individuals collected in each embayment. In comparison to otolith growth estimates, model predictions overestimated yellow perch growth and did produce the same ordering of faster to slower growing embayment populations. While growth projects varied, the same ordering of systems indicates that morphology and habitat structure were important internal factors in the study systems.

Applications in Conservation and Management

Great Lakes coastal wetlands are seeing increased management attention and interest in developing indicators for monitoring. We collected and analyzing data in conjunction with other Great Lakes projects to define a long-term basin-wide monitoring plan. The indicators included several methods for invertebrate data collection, several for fish data collection, submerged and emergent wetland plant identification, water chemistry, water levels, sediment flow, and landscape attribute assessment. We further augmented our data by performing collections in all study waters multiple times over the summer season to gain a better understanding of temporal changes of indicators within wetlands. After our field study, we evaluated indicators of wetland degradation using data from six teams of investigators that conducted standardized collections of field data on the same flora, fauna, physical, and landscape level data using standardized protocols. Based on overall indicator evaluation, we recommend that monitoring programs sample water chemistry, fish, macroinvertebrates, and landscape attributes as these indicators have the highest degree of sensitivity, applicability, measurability and complementary data availability for the lowest cost.

We expanded our analyses of wetland quality monitoring by developing and demonstrating a geographic information system (GIS) modeling method for coastal zone environmental assessment. Hydrologic and habitat mapping was conducted to identify land areas and stream segments associated with highly altered streamflows, degraded habitats, and fragmented stream courses. These locations were then identified on a large scale GIS for targeting enhancement actions (see example in Figure 8). To this we added government and community attributes indicative of successful watershed conservation to determine the capacity of local governments and communities to undertake land planning activities. The combination of hydroecological enhancement need and community capacity maps provide a visual guide for allocating funding for enhancing the ecological health of the Great Lakes Basin. Only recently has all the information been compiled and mapped. However, it is clear that our modeling procedures have considerable potential to identifying likely environmental enhancement sites in the context of coastal zone management. Further study is also underway on GIS-based methods so we can identify pollution sources for better impairment projections.

Change in land use through time

In both the eastern and western watersheds, the land area in open water, tree plantations, mixed herbaceous wetlands, and farm buildings has not changed more than 1% between 1938 and the late 1990s (Table 1). Commercial holdings, barren lands, golf courses, and shrub-dominated wetlands have always existed in proportions of 1% or less in all periods between the 1930s and the late 1990s in both watersheds. Residential areas have increased in the western watersheds from negligible to 3% between 1938 and 2000 and forested lands have increased from 18% to 46% over the same time period. Meanwhile, pasture and croplands in the western watersheds have decreased in area from 6% to 2% and 66% to 33%, respectively and woody wetlands and abandoned lands have varied over time between 2% and 8% and 2% and 6%, respectively in the western watersheds. The eastern watersheds have experienced somewhat different trends. In this region, cropland has decreased from 37% to 18% and land under pasture has decreased from 4% to negligible. Forest has increased from 40% to 61%. Residential and urban lands have increased from 1% in 1938 to 5% in 1995. Woody wetlands have varied through time from as low as 4% to a high of 13%. Abandoned lands increased from 3% to 6% between 1938 and 1955 but then decreased again to 3% in 1975 and have remained at that percentage since then.

back to top

---

Embayment hydrodynamics

We studied in great detail two canonical lake-embayment systems, Little Sodus Bay (LSB) and Sterling Pond (SP). These were identified as canonical systems in that LSB has essentially no watershed and its exchange physics with Lake Ontario (LO) is enteriley determined by transport processes through its permanent connection. SP on the other hand has a significant watershed and is a relatively small waterbody hence its exchange with LO is a balance of watershed and permanent connection transport physics. We note that LSB is relatively unaffected by macrophytes while LSB is at times dominated by macrophytes.

In LSB we analyzed the exchange between a weakly forced lacustrine embayment and a large lake through a long, shallow channel. Exchange in the channel is the result of a multiple and subtle balance in which spatial thermal variations (baroclinic forcing), oscillations in the water level (barotropic forcing), bottom friction, diffusion, wind forcing, and the effects of unsteadiness are all important. Temperature gradients across the channel result from differences in thermal inertia of the embayment and lake at seasonal time scales and differences in the wind-driven internal dynamics of the lake and the embayment at diurnal to synoptic time scales. These gradients are, in general, weak and barotropic forcing dominates the channel momentum balance; however, episodic upwelling in the lake can shift the balance in favor of baroclinic dominance. A combination of scaling, analysis of observational data, and three dimensional simulations was used to demonstrate that bed stress, vertical turbulent diffusion, wind stress, and unsteadiness effect exchange relative to the predictions of internal hydraulic theory; the quasisteady inviscid theory that describes the motion resulting from a purely barotropic/baroclinic force balance. The average length of time water remains within the boundaries of an aquatic system is a key parameter controlling the system's biogeochemical behavior. This timescale, generally referred to as the hydraulic residence time, provides a first-order description of multiple and complex processes that drive transport. The procedures to estimate these transport timescales were reviewed a common language amonst the interdisciplinary team developed around residence time scales. Through the analysis of numerical simulations, the links between residence timescales and the underlying hydrodynamic processes in LSB were explored. LSB has negligible through-flow and is connected permanently to LO through a narrow and shallow channel. Exchange in the channel is the result of a multiple balance where baroclinic forcing, barotropic forcing, frictional mixing, wind, and the effects of unsteadiness are all important. The simulations indicate that baroclinic processes are the dominant exchange and mixing mechanisms in embayments like LSB. The largest density gradients across the channel are caused by episodic upwelling events in LO during the stratified season, when exchange rates increase by at least an order of magnitude. The mean residence timescales undergo dramatic variations in time and space and, in general, are comparable to the timescales of the systems variability itself. The simulations reveal that temporal variations of mean residence timescales occur at inter-annual, seasonal, and down to synoptic timescales, and are closely related to the occurrence and frequency of upwelling events. The major findings of the LSB line of research can be found in Rueda and Cowen (2005a,b).

Sterling Pond allowed us to extend our analysis to include the effects of high watershed inflow and strong macrophyte forcing. We collaborated strongly with Robert Johnson’s group who made extensive field macrophyte surveys, and carried out field dye studies and numerical simulations. These field and numerical studies allowed us to decipher the qualitative effects of macrophyte canopies on the hydrodynamics and residence time scales of SP. These results are now reviewed in more detail as the manuscript describing them is in final preparation.

Johnson’s research group conducted surveys of macrophyte species composition, dry weight, abundance, surface area, and height from the bottom on a sampling grid of 100 m x 100 m quadrates. Biomass and species abundance were measured in each quadrate by different methods in 2002 and 2003. Figure 9 shows the quadrates that had high macrophyte aboveground biomass (gdw/m2) on the sampling dates near each of the dye studies.

Sterling Pond is a small and shallow embayment that is on average about 1.5 m deep, 600 m long (in the direction of net through flow), and 400 m wide. SP is a natural retention basin that drains Sterling Creek, bordering wetlands, and a large watershed (200 km2) into LO through a long, narrow, maintained channel (2 m deep, 17 m wide, and 100 m long). From late spring through early fall, SP is home to diverse populations of aquatic macrophytes which experience one or more dense blooms (Figure 9). As there is some variability in terminology between the terrestrial and aquatic vegetated flow research communities and the aquatic macrophyte ecology research community we will now precisely define our terminology:

Our research group carried out three dye studies to measure the residence time distribution, or RTD (for one specific release point in space and time – see Rueda & Cowen, 2005b for details), during periods of different hydrodynamic forcing and macrophyte abundance. In each field experiment, a passive tracer was released locally in space and time where Sterling Creek enters SP. By monitoring the velocity profile and the dye concentration in the channel connecting SP to LO, the dye flux out of SP was determined. The first moment of the dye flux temporal history out of the pond (expressed as the fraction of total dye per unit time, which is the residence time distribution, or RTD) is the mean hydraulic residence time.

The setup for the three experiments is illustrated in figure 10. One or two gallons of Rhodamine WT (20% by weight solution) fluorescent dye was released through a vertical line-source diffuser. An RDI 1200 KHz Workhorse Monitor acoustic Doppler current profiler (ADCP) recorded the velocity profile at mid-channel using a highresolution pulse-coherent mode and facing upward from the bed. In the May 2002 and fall 2003 experiments, a calibrated Turner Designs 10-AU flow-through fluorometer was positioned with its intake tube in the center of the channel to record dye concentration. In the fall 2002 experiment, two Wetlab FlashPak submersible flow-through fluorometers were positioned in the upper and lower halves of the channel.
Temperature profiles were recorded with Sea Bird SBE-39 thermistors every 1-2 minutes at the locations shown in figure 10. Water surface elevation was measured in SP at the dye release site using a bottom-mounted SBE-39 temperature / pressure logger. A Campbell Scientific meteorological station mounted at LSB (1 km west of SP) recorded solar radiation, air temperature, atmospheric pressure, relative humidity, and wind speed and direction at 15-minute intervals.

The three dye experiments were simulated using the numerical model Si3D (discussed in Rueda and Cowen, 2005b). The model does not presently account for the presence of vegetation. The resolution was 17.5 m in the horizontal and 0.5 m in the vertical. Time-varying boundary conditions, including stream flow and temperature in Sterling Creek, water surface elevation and temperature in LO, and wind and meteorological data over the surface of SP, were measured continuously during the dye experiments and used to specify the model boundary conditions. In each of the experiments, the initial temperature field was specified, and the system was allowed to warm up for a 24-hour period (the dominant barotropic forcing time scale at SP is order 20 minutes) before the dye release. A uniform drag coefficient of 0.002 was specified for the mud bottoms. The dye release was simulated within Si3D as an instantaneous release of a 3-D Gaussian plume of passive tracer at the mean time of the actual dye release. Measured and modeled residence time distributions are plotted in figure 11.

During the May 2002 dye experiment, flow in Sterling Creek was high, and the aquatic vegetation was sparse and deeply submerged. 90% of the dye exited SP within 7 hours, and the RTD is well approximated by classic plug flow with longitudinal dispersion (results not shown). In the fall 2003 dye experiment, stream flow was low, and exchange with LO drove SP hydrodynamics, mediated by a thick, somewhat uniform (in density, though not in species – see figure 9) canopy of emergent macrophytes filling the entire pond. As expected, because it does not account for macrophytes, Si3D underpredicts the arrival time and overall residence time of the dye, but after the dye arrives at the channel in the simulation, Si3D accurately reproduces the phase, though not the amplitude, of the residence time curve (figure 11). In summary, the uniform canopy of macrophytes damped the barotropic and baroclinic forcing and lengthened the mean residence time, but did not change the shape of the residence time distribution.

As in 2003, stream flow in 2002 was low and exchange with LO drove SP hydrodynamics. A dense macrophyte canopy again filled the pond, but in this case distinct patches of macrophytes were visible and surveys confirmed that the canopy density was highly heterogeneous (figure 9). Even after the initial arrival of the dye front, the RTD is not well-predicted by Si3D. Based on the comparison of field observations to the numerical simulations, it is clear that patchiness of the canopy during the fall 2002 experiment played an important role. Transects across the pond showed that exchange flow (between SP and LO) appeared to occur dominantly in the less densely vegetated sections of SP, a process similar to the short-circuiting described by Andradóttir & Nepf (2001), leaving considerable mass of dye residing within the thick patches of macrophytes. This mass was observed to move from patch to patch with changes in wind-driven circulation. While the numerical model simulated the wind stress on the water surface and the barotropic and baroclinic exchange flows with LO, it was unable to capture even their qualitative effect on residence time (figure 11) and mixing within SP. We hypothesize that this is due to the heterogeneous structure of the macrophyte canopy – a topic we are studying under current NSF (CBET – 0626164) funding as motivated by this research project.

Finally, we made limited measurements and conducted limited modeling exercises in a third embayment – Blind Sodus Bay (BSB) which is characterized by small seasonal connection to LO and strong watershed forcing. Here we found that even with the small connection to LO that the barotropic forcing of LO on BSB contributed significantly to the exchange between lake and embayment even during high tributary flow years. Inter-annual variability in watershed forcing led to a factor of 2 change in the residence time.

back to top

---

Wetland plants

Influence of Species Composition on Groundwater Solute Concentrations in a Coastal Medium Fen

Wetlands are important ecosystems for the storage, transformation and uptake of nutrients. Excesses in nutrient loading and availability have been shown to affect species composition and plant community structure. Nitrogen has received considerable attention in relation to its effects on plant communities and water quality. Wetland plant communities dominated by nitrogen-fixing species may locally influence groundwater chemistry, thereby impacting nitrogen dynamics on the landscape and possibly water quality of adjacent surface waters.

Few studies exist investigating the effects of N-fixing species on groundwater in medium fens, which harbor a large number of plant species including some plant species of concern for conservation. In order to quantify differences in local groundwater chemistry among communities dominated by two N-fixing species (Alnus incana ssp. rugosa and Myrica gale) and communities absent of N-fixing species, groundwater samples were collected and their chemistry analyzed throughout two growing seasons. The study was conducted at South Sandy Fen, a medium fen adjacent to the southeastern portion of Lake Ontario, New York.

In June of 2004, wells were installed at two depths (30 and 80 cm) at six locations within the fen (n = 12). Each pair of wells, replicated twice, were placed approximately 30 cm apart in the three distinct plant community assemblages (Alnus, Myrica, and graminoid-dominated). Wells were sampled at monthly intervals from July 2004 to December 2004. Sampling resumed in June 2005 and continues at present. Nitrate was significantly different among species, whereas ammonium was not. Nitrate and ammonium were not significantly different between well depths. There was no significant interaction between plant community type and well depth. These results indicate that N-fixing species locally influence nitrogen inputs to groundwater in medium fen communities.

Chemical and physical characteristics of Lake Ontario wetlands

Chemical and physical characteristics of eight wetlands adjacent to the southeastern and eastern portions of Lake Ontario were collected from 2001-2005. Sites were classified based on plant community type and their physical connection to Lake Ontario. With these groupings as a foundation, chemical and physical data were analyzed in order to better understand ecosystem function of the wetlands. All sites were determined to be phosphorus-limited based on N:P ratios, and all shared similar pH levels. Connectivity to the lake appeared to have no significant influence on wetland groundwater chemistry. Community types had significantly different levels of total nitrogen, calcium, magnesium and conductivity. Conductivity and magnesium were the only measured variables that showed significant differences based on wetland area. Wetland type (fen vs. marsh) appears to be the most important factor in explaining differences in groundwater chemistry among sites. Results from wetland analyses are being incorporated to data from research groups at Cornell and Syracuse Universities studying adjacent lake and upland sites.

Paleoecological Investigation of Lake Ontario Wetlands

To date, peat cores from South Sandy Pond wetland and Juniper Pond wetland have been analyzed for plant macrofossil composition, and stratigraphic (less detailed) analysis has been undertaken on two additional cores from South Sandy Pond to better characterize the spatial development patterns of this wetland. Additional analyses are still in progress, including radiocarbon dating of these cores, C:N analysis of core material, and refined identification of bryophyte fossils. Detailed and stratigraphic analyses from additional coastal and inland wetlands will also be undertaken as part of a related project, finishing by the end of 2008.

Results of this investigation are preliminary, and the lack of chronological information at this time particularly limits the scope of inference possible from these data. Nonetheless, the preliminary results give some insight into the development of coastal wetlands of eastern Lake Ontario and, in particular, the stability and longevity of meadow-marsh / fen vegetation in Great Lakes Wetlands, which is of particular interest due to the high plant diversity and numbers of rare elements associated with this vegetation type.

The macrofossil analyses from South Sandy Pond suggest initiation of a wet forest (Hemlock-Birch) vegetation after an erosional event or dune breaching. After this, accumulation of peat was associated with a longer period of Alder shrub swamp. During these periods, the higher percentage of true mosses implies a relatively minerotrophic and possibly more nutrient-rich state. The wetland underwent important changes once about 4 meters of peat had built up over the deepest areas; the peatland began extending out of the deepest trough and into the surrounding terrace, and the thickening peat began to isolate the vegetation from groundwater influence, producing a more acidic, nutrient poor wetland. Vegetation changed from shrub thicket to meadow-marsh / fen vegetation, dominated by sedges but also including leatherleaf (Chamaedaphne calyculata), sundew (Drosera intermedia), beakrush (Rhynchospora), and Sphagnum species dominating the bryophyte community. This community, with variation in the importance of different species, persisted through the accumulation of 3 additional meters of peat, likely representing over 1000 years. During this time, fire also appears to have been an important disturbance throughout the wetland, possibly reducing the recruitment of tall shrubs and trees. In the top meter, much of which represents post-colonial time, fire frequency was reduced and changes in vegetation are seen in various parts of the wetland. Swamp forest expanded into the western (now forested) portion of the wetland, possibly due to ditching and road construction in the middle of the 20th century and related hydrological changes. Typha became dominant in the north-central portion of the wetland (now dominated by Typha x glauca) during this period, and in several places, a spike in tamarack (Larix laricina) needles is seen, followed by a rapid decline. This may be a sign of the reduced effects of fire at the beginning of this period, followed by further, possibly hydrological changes.

The macrofossil record from Juniper Pond to the south and west shows some similar patterns. As in South Sandy Pond wetland, the wetland appears to have progressed from an alder shrub swamp stage toward a more acidic fen stage characterized by greater dominance of Sphagnum and acid-tolerant, low-nutrient fen species. As in South Sandy Pond, this period was associated with increased frequency of fires. This stage lasted through the accumulation of 3 meters of peat, likely over 1000 years. Unlike South Sandy Pond, however, this stage was ended by a major hydrological perturbation indicated by the presence of a silt lens and the subsequent emergence of a more shrubby minerotrophic flora. This disturbance may represent a dune breaching event or an erosional event, possibly associated with European colonization and land-use change in the surrounding uplands.

In general, macrofossil analysis suggests that coastal wetlands have gone through several stages in their development, beginning with wet-forest or shrub swamp communities and gradually building up sufficient peat to produce the more oligotrophic, acid peatland communities which today support some of the rarest plants and animals in New York State. These fen or meadow-marsh communities have been significantly changed in recent decades through reduction in fire frequency and changes in hydrology and sedimentation. These impacts have led to increase in forest cover and cattail dominance in South Sandy Pond wetland and reversion to minerotrophic shrub swamp in Juniper Pond. The degree to which these are do to human impacts will be more readily determined once the chronology of these cores is more accurately assigned.

back to top

---

Aquatic plants

Macrophyte Phenology in Sterling Pond

In 2002 and 2003, we attempted to describe the aquatic macrophyte phenology in Sterling Pond using common methods to describe plant abundance both temporally and spatially. We conducted our sampling for plant species presence and abundance of aquatic macrophytes in a manner to identify embayment-wide trends in species diversity and community structure. To choose sampling points we used the line intercepts of 100m X 100m UTM transect grids (NAD 27datum and true north) for predetermined sampling points located by GPS. Figure 12, below shows sampling points on a grid pattern 100 meters apart where we recorded relative abundance and biomass measures in Sterling Pond. We used a method of point sampling on the line intercept for this study to determine biomass by traditional quadrat collection and total plant and individual species relative abundance with a rake-toss sampling. At each sampling point we used a grapple hook (throw-rake) formed by connecting the “heads” of two garden rakes back-to-back attached to a line and tossed approximately 10m from the boat to sample the plants on the embayment bottom. At each sampling point our crew threw three rake tosses to record plant species presence and make an overall estimate of plant abundance on the rake as “dense”, “medium”, “sparse”, “trace” or “zero, then went further to estimate the percentage of individual species on each rake. We transcribed, all information onsite, onto data sheets for later entry into a data spreadsheet. Figure 12 below shows sampling points on a grid pattern 100 meters apart where we recorded relative abundance as depicted below and used the same approximate points to measure biomass in Sterling Pond.

To determine biomass at each sample point location we tossed a 0.5m X 0.5m quadrat into the water, from the boat to sample an area of 0.25 sq m. After locating the quadrat on the bottom, the diver collected all plants growing within the 0.25 sq m frame by cutting them off at the substrate-water interface. Alternatively, plants pulled from the substrate with below sediment plant material had that material removed in the plant processing before placed in the drying oven. Crew members placed the collected plant material into labeled plastic bags and stored on ice until returned to the laboratory where samples are stored in refrigerators until processed. We processed the collected material by washing with tap water to remove soil, animals, weakly adhering algae, and decayed material. We separated the plant mass to individual species, removed below sediment plant material (such as roots), and did not include roots for dry weight determination. Plant turions (winter buds) are vegetative plant parts and if they are not decayed they are included as plant material. Drying of individual species then takes place in ovens at 105oC for at least 48 hours and then plant material is weighted and recorded as species dry weight/0.25 sq m. We then are able to use dry biomass of species to describe macrophyte species and density changes within Sterling Pond as in Figure 13 below.

In Figure 14, we display our biomass data as a mean total biomass of all macrophytes for Sterling Pond using the mean dry weights of all sampled points and we see a typical rapid increase in mid to late summer mass.

By analyzing the individual species data, we show in Figure 15 that the rapid increase in 2002 is due to one species Ceratophyllum demersum a native plant species. Seasonal growth variation is obvious with Potamogeton pusillus, a native that starts out slowly and peaks in late June then all but disappears in mid-August. Another native Vallisneria americana starts very slowly and reaches a peak in August, while Elodea canadensis, a native, increases steadily throughout the growing season. Interestingly, growth of two major submersed non-indigenous invasives, Myriophyllum spicatum, and Potamogeton crispus appear to exert little influence on overall plant growth in Sterling Pond (Figures 15, 17).

We believe, and our macrophyte herbivore data suggests, that the aquatic weevil Euhrychiopsis lecontei feeding on M. spicatum and the aquatic moth Acentria ephemerella feeding heavily on both non-indigenous species P. crispus and M. spicatum in Sterling Pond are depressing growth of these two invasives. This is particularly interesting because other members of this project show data indicating the presence in Sterling Pond of large sunfish populations, known predators of M. spicatum herbivores. The moth, Acentria and the weevil, Euhrychiopsis in Figure 16 are important herbivores that occur at high population densities in Sterling Pond.

back to top

---

Zooplankton and phytoplankton

Routine plankton sampling in the embayments has revealed some interesting patterns in these systems. Some of the systems varied considerably from 2001 to 2002, while some showed very similar dynamics between the two years. Juniper Pond went from being characterized by a medium-level average phytoplankton biovolume in 2001 to having the highest average phytoplankton biovolume of the eight embayments in 2002 (a value which almost tripled the 2001 average biovolume). The phytoplankton community changed dramatically as well, being dominated by a large cyanobacterial bloom early in the season (May) in 2001, to being characterized by edibles during the spring phytoplankton bloom and developing into a cyanobacteria-dominated community as the season progressed into late summer in 2002. Embayments such as Floodwood and Sterling, on the other hand, exhibited similar average phytoplankton biovolume and dynamics in both 2001 and 2002, which is interesting as they are two of the bays we would expect would be most driven by chaotic external forcing.

Overall, Juniper, Sterling, and South Sandy had the highest phytoplankton average biovolume in 2002, while Little Sodus and Floodwood were at the other end of the spectrum with the lowest biovolume (Figure 20). Juniper and South Sandy’s large plankton communities were also characterized by the lowest percentage of edible phytoplankton for the eight sites, although Juniper’s plankton were classified as inedible because of their large size, while South Sandy was dominated by cyanobacteria. Juniper pond was the only embayment to be dominated by copepods, and this could be related to the fact that Juniper is home to the largest average phytoplankton cells of the eight bays. While large cells would clog or not fit in the feeding apparatus of filtering zooplankton like Daphnia, copepods may be better able to utilize these food sources by breaking them apart. Sterling’s large community, on the other hand, joined the small communities in Little Sodus and Floodwood as the embayments whose phytoplankton were most edible. This may be a result of reduced grazing pressure in Sterling and Floodwood, which were also characterized by the lowest zooplankton biomass of the eight embayments (Figure 20). This pattern may also be driven by high nutrient inputs through the extensive tributary systems feeding these two embayments.

Blind Sodus and South Sandy, on the other hand, had the highest average 2002 zooplankton biomass, and were the only two embayments that were dominated by slower-growing large grazing cladocerans like Daphnia (Figure 21). Half of the embayments, including those with the higher flushing rates such as Floodwood and Sterling, were dominated by zooplankton species that typically exhibit high relative growth rates, such as rotifers and smaller cladocerans like Bosmina. While nutrient-rich and slow-flushing systems like Blind Sodus, South Sandy, and Juniper can support the longer life span requirements of larger crustaceans like Daphnia and copepods, the dominance of faster-growing species in the other embayments may be a response to the less-stable environment found in these more rapidly-flushed systems.

Even within an embayment, our preliminary data indicate that different regions of Sterling Pond can support different plankton communities, and these differences may be related to variable flow rates. Early on in the season, the more protected deep-lobe site harbors three-times the zooplankton abundance (and twice the biomass) of the other sites, but as the macrophytes fill in the basin and channel flow to all of the sites but the center site, the center site takes over as the most abundant in zooplankton (Figure 22). The relative consistency of the center site’s macrophyte-surrounded environment may also help to explain why we routinely collected a greater percentage of large-bodied cladocerans at this location, versus the higher-flow rotifer and Bosmina-dominated sites near the mouth of the tributary and the channel connection with Lake Ontario.

These routine sampling patterns suggest that internal dynamics in our embayment systems must be interpreted in the context of external forcing. Data from the three different upwelling events that we sampled in 2002, 2003, and 2004 show that the extent to which these events impacted embayment plankton dynamics depended on embayment morphometry, the duration of the event, and the amount of associated precipitation (forcing from upstream). Upwelling-driven water exchange between Lake Ontario and the embayments can alter late-summer hypolimnionic anoxia, impact water chemistry, and introduce novel plankton species. Preliminary analysis of the long-lasting upwelling event in 2002 shows that extended upwelling in Lake Ontario can fill Little Sodus Bay from the bottom-up, pushing the nutrient-rich hypolimnetic embayment water up into the epilimnion. These events replace the nutrient-depleted photic zone with biologically limiting nutrients such as phosphorus and nitrogen. While the plankton community initially decreases in total abundance, preliminary data suggest that the plankton community positively responds to this late-summer pulse in nutrients.

Short-term, pulsing upwelling events like that of 2003 do not strongly impact Little Sodus, but do replace open Sterling embayment water with nutrient-poor Lake Ontario upwelling water. These events also introduce novel Lake Ontario plankton species to the Sterling Pond ecosystem, some of which remain in the embayment for weeks after the event (Figure 23). One such species is the invasive predatory invertebrate, Cercopagis pengoi, which represents not only a new organism, but also a novel link on the food chain for a system that has no predatory invertebrates. Preliminary results from our set of mesocosm experiments suggest that this and other Lake Ontario species, such as Polyphemus pediculus (which is commonly found in systems like Sterling) are not able to establish successful populations once introduced into Sterling because of the dense macrophyte and fish density in this embayment.

Finally, upwelling events that co-occur with heavy rainfall events will impact Little Sodus as usual--in the case of the medium-duration event in 2004, upwelling Lake Ontario water began to fill the deepest parts of Little Sodus, breaking the anoxia at the sediment-water interface and releasing soil-bound nutrients into the water column. In Sterling however, the large tributary will dominate the external forcing during these events, flushing the embayment out into Lake Ontario and inhibiting the upwelling Lake Ontario water from entering the embayment. Juniper Pond and Floodwood showed a similar differential response to a precipitation event in 2001. Floodwood, which has a large tributary system like Sterling, was completely washed out for almost three weeks in response to an autumn rain event (9/25), while the plankton communities in Juniper Pond, with its very small watershed (like Little Sodus’), increased in abundance in response to the rain event (Figures 24 and 25).

back to top

---

Water chemistry

The Role of Land Use in Regulating Chemical Inputs

The land cover of the sub-watersheds of Lake Ontario embayments is varied, but largely a mixture of forest and agricultural lands. Tributaries draining largely agricultural lands exhibited distinct seasonal patterns, particularly for nutrients. Nitrate and total nitrogen (TN) concentrations were generally low during the summer growing season, concentrations increased markedly during fall and decreased during winter and spring. Total phosphorus (TP), dissolved organic carbon (DOC) and major ion (Na+, K+, Ca2+, Mg2+, Cl-, SO4 2-) concentrations showed variation with change in seasonal flow. We also observed distinct spatial patterns in tributary solute concentrations that closely corresponded with land cover. Solute concentrations increased markedly with increases in the percentage of the watershed occurring as agricultural land. Such a pattern has been commonly observed for nutrients (e.g., TP, TN, NO3-, DOC), but this relationship was also strongly evident for most non-nutrient solutes (e.g., F-, SO4 2-). These observations suggest that agricultural activities mobilize solutes and greatly enhance transport across the temperate landscape. This aspect of the study is detailed in Chen and Driscoll (in review).

Sources of Water to Lake Ontario Embayments.

In this investigation we also examined how linkages between upland watersheds and Lake Ontario control the structure and function of interconnected embayment ecosystems. We hypothesized that the land-cover of the adjacent watersheds and the hydraulic residence time and connectiveness to Lake Ontario of the embayments are critical attributes regulating the chemical composition of embayment waters. We estimated relative changes in the mixing of tributary, direct precipitation inputs and Lake Ontario waters over the study period based on monthly water samples and a three component mixing analysis of a conservative solute (i.e., fluoride). Based on these mixing relationships, we compared the expected concentrations of non-conservative solutes with actual values to quantify the production/loss of these materials within embayments. This aspect of the research is detailed in Chen et al. (in prep.)

Water Chemistry Patterns

Some of the more important physical and water chemistry characteristics of Lake Ontario embayments are summarized in Table 2. Note that there is a range of water chemistry characteristics in these ecosystems. We believe that there are important controllers of embayment water chemistry. First, land use is an important controller. The watersheds draining into these ecosystems are largely a mixture of forest and agricultural land use types. There are contributions of wetlands and urban lands in some watersheds. The greater the watershed that occurs as agricultural lands, the greater the nutrient flux from the watershed to the embayments. The other major factors are the hydraulic residence time and connection with Lake Ontario. Some embayments (e.g., Sterling, Floodwood) have large watershed area to embayment area and therefore short hydraulic residence time. These sites strongly reflect inputs and the water quality draining from the watershed. In contrast, sites with moderate hydraulic residence times reflect some mixing with Lake Ontario, particularly those with strong connections with the lake (e.g., Little Sodus Bay). Note these conditions vary through the year. During the summer when tributary flow is low, mixing with Lake Ontario is maximized.

Embayment element mass balances exhibit a range of patterns. Some solutes are characterized by nearly conservative behavior (e.g., F-). Embayments are net sinks (e.g., NO3-, SO4 2-) or net sources (e.g., dissolved organic N) with respect to some solutes. Still other elements and solutes show mixed patterns, with some embayments acting as net sources while other act as net sinks (e.g., total P, DOC). These mixed patterns may reflect direct or near-shore inputs that are not quantified by inputs from the watershed tributaries or Lake Ontario. Understanding the linkage between water quality of embayments and the biomass and distribution of aquatic biota is an important future research activity.

back to top

---

Models of biological community dynamics

Plankton are often regarded as passive tracers, completely at the mercy of water flow. However, field studies have demonstrated plankton transport against the mean flow. These findings suggest that factors besides advective transport, such as migration behavior, may dominate at smaller scales. Our work examined whether a combination of plankton vertical migration behavior, and exchange flow induced by the temperature difference between Lake Ontario and embayment water could allow plankton to avoid quick washout into Lake Ontario. We performed computational studies using a three dimensional finite-difference computational hydrodynamic code for shallow water flow (SI3D), coupled with a post-processing individual-based model for plankton movement based on computed water velocities and various models for vertical migration behaviors by the plankton, to track the movement of individual plankters and determine how long they were able to persist within the embayment (Figure 26). We also use a global sensitivity analysis, based on nonparametric regression, to determine the most important factors in the fate of the plankton.

In a channel free of macrophytes, the mean flow rate through the embayment by itself accounts for almost a third of the variance in biological residence time, while the initial size of the zooplankton cloud has the next largest effect. The independent contribution of migration type to biological residence time is relatively small (6% of the variance), but interactions of migration type with other parameters were significant and accounted for an additional 46% of the variance in mean retention time. In contrast, for a macrophyte-filled channel, plankton behavior had very little effect and retention time was almost completely determined by the mean flow rate and interactions between flow rate and other parameters.

We also investigated a simple time-averaged approximation to the individualbased plankton model. The approximation replaces moment-by-moment simulation with a partial differential equation model using the average velocity over a 24-hour period, computed using the output of the hydrodynamic model and a deterministic model of plankton migration in which plankton have perfect control over their vertical position at all times. The predicted mean residence time is then determined by the average downstream velocity (averaged over a 24 hour period) and the distance between the release point and washout points (i.e. time=distance/rate). The approximation was successful for macrophyte-free channels, but not very accurate with macrophyte-filled channels.

This research makes two significant contributions to ecology:

1. It contributes to the resolution of the longstanding 'drift paradox' – the ability of plankton to persist in the face of rapidly moving currents – by showing clearly the importance of variability in flow regimes. Where previous work has assumed completely homogenous flow (downstream velocity constant in space and time), the use of a computational hydrodynamic model allowed us to show the importance of vertical inhomogeneity in flow rates resulting from temperature gradients and/or macrophytes. Specifically, with homogeneous flow assumed, there would be no impact whatsoever of plankton vertical migration behavior, but with realistic hydrodynamics plankton behavior was nearly as important as mean flow rate in determining plankton persistence time.

2. More generally, it contributes to the problem of "scaling up" from individual behaviors (which are the things we can measure) to population-level models that are amenable to analytic study – so we can try to understand what a model is doing, rather than just getting computational results and trying to interpret them. It is encouraging that a highly simplified scaling up based on very coarse time-averaging was so successful in the no-macrophyte case, and informative that the same approximation failed for high macrophyte density (it will be more informative when we figure out why the approximation failed, but that will have to come later).

back to top

---

Watershed modeling

This research applies and evaluates two different models for predicting the runoff from eight watersheds adjacent to Lake Ontario. These models are called the HP Manning model and the Generalized Watershed Loading Functions (GWLF) model. This assessment is accomplished by first testing the behavior of both non-point source (NPS) models in two hypothetical cases: impervious watershed and another pervious watershed typical of a deciduous forest. Sensitivity analysis of the critical parameters was performed. Then these models were applied to the Lake Ontario watersheds. The outputs of the simulations include streamflows, sediment yield, total dissolved phosphorus and total nitrogen. The simulated outputs of both models are compared with the observed data using three performance model statistics. The statistics shows low correlations between the simulated output and the observed data. However, the correlations were relatively high between the outputs of both simulation models. The HP-Manning streamflow and total nitrogen output was relatively lower than that of the GWLF output.

Additional work was performed to model Lake Ontario watersheds using an integrated modeling system. Jonghwa Ham, working with Pete Loucks in the Department of Civil and Environmental Engineering, worked on a system using BASINS and HSPF to evaluate the effects of watershed management on embayment water quality and ecosystem health. An Integrated modeling system, called BASINS, was used to comprehensively assess the watershed and receiving waterbody. In BASINS, GIS data such as landuse, landcover, dem and boundaries, monitoring data, and sources of pollution were used. Using these data and assessment tools, we can simulate watershed and water quality. HSPF was also used to model flow rate in 18 subwatersheds. Future research will involve: 1) Stream water quality modeling by HSPF using point source pollutant loading data and stream water quality monitoring data; 2) Linking water quality model to analyze embayment water quality using BASINS and HSPF and 3) Evaluating the effects of watershed management (including changing land uses and pollutant loading, constructing WWTP) on embayment water quality and ecosystem.

back to top

---

Human Dimension of Biocomplexity

This project aims to complete an incomplete circuit in the thinking about the relationship between humans and the natural environment. Most ecological literature on the subject, especially on the relationship between land use and water quality, focuses on how urbanization and suburbanization degrade water quality. Very little, by contrast, is known about whether, how, and why land-use patterns change in response to water quality. We do, however, have some important findings already from this research and are poised to learn more.

We know that—consistent with past findings and conventional wisdom about amenities and property values—land along the shoreline is more valuable than upland sites. This means that the shoreline attracts many users at high densities, and that, absent some institutional response (public or private sector), the amenity will be degraded. In particular, water quality will be threatened by runoff and untreated sewage, because most new development has occurred in areas without public sewage treatment. Some such responses might accommodate increased density through the construction of public infrastructure for water and sewage treatment and the development of more sophisticated site-planning techniques; others might discourage new building through public or private open-space purchase or public prohibitions of high-density development (zoning).

A complex set of factors has, however, precluded the full development of either growth-accommodating or growth-discouraging measures in the embayment area (Figure 27). First and foremost, land development has been very slow, although steady, and population growth has been negligible. Hence, development has not appeared yet to pose the primary threat to coastal resources, even though septic systems near the embayments are likely an important contributor to poor water quality where it does exist. Second, we also know that the area is politically quite conservative in the sense that its residents have traditionally respected private property rights and resisted paying higher property taxes. This social setting makes it unlikely that local residents will accept measures either to accommodate growth (which require capital investment) or to limit it (which require regulation). Indeed, in Huron, we found that local elections turned on proposals to impose modest regulations on septic systems. Third, the area is very politically fragmented; it has many small local governments, each with control over its own land use and often its own infrastructure systems; the small, mainly rural local governments of the embayment area are generally too small to have the resources to plan for and regulate growth, even if there were a political constituency that supported planning and regulation. Since New York State has few or no controls, incentives, or mandates from higher levels of government (counties or state) and very limited technical assistance for lower units of government, the only “signal” that local governments pay attention to is from the voter-taxpayers within their boundaries.

How, then, do local residents respond to the environment, in this slowly growing, conservative, home-rule environment? Faced with real or hypothetical threats, what will they do themselves, and what actions by others will they support? An ongoing survey will help answer these questions as it reveals people’s knowledge about water quality in the embayments; their priorities for protection; and their responses to individual, collective private, and collective public responses to environmental threats. We will learn whether perceptions of poor environmental quality match scientific knowledge about environmental degradation; whether “user” priorities for water are consistent with the maintenance of high water quality (we know already that people prefer to modify the shoreline and remove aquatic vegetation, both of which can degrade the biological integrity of the embayments); and whether residents’ preferred institutional responses are appropriate for addressing the problems they perceive. We will also learn what kinds of people are most inclined to perceive and respond to environmental threats; based on differences in age, gender, household status, length of residence, and workplace location, we will advance our simulations of how environmental organizations appear, evolve, and respond to environmental threats.

In short, there is substantial complexity in the chain of events that lead from amenity, to settlement, to environmental change, to perceptions about change, to action. And once action is taken, there are usually complex feedbacks to both the environment and to the composition of local communities that result in instabilities in the coupled social-environmental system. While present in all social-environmental systems, this complexity is even more evident in amenity-rich “consumption” regions like our embayment communities, whose economies no longer depend on local resource extraction but rather rely on the metropolitan areas and small cities outside the watersheds entirely.

back to top

---

Paleoecology

Lake Ontario’s Dynamic Coast: Analyzing Ecosystem History for Sustaining Environmental Health This is a project to use paleoecological methods to study long-term changes in the coastal embayments along the south and east shores of Lake Ontario, specifically the effects of changes in land-use and water-level regulation. The project is a collaboration between scientists at Cornell University, Syracuse University, SUNY ESF, and Columbia University. Wolin at Cleveland State University and John Peck at University of Akron joined after the project was underway.

Three embayments were chosen for study that are also part of the NSF Biocomplexity project: South Sandy Pond, Juniper Pond and Little Sodus Bay. 1-mlong sediment cores were collected at each embayment in August 2003 and analyzed for (a) 210-Pb dates indicating that the bottom of each core goes back at least 300 years (pre-European settlement)., (b) nitrogen, carbon and mercury concentrations as indicators of embayment trophic conditions and pollutant inputs, (c) diatom microfossils as indicators of both environmental conditions and connection to L. Ontario, (d) paleomagnetism as indicator of changes sediment origins, (e) pollen and spore stratigraphy as indicators of broader changes in land use, and (f) changes in wetland vegetation. There is also an outreach component to this project with local stake-holders contributing knowledge about pond history and management concerns, while also being end users of the management related conclusions.

The sediment records show patterns distinctive to each embayment in terms of overall sediment accumulation rates and changes in those rates over time. Elemental (C&N) and diatom stratigraphy appear to show periods, especially in South Sandy Pond (SSP) when the embayments were exchanging water directly with L. Ontario and periods when such exchange was much reduced. The paleomagnetism data for SSP support this interpretation. Ongoing studies of the core from Juniper Pond (JP) indicate that the forests surrounding this region prior to human impact (1800 AD) were represented by more abundant oak, pine, hemlock and beech than today. Major increases in ragweed and grass beginning at 70 cm (around 1750 AD) indicate human impact on the region, along with the increases in weedy species. Cattail also increases in the wetland with the advent of Europeans. Charcoal, indicative of forest clearing for farmland, is highest between 70 and 55 cm (1750 to 1850), then fluctuates and drops to 0 in the upper 10 cm. Pollen stratigraphy from Juniper Pond shows some very interesting correlations in wet periods with the diatom data. Research that remains to be completed includes analyses of upland vegetation for JP and SSP and terrestrial vegetation for SSP vegetation. The outreach component of the project has begun to show an interesting pattern among residents living near the embayments with people genuinely concerned about water quality but likely to assign causation of problems to others or to the condition of Lake Ontario. We are using the record of environmental change to explain current conditions in the embayments and to focus discussion on how past change provides insight for future management. We are working with citizens and policy makers to understand what our science tells them in terms of dispelling public misconceptions, determining public expectations, and defining management trade-offs. Issues that are clearly critical and to which our data can contribute are background for understanding the history of lake-level fluctuations and the impacts of development on eutrophication.

back to top