Search

Macroinvertebrate and fish metrics are an essential tool in freshwater biomonitoring. Yet, many well-known stream metrics do not perform well as indicators of anthropogenic stress in lakes and wetlands, and there is a need to better understand the role of potential confounding factors such as latitude or environmental variables acting at regional scales. We attempt to untangle the relationships among metrics and watershed stressors while accounting for potential confounding factors using hierarchical partitioning, and test metric consistency across time, focusing on the Laurentian Great Lakes wetland monitoring programs conducted 10 years apart. Our results show that many frequently used metrics have high temporal variability and are significantly affected by spatial factors, most notably richness-type metrics used in numerous regional monitoring studies. Only a few metrics (invasive fish species richness and relative abundance, mayfly family richness and the relative abundance of Ephemeroptera, Trichoptera, Sphaeriidae and Odonata) had slightly better and more consistent correlations with watershed land use. We suggest that explicit consideration of confounding factors is essential in the context of large-scale monitoring programs, and focus on the less general (such as habitat-specific) metrics may be a more promising approach.

We describe development anthropogenic stress indices for coastal margins of the Laurentian Great Lakes basin. Indices were derived based on the response of species assemblages to watershed-scale stress from agriculture and urbanization. Metrics were calculated for five groups of wetland biota: diatoms, wetland vegetation, aquatic invertebrates, fishes, and birds. Previously published community change points of these assemblages were used to classify each watershed as ‘least-disturbed’, ‘at-risk’, or ‘degraded’ based on community response to these stressors. The end products of this work are an on-line map utility and downloadable data that characterize the degree of agricultural land use and development in all watersheds of the US and Canadian Great Lakes basin. Discrepancies between the observed biological condition and putative anthropogenic stress can be used to determine if a site is more degraded than predicted based on watershed characteristics, or if remediation efforts are having beneficial impacts on site condition. This study provides a landscape-scale evaluation of wetland condition that is a critical first step for multi-scale assessments to help prioritize conservation or restoration efforts.

Stressor gradients and spatial narratives of the St. Louis River Estuary, a joint Minnesota and Wisconsin Sea Grant study, connected aquatic science research with spatially-explicit stories of local resource issues and place-based geo-quests to enhance spatial awareness and stewardship of the estuary. The goal of this paper is to report and reflect on an integrated study that combined environmental humanities and technology with aquatic science in a spatial context. Our study was organized into three objectives around research, outreach, and evaluation. First, we summarized anthropogenic stressors within high resolution watersheds and linked the watershed stress estimates to aquatic habitats within the estuary. Second, we designed tools to deliver place-based environmental science and technology to targeted users to increase awareness, learning, and the potential for long-term stewardship. And third, we evaluated the responses of targeted end users to their interaction with the project’s integrated science and innovative delivery methods. Finally, central to all three objectives, we created a dynamic website to facilitate regional to national coastal outreach and education goals. We found significant correlations between the stressor index and the water quality and biotic data, along with variability attributed to landscape elements. Connecting this science with the place-based experiences we collected is expected to expand the scope and reach of state, bi-national and non-governmental outreach programs. The project also has direct applications to classroom science education. Developing this integrated project contributed to our shared knowledge of environmental and cultural aspects of the estuary for place-based education, and offers several lessons for future work of this nature.

The St. Louis River drains an area of 9412 km 2 into the western arm of Lake Superior. The river's lower section, including a 48.5 km 2 estuary, was designated as a Great Lakes Area of Concern due to degradation from industrial activities. Part of the estuary is occupied by the largest port in the Great Lakes. A GIS-based stressor index was previously developed to characterize anthropogenic stress within the watershed. The components of the stressor index were road density, point-source pollution permit density, population density, and percent agricultural and developed land. Water quality sampling was conducted at 27 sites in the estuary in tributaries and associated nearshore areas during multiple flow regimes in 2010–2011. Additional data were analyzed from 34 upper watershed sites sampled in 2009–2010. Stressor scores were significantly (p 2 − /NO 3 − -N, dissolved oxygen saturation, pH, specific electrical conductivity, chloride, sulfate, and E. coli in the upper watershed. In the estuary, the index was significantly and positively correlated with NO 2 − /NO 3 − -N, NH 4 + -N, and chloride at multiple flow regime and location combinations. Soil K factor (an erosivity index from recent NRCS SSURGO soil surveys) was found to have stronger relationships with sediment related parameters than the stressor gradient. Although originally designed to help stratify sampling across a gradient of landscape stress and identify reference areas for restoration projects, the stressor index was shown to have substantial predictive power for multiple water quality parameters.

Context Green infrastructure may improve water quality and mitigate flooding in forest-urban watersheds, but reliably quantifying all benefits is challenging because most land cover maps depend on moderate- to low-resolution data. Complex and spatially heterogeneous landscapes that typify forest-urban watersheds are not fully represented with these types of data. Hence important questions concerning how green infrastructure influences water quality and quantity at different spatial scales remain unanswered. Objectives Demonstrate the feasibility of creating novel high-resolution land cover maps across entire watersheds and highlight deficiencies of standard land cover products. Methods We used object-based image analysis (OBIA) to create high-resolution (0.5 m) land cover maps and detect tree canopy overlapping impervious surfaces for a representative forest-urban watershed in Duluth, MN, USA. Unbiased estimates of accuracy and area were calculated and compared with similar metrics for the 30-m National Land Cover Database (NLCD). Results Mapping accuracies for the high-resolution land cover and canopy overlap maps were ~90 %. Error-adjusted estimates of area indicated that impervious surfaces comprised ~21 % of the watershed, tree canopy overlapped ~10 % of impervious surfaces, and that three high-resolution land cover classes differed significantly from similar NLCD classes. Conclusions OBIA can efficiently generate high-resolution land cover products of entire watersheds that are appropriate for research and inclusion in the decision-making process of managers. Metrics derived from these products will likely differ from standard land cover maps and may produce new insights, especially when considering the unprecedented opportunity to evaluate fine-scale spatial heterogeneity across watersheds.

Total suspended solids (TSS) and total phosphorus (TP) have been shown to be strongly correlated with turbidity in watersheds. High-frequency in situ turbidity can provide estimates of these potential pollutants over a wide range of hydrologic conditions. Concentrations and loads were estimated in four western Lake Superior trout streams from 2005 to 2010 using regression models relating continuous turbidity data to grab sample measures of TSS and TP during differing flow regimes. TSS loads estimated using the turbidity surrogate were compared with those made using FLUX software, a standard assessment technique based on discharge and grab sampling for TSS. More traditional rating curve methodology was not suitable because of the high variability in the particulates vs. discharge relationship. Stream-specific turbidity and TSS data were strongly correlated (r2 = 0.5 to 0.8; p

Non-native earthworm invasions in north-temperate North America cause substantial adverse effects to hardwood forest ecosystems. Quantification of invasions is necessary for understanding impacts and identifying remnant earthworm-free areas, but existing sampling techniques are effort-intensive and/or environmentally damaging. We: (1) developed and applied a protocol that allows rapid classification of earthworm invasion into five stages based primarily on visual assessment of the forest floor, (2) sampled earthworms to test whether the protocol's stages can predict invasion by different species, and (3) assessed relationships between individual forest floor characteristics and presence of different earthworm species. Based on differences in biomass among points assigned to different stages, the 5-stage classification protocol accurately identified the onset of invasion by Lumbricus rubellus and Lumbricus terrestris, the species of greatest management concern in the northern Midwest. Except for middens as a predictor of L. terrestris presence, no forest floor variable was useful by itself for assessing invasions. The 5-stage protocol provides an efficient approach for assessing earthworm invasions in hardwood forests of the U.S. northern Midwest, can be implemented with minimal training, and serves as a blueprint for similar protocols in other regions experiencing earthworm invasions.

Many streams along the Minnesota coast of Lake Superior have been listed as impaired from either high turbidity or high fish mercury concentrations or both. Both turbidity and total mercury have been shown to be strongly correlated to total suspended sediment in many disturbed watersheds. Turbidity and total mercury concentrations and loads were estimated in four western Lake Superior watersheds from 2005–2006 using automated in-stream turbidity measurements. Regression models were developed relating this near-continuous turbidity data to grab sample measures of mercury during differing flow regimes. Total mercury values ranged from 1 to 28 ng l−1 throughout the open water season and showed a close relationship to total suspended sediment (r2 = 0.85, n = 23; p < 0.001) and a less robust but still significant relationship with turbidity (r2 = 0.40, n = 34; p < 0.001) for all four streams. Mercury loads to Lake Superior were estimated to range from 8 to 97 g yr−1 with watershed yields ranging from 0.5 to 4.3 μg m−2 yr−1. Continuous turbidity monitoring appears to be a reasonable surrogate for both suspended sediment and total mercury concentration, providing information when manual sample collection is cost-prohibitive or logistically difficult, and across a wide range of flows.

Biological indicators can be used both to estimate ecological condition and to suggest plausible causes of ecosystem degradation across the U.S. Great Lakes coastal region. Here we use data on breeding bird, diatom, fish, invertebrate, and wetland plant communities to develop robust indicators of ecological condition of the U.S. Lake Superior coastal zone. Sites were selected as part of a larger, stratified random design for the entire U.S. Great Lakes coastal region, covering gradients of anthropogenic stress defined by over 200 stressor variables (e.g. agriculture, altered land cover, human populations, and point source pollution). A total of 89 locations in Lake Superior were sampled between 2001 and 2004 including 31 sites for stable isotope analysis of benthic macroinvertebrates, 62 sites for birds, 35 for diatoms, 32 for fish and macroinvertebrates, and 26 for wetland vegetation. A relationship between watershed disturbance metrics and 15N levels in coastal macroinvertebrates confirmed that watershed-based stressor gradients are expressed across Lake Superior's coastal ecosystems, increasing confidence in ascribing causes of biological responses to some landscape activities. Several landscape metrics in particular—agriculture, urbanization, human population density, and road density—strongly influenced the responses of indicator species assemblages. Conditions were generally good in Lake Superior, but in some areas watershed stressors produced degraded conditions that were similar to those in the southern and eastern U.S. Great Lakes. The following indicators were developed based on biotic responses to stress in Lake Superior in the context of all the Great Lakes: (1) an index of ecological condition for breeding bird communities, (2) diatom-based nutrient and solids indicators, (3) fish and macroinvertebrate indicators for coastal wetlands, and (4) a non-metric multidimensional scaling for wetland plants corresponding to a cumulative stress index. These biotic measures serve as useful indicators of the ecological condition of the Lake Superior coast; collectively, they provide a baseline assessment of selected biological conditions for the U.S. Lake Superior coastal region and prescribe a means to detect change over time.

Quantifying gradients of anthropogenic stress can inform the development of sample designs, provide an important covariate in modeling relationships of response variables, identify reference and highly-disturbed sites, and provide a baseline and guidance to restoration and remediation efforts. We describe development of SumRel, a composite index of anthropogenic stress, for the U.S. and Canadian Lake Superior basin. Key elements of the project include development of high-resolution watersheds throughout the basin, summarization of the major point and non-point stressors within these watersheds, and creation of tools for scaling the watersheds and stressor summaries. SumRel was calculated at two spatial scales: for high resolution subcatchments within the Lake Superior basin (mean watershed area = 93 ha) and for coastal watersheds of Lake Superior. An assessment of subcatchments within Minnesota's St. Louis River watershed showed a correlation between the degree of disturbance, as indicated by SumRel, and impaired water quality, as evidenced by in-stream conductivity. These data and tools allow identification and visualization of reference and highly-disturbed sites at multiple spatial scales, providing decision support for individual agency and binational monitoring, assessment and restoration initiatives across the Lake Superior basin.

We summarized landscape approaches used in the study of freshwater ecosystems, updated recent literature reviews on interactions between terrestrial and lotic ecosystems, and discussed the influence of J-NABS on developments in the field. We focused primarily on studies of freshwater ecosystems done at or above the catchment scale. Special issues of J-NABS and other journals have advanced our understanding of the effects of spatially distributed characteristics and phenomena on aquatic ecosystems. Topics that have been well covered in J-NABS include use of classification to predict biotic assemblages and impacts of human disturbance (especially urbanization) on stream structure and function. Early work focused on correlative relationships between landscape variables and various biotic components of stream systems, whereas later studies addressed causal linkages between landscape and biota, including landscape effects on hydrology, habitat at various spatial scales, and ecosystem processes. At larg...

We developed ecological indicators of stream macroinvertebrates in two regions of the Midwestern USA dominated by row-crop agriculture. Indicators were identified in a hierarchical fashion. Reach-scale variables related to macroinvertebrate attributes were first identified, and then catchment-scale variables related to those reach-scale variables were identified. Reach-scale indicators common to both regions were % fine sediments, number of habitats, and width:depth ratio. SD of elevation and % commercial land use were selected as catchment-scale indicators in both regions. Our analyses revealed a multi-scale mechanistic relationship between macroinvertebrate attributes associated with degraded conditions (i.e., fewer taxa of Plecoptera and Trichoptera, and a higher proportion of chironomids, burrowers, and depositional taxa) and % fine sediments in stream reaches, which, in turn, was negatively related to catchment characteristics (i.e., SD of elevation) in one region. Understanding how natural variables such as topography influence channel shape and within-channel structure can help guide management options and expectations for different regions. We suggest that developing multi-scale indicators in a mechanistic fashion will be more effective than developing indicators at only one spatial scale for protecting and restoring stream structure and function.

We used General Land Office survey data (1860–1890) and interpreted aerial photography from the 1930s, 1970s, and 1990s to quantify forest disturbance frequency and spatial patterns for four time periods in the Mixed Forest Province of Minnesota. The study region included eight subsections within the Mixed Forest Province of Minnesota’s Ecological Classification System. Presettlement disturbance and spatial pattern estimates varied across the eight subsections indicating a strong relationship to soil and landform characteristics. Land surveyors primarily recorded higher severity disturbances that resulted in significant tree mortality. The 1900–1940 era was characterized by a short-term increase in fire frequency that was relatively uniform across the study region, in contrast to the variability of the presettlement (1860–1890) landscape. In the postsettlement period (1940–1995), timber harvest replaced fire as the dominant disturbance factor. Similar management practices among subsections created similar harvest rates throughout the study region. These management practices imposed a more homogeneous pattern dominated by small (10–25 ha) patches. Management practices now have a greater influence than natural processes in the generation of landscape pattern in the Mixed Forest Province of Minnesota. Information on presettlement forest conditions and subsequent changes can be used by land managers to restore spatial pattern variability in managed forest landscapes.

Developing indicators of ecosystem condition is a priority in the Great Lakes, but little is known about appropriate spatial scales to characterize disturbance or response for most indicators. We surveyed birds, fish, amphibians, aquatic macroinvertebrates, wetland vegetation, and diatoms at 276 coastal wetland locations throughout the U.S. Great Lakes coastal region during 2002-2004. We assessed the responsiveness of 66 candidate indicators to human disturbance (agriculture, urban devel- opment, and point source contaminants) characterized at multiple spatial scales (100, 500, 1,000, and 5,000 m buffers and whole watersheds) using classification and regression tree analysis (CART). Non- stressor covariables (lake, ecosection, watershed, and wetland area) accounted for a greater proportion of variance than disturbance variables. Row-crop agriculture and urban development, especially at

Land use and geology are two important extrinsic factors regulating the structure and function of stream ecosystems. The interactions among these two landscape-scale factors on streams are, however, poorly understood. To determine the effects of these factors on stream flow, sediment, and nutrients, we analyzed 72 ungaged, agricultural watersheds in Minnesota and Michigan using the hydrologic model SWAT (the Soil Water Assessment Tool). The watersheds differed in surficial geology (landform) and land use, but were of similar size, with streams ranging from 2nd to 3rd order. SWAT was developed for use on ungaged basins, but to improve the outputs we used US Geological Survey discharge data from sites near our study watersheds for calibration. We found seasonal and annual differences in flow and nutrient and sediment loading across different land forms and land use types. Watersheds with greater amounts of row-crop agriculture and watersheds dominated by morainal landforms were associated with more sediment and nutrients and greater flow volumes and flashiness. Multivariate analyses identified at least nine landscape variables which were related to nutrients, sediment, and flow, although the responses varied between Minnesota and Michigan. Results also indicated the possibility of a threshold effect for row crop agricultural. Increases in this land use had little additional effect on nutrients or flow when percent row crop exceeds the threshold value. At moderate to high levels of row crop agriculture, watersheds appeared to show greater sensitivity to differences in landform.

We investigated impacts of harvesting on soil disturbance and vegetation in the aspen cover type of northern Minnesota, United States. The soil disturbance (resistance to penetration) and understory vegetation were characterized for 19 sites on five 60-m 2 plots placed along a disturbance gradient: landings (high harvesting traffic), skid trails (inter- mediate harvesting traffic), and areas off skid trails (low to no harvesting traffic). Penetration levels were quite variable, but they still indicated that within-site responses to disturbance patterns created by clear-cut harvesting were not uniform. In general, soil disturbance and understory species composition within landings were similar to those with skid trails. The soil disturbance and vegetation composition of these two levels differed from those of the low-disturbance plots (off skid trails), indicating that removing trees alone did not affect vegetation composition as much as creating an es- tablished skid trail, regardless of harvest timing. However, sites with more variable species composition (win- ter-harvested sites) and lower disturbance levels were less altered than sites with likely lower initial diversity (summer-harvested sites). The results suggest that it is important for recovery of understory plant communities to not only limit the amount and level of disturbances but also consider the spatial layout of harvesting, thus maintaining a spatially connected network of remnant forest patches large enough to contain interior forest species. Resume : Les auteurs etudie l'impact des activites de recolte sur la perturbation du sol et la vegetation dans les forets de peuplier faux-tremble du nord du Minnesota aux Etats-Unis. La perturbation du sol (resistance a la penetration) et la vegetation de sous-bois ont ete caracterisees pour 19 sites dans cinq places-echantillons de 60 m 2 etablies de long d'un

To improve source–sink relationship based carbon-allocation models, the basic proportional model was extended to account for a well-known effect of individual source to sink distances: among different sinks of similar characteristics, the more distant from the source, the lower the allocation coefficient. This was achieved through multiplication of the sink strength value by a coefficient that is proportional to a decreasing, simple function of distance, f; the power form was chosen for both simplicity and theoretical reasons. The resulting model was parameterized and evaluated on the empirical allocation matrix of the ECOPHYS model, after grouping together several individual, small sinks of similar nature and close location to remove any phyllotaxy-related bias. Both goodness of fit and predictive value were significantly improved compared with the basic proportional model (f = constant). The f-extended model yielded even better results if segments of different nature or age on the source to sink pathway were assigned different weights in the expression of distance, whereas the default expression of f, with an exponent of –1 and no additive constant, was optimal with no further parameter required. Thus, only 7 parameters (3 for pathway segment weights and 4 for sink strength values) were sufficient to retrieve the original 68 independent experimental allocation coefficients with a reasonable degree of accuracy. Pathway segment weights likely reflect both intrinsic transport pathway properties and situation within the plant architecture; this is discussed in relation to the possibilities of generalization and practical use of the model.

The interaction between physical environment and land ownership in creating spatial heterogeneity was studied in largely forested landscapes of northern Wisconsin, USA. A stratified random approach was used in which 2500-ha plots representing two ownerships (National Forest and private non-industrial) were located within two regional ecosystems (extremely well-drained outwash sands and moderately well-drained moraines). Sixteen plots were established, four within each combination of ownership and ecosystem, and the land cover on the plots was classified from aerial photographs using a modified form of the Anderson (U.S. Geological Survey) land use and land cover classification system. Upland deciduous forests dominated by northern hardwoods were common on the moraines for both ownerships. On the outwash, the National Forest was dominated by pine plantations, upland deciduous forests, and upland regenerating forests (as defined by

Biological attributes of ecosystems often change nonlinearly as a function of anthropogenic and natural stress. Plant and animal communities may exhibit zones of change along a stressor gradient that are dispropor- tionate relative to the incremental change in the stressor. The ability to predict such transitions is essential for effective management intervention because they may indicate irreversible changes in ecological processes. De- spite the importance of recognizing transition zones along a stressor gradient, few, if any, investigators have examined these responses across multiple taxa, and no community threshold studies have been reported at large geographic scales. We surveyed benthic macroinvertebrate, fish, bird, diatom, and plant communities in coastal wetlands across a geospatially referenced gradient of anthropogenic stress in the Laurentian Great Lakes. We used Threshold Indicator Taxon Analysis (Baker and King 2010) to analyze each community's response to iden- tify potential zones of disproportionate change in community structure along gradients of major watershed-scale stress: agriculture and urban/suburban development. Our results show surprising congruence in community thresholds among different taxonomic groups, particularly with respect to % developed land in the watershed. We also analyzed uncertainty associated with the community-specific thresholds to understand the ability of different assemblages to predict stress. The high and congruent sensitivity of assemblages to development dem- onstrates that watershed-scale stress has discernible effects on all biological communities, with increasing poten- tial for ecosystem-scale functional changes. These findings have important implications for identifying reference- condition boundaries and for informing management and policy decisions, in particular, for selecting freshwater protected areas.

1. Landscape characteristics of sixty-two subcatchments within the Saginaw Bay Catchment of central Michigan were examined to identify relationships with stream water chemistry. Land use, land cover and elevation were quantified for both entire catchments and the upland–river ecotone (100 m stream buffer strip). Catchment and ecotone data were then empirically compared with stream water chemistry using multivariate and regression analyses. Redundancy analysis was used to partition variance among land use, geology, and the shared influence of land use and geology. 2. Major catchments dominated by rowcrop agriculture had the highest alkalinity, total dissolved solids and nitrate + nitrite concentrations. 3. Strong seasonal differences were observed in total nitrogen and nitrite + nitrate, but not in total phosphorus or suspended solids. Land use and landscape structure factors such as slope and patch density (number of land use patches per km2) accounted for most of the observed variance in summer. 4. In both autumn and summer, landscape factors accounted for much of the observed variation in total dissolved solids and alkalinity. During autumn, geological factors and the shared influence of geology/landscape structure plus land use exerted more influence than did land use alone. 5. Total phosphorus and total suspended solids were much better explained by land use within the stream ecotone in summer than in other seasons. However, total nitrogen, nitrate, orthophosphate and alkalinity were equally well explained by land use within the ecotone and throughout the whole catchment. Only total dissolved solids in summer and ammonium in autumn were explained better by the whole catchment than the ecotone. 6. Our results show that relatively coarse spatial databases can provide useful descriptors of regional water quality.

1. We used catchment and reach-scale physical properties to predict the occurrence of specific species life history and behaviour traits of aquatic insects across fifty-eight catchments in a mixed land use basin. Catchment-scale attributes were derived using a geographical information system (GIS). Logistic regression techniques were used to model the relationships. 2. The reach-scale properties were highly predictive of species traits. Fourteen of the fifteen traits had significant models with concordance values greater than 68%. Cross-sectional area at bank full discharge, % shallow, slow-water habitats, and % fines were the most important variables. 3. Life history and behavioural attributes were best related to reach-scale physical features. This suggests that species traits exhibit strong relationships to local environmental conditions. 4. Catchment-scale variables had fewer significant models with species traits (four of fifteen), however these variables may have direct or indirect influence on reach-scale properties. 5. Catchment features, in particular surficial geology, influence macroinvertebrate assemblages through their control over channel morphology and hydrologic patterns. 6. The effects of land use were masked by geology (i.e. lacustrine clay geology and rowcrop agriculture were correlated), lack of detail in land use data and the aggregation of the species data. 7. These models reflect the coupling of local environmental conditions and the set of adaptations among the local taxa. These observations underscore the idea that habitat plays a major role in organizing stream assemblages. 8. Using these approaches, predictions can be made about the ability of various taxonomic groupings to track environmental change through time, or for projecting the impact of alternative land management scenarios. Identifying fundamental life history and other traits can improve the selection and evaluation of such indicators.

Ecological land classification systems have recently been developed at continental, regional, state, and landscape scales. In most cases, the map units of these systems result from subjectively drawn boundaries, often derived by consensus and with unclear choice and weighting of input data. Such classifications are of variable accuracy and are not reliably repeatable. We combined geographic information systems (GIS) with multivariate statistical analyses to integrate climatic, physiographic, and edaphic databases and produce a classification of regional landscape ecosystems on a 29 340-km2 quadrangle of northwestern Wisconsin. Climatic regions were identified from a high-resolution climatic database consisting of 30-yr mean monthly temperature and precipitation values interpolated over a 1-km2 grid across the study area. Principal component analysis (PCA) coupled with an isodata clustering algorithm was used to identify regions of similar seasonal climatic trends. Maps of Pleistocene geology and major soil morphosequences were sued to identify the major physiographic and soil regions within the landscape. Climatic and physiographic coverages were integrated to identify regional landscape ecosystems, which potentially differ in characteristic forest composition, successional dynamics, potential productivity, and other ecosystem-level processes. Validation analysis indicated strong correspondence between forest cover classes from an independently derived Landsat Thematic Mapper classification and ecological region. The development of more standardized data sets and analytical methods for ecoregional classification provides a basis for sound interpretations of forest management at multiple spatial scales.

A strategy for process modelling of short-rotation Salix coppice plantations is presented. To integrate effectively both large- and small-scale processes, this model must simulate critical processes at scales ranging from the individual leaf to the plantation. This multiscale modelling strategy extrapolates the detail of mechanistic physiological models to the patch and plantation scales. To accomplish this scaling, regional climatic and soil databases are used to generate model inputs; i.e. site and climatic information is passed from the regional scale to drive models operating at the stand or tree level. Key input variables at the individual tree level include genetically determined factors such as leaf morphology and branch architecture, as well as phenological and physiological parameters. Outputs such as carbon fixation, leaf area index, biomass by component, and yields per hectare are then passed back up to address regional questions. This strategy requires a parallel programme of modelling and experimentation. While this strategy focuses on Salix , the approach should also be applicable to species with similar life history attributes, such as Betula and Populus .

Geographic Information Systems (GIS) were used to assess the relationships between land use patterns and the physical habitat and macroinvertebrate fauna of streams within similar sized watersheds. Eleven second or third order watersheds ranging from highly urbanized to heavily forested were selected along Lake Superior's North Shore. Land use patterns within the watersheds were quantified using readily available digital land use/land cover information, with a minimum mapping resolution of 16 ha. Physical habitat features, describing substrate characteristics and stream morphology, were characterized at sample points within each stream. Principle component and correlation analyses were used to identify relationships between macroinvertebrates and stream physical habitat, and between habitat and land use patterns. Substrate characteristics and presence of coarse woody debris were found to have the strongest correlations with macreinvertebrate assemblage richness and composition. Agricultural and urban land use was correlated with substrate characteristics. Algal abundance, associated with macroinvertebrate compositional differences, was correlated with housing density and non-forest land covers. The use of readily available spatial data, even at this relatively coarse scale, provides a means to detect the primary relationships between land use and stream habitat quality; finer-resolution GIS databases are needed to assess more subtle influences, such as those due to riparian conditions.

SUMMARY 1 Patterns of macroinvertebrate community composition were examined in streams within a 40000-km2 catchment in central Michigan, U.S.A., to identify the major environmental gradients influencing community variation. Agriculture and associated clay and sandy soils predominated in much of the region. 2 Eighty macroinvertebrate taxa were collected from stream surveys conducted during May and August 1990. Community composition varied primarily by the proportions of Plecoptera and Ephemeroptera. Benthic communities from the heaviest agricultural zones were most different from those at other sites. 3 Chemical composition among the sites varied most in relation to nutrients (NH3, NO3, PO4). Other parameters were relatively similar. Physical characteristics of the sites were scored in six habitat categories: (i) substrate characteristics, (ii) instream cover, (iii) channel morphology, (iv) riparian zone and stream-bank conditions, (v) riffle/run quality, (vi) pool quality. Most physical habitat scores were lowest in the intense agriculture zones. 4 The relative importance of physical and chemical variables in explaining variation in macroinvertebrate communities was quantified using canonical correspondence analysis. Substrate characteristics were most important in both surveys. Significant correlations (P

Ground-flora and overstory composition and abundance patterns in 76 upland forest stands were related to the glacial geomorphology of a five-county area in northwestern lower Michigan. Nine classes of landforms were identified based on parent material and surface configuration. Chi-squared analyses of standardized residuals indicated that spatial distribution patterns of the ground flora were significantly associated with glacial landform. Detrended correspondence analyses showed that specific overstory–ground flora assemblages recur in characteristic landscape positions. The Interlobate Moraine, a predominant landform in the northeastern portion of the study area, was characterized by a northern hardwood canopy, primarily sugar maple (Acersaccharum Marsh.), basswood (Tiliaamericana L.), and red oak (Quercusrubra L.), with herbaceous annuals, perennials, and ephemerals forming the ground flora. Other morainal and glaciofluvial landforms supported oak (Quercus spp.) overstories with a woody ericaceous or graminoid ground flora. Ground-flora and overstory composition were highly correlated, indicating that they respond similarly to variation in environmental and historical factors. Moisture availability, as evidenced by soil texture and the presence of subsurface textural discontinuities, appeared to be a predominant factor influencing species distribution patterns. Since moisture availability and related nutrient dynamics are functions of depositional and postglacial history, regional-scale geomorphic patterns constrain forest composition in a probabilistic manner.

In this paper we present the proposed conceptual approach and essential subsystems of SYLVATICA, a landscape-based simulation model which will integrate several resource-management technologies in a visual interactive environment. We believe this approach incorporates several critical technologies which should be common to any integrated landscape-based simulation and visualization environment.

Multivariate classification and ordination methods were used to develop ecological species groups for upland forest ecosystems of northwestern Lower Michigan. Species groups were based on similarities in ground-flora composition and abundance patterns among 76 sample stands. Nine ecological species groups were identified using 48 herbaceous, woody and moss species. The divisive classification produced six classes of stands: two dominated by northern hardwoods and four dominated by oaks. The distribution patterns typified by Deschampsia flexuosa, Vaccinium angustifolium, Viburnum acerifolium, and Desmodium spp. were important in discriminating among oak stands, whereas patterns typical of Osmorhiza claytonii and Maianthemum canadense distinguished northern hardwood stands. The species groups exhibited a wide range of ecological amplitudes, from highly specific associations with individual ecosystems to broad distributions across a range of site conditions. Within the geographic limits imposed by macroclimate and regional physiography, ecological species groups act to integrate site attributes and can simplify the process of mapping ecological land units.

Functional–structural plant models (FSPMs) typically integrate suites of detailed physiological and phenological processes to simulate the growth of individual plants. Recent advances in high-performance computing have allowed FSPMs to be extended to patches of interacting trees. Here, we describe a parallel modelling strategy to run simultaneous individual tree models across an 8 × 8 patch of trees. The 64 ‘core’ trees are surrounded by multiple rings of neighbour trees to remove edge effects. A sensitivity analysis of the patch model demonstrates that computational factors such as the number of independently simulated trees (9 v. 36) or number of neighbour rings (3 v. 6) did not significantly influence model estimates of tree volume growth. Updated submodels for phenology and redistribution of overwinter carbohydrate storage allow the simulation to be more responsive to above ground competition among trees in a patch over multiple growing seasons. An 8-year patch-scale simulation of aspen clones 216 and 259 was conducted using high-resolution environmental data from the Aspen FACE Experiment, a long-term free-air carbon dioxide enrichment (FACE) study. Tree heights and volumes were comparable to 8-year growth measurements made at the Aspen FACE site.

Watersheds represent spatially explicit areas within which terrestrial stressors can be quantified and linked to measures of aquatic ecosystem condition. We delineated thousands of Great Lakes watersheds using previously proven and new watershed delineation techniques. These were used to provide summaries for a variety of anthropogenic stressors within the Great Lakes. All delineation techniques proved useful, but each had applications for which they were most appropriate. A set of watershed delineations and stressor summaries was developed for sampling site identification, providing relatively coarse strata for selecting sites along the U.S. Great Lakes coastline. Subsequent watershed delineations were used for high-resolution site characterization of specific sites and characterizing the full coastal stressor gradient. For these delineations we used three general approaches: 1) segmentation of the shoreline at points midway between adjacent streams and delineation of a watershed for each segment; 2) specific watershed delineations for sampled sites; and 3) a Great Lakes basin-wide, high-resolution approach wherein sub-basins can be agglomerated into larger basins for specific portions of the coast. The third approach is unique in that it provides a nested framework based on hierarchies of catchments with associated stressor data. This hierarchical framework was used to derive additional watershed delineations, and their associated stressor summaries, at four different scales. Providing anthropogenic stressor metrics in such a format that can quickly be summarized for the entire basin at multiple scales, or specifically for particular areas, establishes a strong foundation for quantifying and understanding stressor-response relationships in these coastal environments.

Integrated, quantitative expressions of anthropogenic stress over large geographic regions can be valuable tools in environmental research and management. Despite the fundamental appeal of a regional approach, development of regional stress measures remains one of the most important current challenges in environmental science. Using publicly available, pre-existing spatial datasets, we developed a geographic information system database of 86 variables related to five classes of anthropogenic stress in the U.S. Great Lakes basin: agriculture, atmospheric deposition, human population, land cover, and point source pollution. The original variables were quantified by a variety of data types over a broad range of spatial and classification resolutions. We summarized the original data for 762 watershed-based units that comprise the U.S. portion of the basin and then used principal components analysis to develop overall stress measures within each stress category. We developed a cumulative stress index by combining the first principal component from each of the five stress categories. Maps of the stress measures illustrate strong spatial patterns across the basin, with the greatest amount of stress occurring on the western shore of Lake Michigan, southwest Lake Erie, and southeastern Lake Ontario. We found strong relationships between the stress measures and characteristics of bird communities, fish communities, and water chemistry measurements from the coastal region. The stress measures are taken to represent the major threats to coastal ecosystems in the U.S. Great Lakes. Such regional-scale efforts are critical for understanding relationships between human disturbance and ecosystem response, and can be used to guide environmental decision-making at both regional and local scales.

Presents technical documentation for development of an Ecological Classification System for the Manistee National Forest in northwest Lower Michigan, and suggests procedures applicable to other ecological land classification projects. Includes discussion of sampling design, field data collection, data summarization and analyses, development of classification units, and validation.

The ECOPHYS model is an ecophysiological growth process model of juvenile poplar clones growing under near optimal conditions. The theoretical basis for the ECOPHYS model is that (1) individual leaves drive and control growth; (2) the microenvironment at the leaf exerts primary control of photosynthetic rates; (3) leaf orientation is a major determinant of that microenvironment, (4) photosynthates produced by leaves are allocated among meristematic and respiratory sinks: and (5) the plant's genome and microenvironment regulate photosynthate allocation. The major driving variables are solar radiation, temperature, and clonal morphological and physiological factors. The user can interact or override any or all of the input variables to examine the effects of such changes on photosynthetic production and growth. Verification and sensitivity analyses of ECOPHYS are presented and discussed. The use of ECOPHYS as a research tool is illustrated with several examples. Model potential and limitations are discussed.

During the week of 24 October, 1994, an international group of scientists met in Uppsala, Sweden, to discuss the problems of understanding and modelling the growth of short rotation forests for energy purposes. Three working groups were formed to study specific aspects of the problem: ecophysiology, competition and modelling strategies. This special issue of Biomass and Bioenergy reports the findings of those groups and presents a series of papers related to the ecophysiology and process modelling of Salix .

Potential nitrogen mineralization and nitrification were studied in three upland forest ecosystems to develop an understanding of nitrogen turnover on a landscape basis. The northern Michigan forests studied were an oak ecosystem primarily associated with glacial outwash features and two sugar maple ecosystems that occurred on morainal landforms but differed in the diversity and abundance of ground flora species. Four randomly chosen stands separated by at least 6 km were sampled within each of the three ecosystems. Potential net nitrogen mineralization and nitrification were determined by an aerobic laboratory incubation. Litter was collected from all ecosystems during autumn. Litter production, nitrogen returned to the forest floor, and net mineralization differed by a factor of two between the oak and sugar maple ecosystems. The species-rich sugar maple ecosystems exhibited a fourfold increase in potential nitrification compared with the species-poor sugar maple ecosystem. Nitrification was virtually absent in the oak ecosystem. The distribution of ecosystems could be used to predict differences in potential mineralization and nitrification. Areas susceptible to nitrate loss following intensive forest management practices may be related to the occurrence of plant associations. In this upland landscape, high nitrification potentials appear to be confined to species-rich sugar maple forests.

Variation in overstory biomass and mean annual biomass increment (MABI) of upland forest stands was studied at two spatial scales: glacial landforms (1: 250 000–1: 000 000) and ecological land units (1: 10 000–1: 80 000). Ecological land units were defined based on combinations of ground-flora vegetation, soil, and physiography. Biomass estimates were based on allometric regression equations developed in the Lake States area. Analyses of covariance were used to study the patterns of total biomass and biomass increment among glacial landforms and among ecological land units; stand age was used as the covariate. Overstory biomass ranged from 105 t/ha (MABI = 1.5 t ha−1 year−1) on glacial outwash landforms to 208 t/ha (MABI = 3.2 t ha−1 year−1) on morainal landforms; 37% of the total variation in biomass was accounted for by landform. Analysis at the ecological land unit scale accounted for a higher proportion of variation, approximately 60% of the total. Overstory biomass among ecological land units ranged from 85 t/ha (MABI = 1.3 t ha−1 year−1) for oak-dominated forests occurring on xeric sandy outwash sediments to almost 250 t/ha (MABI = 3.6 t ha−1 year−1) for sugar maple–red oak forests occurring in mesic morainal positions. Variation in biomass appeared to be strongly related to differences in species composition and variation in soil moisture availability, as evidenced by differences in soil texture and the presence or absence of deep-lying textural bands. Since ground-flora composition often reflects variation in soil properties, we examined the relationship between ground-flora composition and biomass increment. Detrended correspondence analysis was used to ordinate stands along a floristic gradient. Regression of stand ordination scores against MABI was significant (r2 = 0.65), indicating a relatively strong association between ground-flora composition and productivity. One advantage of understanding spatial variation in site productivity is that ecological land units can be readily mapped, whereas traditional estimates of site potential, such as site index, are point specific.

Potential net nitrogen (N) mineralization, potential net nitrification, and overstory (boles and branches) biomass were measured in nine forest ecosystems commonly found within the well-drained uplands of northern Lower Michigan. The ecosystem types ranged from oak-dominated forests on coarse-textured outwash sands to mesic northern hardwood forests on sandy glacial till. Overstory biomass was calculated using species-specific allometric equations developed for Lake States hardwood species. Potential net N mineralization and potential net nitrification were measured by a 30-day aerobic laboratory soil incubation. Analyses of (co)variance were used to determine differences in potential N mineralization, net nitrification, overstory biomass, and biomass increment among the nine ecosystem types. Linear and nonlinear regression analyses were used to predict overstory biomass and biomass increment using potential net N mineralization as the independent variable. Overstory biomass ranged from 92 t•ha−1 in a xeric oak ecosystem to 243 t•ha−1 in a northern hardwood ecosystem; annual biomass production ranged from 1.3 to 3.5 t•ha−1 year−1, respectively. Potential net N mineralization was lowest in the xeric oak ecosystem (52.0 μg N•g−1) and greatest in the mesic northern hardwood ecosystem (127.8 μg N•g−1). Potential net nitrification was 45.5 μg NO3−-N•g−1 in the northern hardwood ecosystem; 10 to 230 times greater than in other ecosystems. A saturating exponential model (y = a(1–e−kx) + c) produced the smallest residual mean squares in predicting overstory biomass (R2 = 0.822) and annual biomass increment (R2 = 0.847) from potential net N mineralization. Maximum overstory biomass and biomass increment predicted from this equation were 247 t•ha−1 and 3.7 t•ha−1, respectively. In addition, laboratory net N mineralization potentials were highly correlated with annual rates of N mineralization determined by insitu incubation (r2 = 0.849). Overstory biomass and woody biomass increment were poorly correlated with potential net nitrification. The exponential function used to predict biomass increment from N mineralization suggests that the productivity of some northern hardwood forests in northern Lower Michigan is not limited by N availability.