Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Rocky Mountain Inventory & Monitoring Network 2008-2010 Stream Monitoring Report

National Park Service U.S. Department of the Interior

Natural Resource Stewardship and Science Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Rocky Mountain Inventory & Monitoring Network 2008-2010 Stream Monitoring Report

Natural Resource Technical Report NPS/ROMN/NRTR—2014/881 ON THIS PAGE Billy Schweiger (NPS, ROMN) collects a water sample from the Clark Fork River, GRKO during a May 2008 sample event Photograph by: J. Burke, ROMN

ON THE COVER Jason Smith (NPS, GRKO) collects a periphyton sample from the Clark Fork River, GRKO during a June 2008 sample event Photograph by: B. Schweiger, ROMN Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Rocky Mountain Inventory & Monitoring Network 2008-2010 Stream Monitoring Report

Natural Resource Technical Report NPS/ROMN/NRTR—2014/881

Authors

E. William Schweiger,1 Laura O’Gan,1 Donna Shorrock,1 and Mike Britten1

With contributions from: Phil Kaufmann,2 Jason Smith,3 Eric Mason,3 Isabel Ashton,4 Laura Atkinson,1 and Kristin Long1

1Rocky Mountain Network National Park Service Fort Collins, Colorado

2Office of Research and Development Western Ecology Division U.S. Environmental Protection Agency Corvallis, Oregon

3Grant-Kohrs Ranch National Historic Site National Park Service Deer Lodge, Montana

4Rocky Mountain National Park National Park Service Estes Park, Colorado

Editor

Nina Chambers Northern Rockies Conservation Cooperative Jackson, Wyoming

June 2014

U.S. Department of the Interior National Park Service Natural Resource Stewardship and Science Fort Collins, Colorado The National Park Service, Natural Resource Stewardship and Science office in Fort Collins, Colorado, publishes a range of reports that address natural resource topics. These reports are of interest and applicability to a broad audience in the National Park Service and others in natural resource management, including scientists, conservation and environmental constituencies, and the public.

The Natural Resource Technical Report Series is used to disseminate results of scientific studies in the physical, biological, and social sciences for both the advancement of science and the achievement of the National Park Service mission. The series provides contributors with a forum for displaying comprehensive data that are often deleted from journals because of page limitations contain NPS policy considerations or address sensitive issues of management applicability.

All manuscripts in the series receive the appropriate level of peer review to ensure that the information is scientifically credible, technically accurate, appropriately written for the intended audience, and designed and published in a professional manner. This report received formal peer review by subject-matter experts who were not directly involved in the collection, analysis, or reporting of the data, and whose background and expertise put them on par technically and scientifically with the authors of the information.

Views, statements, findings, conclusions, recommendations, and data in this report do not necessarily reflect views and policies of the National Park Service, U.S. Department of the Interior. Mention of trade names or commercial products does not constitute endorsement or recommendation for use by the U.S. Government.

This report is available from theRocky Mountain Network website (http://www.nature.nps. gov/im/units/romn), and the Natural Resource Publications Management Web site (http:// www.nature.nps.gov/publications/nrpm/) on the Internet. To receive this report in a format optimized for screen readers, please email [email protected].

Please cite this publication as:

Schweiger, E. W., L. O’Gan, D. Shorrock, and M. Britten. 2014. Stream ecological integrity at Grant-Kohrs Ranch National Historic Site: Rocky Mountain Inventory & Monitoring Network 2008-2010 stream monitoring report. Natural Resource Technical Report NPS/ ROMN/NRTR—2014/881. National Park Service, Fort Collins, Colorado.

NPS 451/124876, June 2014 ii Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Contents Page

Figures vii Tables ix Appendices xi Executive Summary ...... xiii GRKO SEI Summary Condition Table xv Acknowledgements xix Introduction ...... 1 The National Park Service Inventory and Monitoring Program 1 Purpose of Report 1 Stream Ecological Integrity: Synthetic Index to General Condition 2 The Stream Resource at Grant-Kohrs Ranch National Historic Site 3 Why Monitor the Clark Fork River? 3 Current Importance of the Clark Fork to GRKO and the Deer Lodge Valley 6 Historical Role of the Clark Fork River 6 Ecological Monitoring Considerations 7 Regulatory Considerations 7 NPS Policies 7 Clean Water Act 8 Interpreting ROMN Stream Ecological Integrity Monitoring Data 8 Legal/Regulatory Status of the Clark Fork at GRKO 8 Clark Fork Superfund Project 10 Climate Change 11 Aquatic and Riparian Invasive Species 12 Monitoring Objectives 12 Methods ...... 15 Sample Design 15 Sample Events 15 Field Data 16 Site Layout 17 Auxiliary Data 17 Water Physiochemistry and Streamflow 17 Biological 19 Habitat 20 Quality Assurance and Quality Control 20

Contents iii Contents (continued) Page Signal to Noise Ratio 20 Analyses 21 Status 21 Water and Sediment Physiochemistry 21 Physical Habitat 22 Biology 23 Trend 23 Water Physiochemistry 24 Assessment: Reference Condition, Thresholds and Interpretation 25 Reference Condition 25 Assessment Points 26 ROMN Assessment Process at GRKO 28 Summary 31 Results and Discussion 33 Sample Sites and Events 33 Water and Sediment Physiochemistry 33 Summary 33 Measures, Criteria, and Assessment Points 34 Quality Assurance and Quality Control 37 In situ Parameters 37 Major Constituents 38 Nutrients 39 Metals in Water 40 Metals in Sediment 41 Special Case: Water Temperature 42 Select Trends in Water Physiochemistry 46 Physical Habitat 48 Summary 48 Habitat Metrics and Assessment Points 48 Quality Assurance and Quality Control 51 Habitat Volume 51 Habitat Complexity 52 Cover for Stream Biota 53 Substrate 54 Relative Bed Stability 56 iv Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Contents (continued) Page Floodplain Interaction 57 Hydrologic Regime (Habitat-Based) 58 Riparian Vegetation Cover 59 Human Disturbance 59 Hydrology: Stream Flow 63 Biology 63 Summary 63 Biological Metrics and Assessment Points 64 Quality Assurance and Quality Control 65 Macroinvertebrates 65 Periphyton (Diatoms) 71 Aquatic Invasive Species 75 Confidence 77 Summary and Conclusions ...... 77 Water Physiochemistry 78 Habitat 78 Biology 78 Management Implications 79 Climate Change 79 Future Work 80 Literature Cited ...... 81

Contents v

Figures Page Figure 1. Grant-Kohrs Ranch National Historic Site in relation to other parks in the ROMN. 2 Figure 2. Grant-Kohrs Ranch National Historic Site showing the SEI sample site and USGS stream gauge location in Deer Lodge. The green squares show transect locations sampled during a full sample event, the red star gives the location of sentinel sample events and the blue triangle shows the USGS gauge. Inset shows the upper Clark Fork River watershed from the Blackfoot River near Missoula. 4 Figure 3. Oblique view (looking North) of the Clark Fork River at Grant-Kohrs Ranch National Historic Site (red border) and the town of Deer Lodge, MT. Background imagery is early fall 2010 (courtesy Google Earth). Note the park boundary shown is an unofficial ‘fence line’ boundary. The green squares show transect locations sampled during a full sample event, the red star gives the location of sentinel sample events and the blue triangle shows the USGS gauge. The Deer Lodge water treatment ponds are visible in the upper portion of the ranch. Stream flow is from the foreground of the image. The inset shows the upper Clark Fork region with Anaconda, Butte, and the Milltown ponds (source of contaminated sediments) in the foreground. 5 Figure 4. The Clark Fork River in GRKO during a 2008 SEI sentinel water and sediment sampling event. Stream flow is from the foreground of the image. 5 Figure 5. Robert Warren on Clark Fork River Bridge, ca. 1936. GRKO PHOTO #16054, Grant-Kohrs Ranch NHS Historic Archives. 6 Figure 6. Clark Fork Operable Unit. GRKO is located next to the town of Deer Lodge. Map courtesy of EPA (2009b). 10 Figure 7. ROMN SEI sample reach in oblique profile. Stream flow is from the top to bottom of the figure. Cross-sections, sub-sample locations, riparian plot, and thalweg profile all shown. Inset pictures show select details for substrate, channel and riparian sampling. 18 Figure 8. MT DEQ, EPA, and STAR monitoring sites in the Middle Rockies ecoregion within Montana as used in comparison to select ROMN SEI results for GRKO. Sites are classified as reference, degraded or unknown (by the programs responsible for each site) with green showing reference sites, red degraded sites, and orange unknown (“test”) sites. Inset shows all sites across the state with symbology based on the source program of each site. Note that additional sites in the ecoregion but outside of Montana are not shown here were used for select habitat and chemistry parameters. 19 Figure 9. Daily stream temperature by water year for Clark Fork River at GRKO bridge crossing: (top) 2008-2009, (middle) 2009-2010, and (lower) 2010-2011. 44 Figure 11. Mean date of last (left panel) or first freeze (right panel) for water years, 2001-2010 at Deer Lodge USGS gauge. Historic and current periods are largely arbitrary. Current is defined by the beginning of ROMN SEI monitoring in 2007 and historic is considered to be 2001 to 2006. 45 Figure 10. Mean seven-day maximum water temperature for water years, 2001- 2010 at Deer Lodge USGS gauge. Horizontal red line depicts maximum summer temperature threshold (USFWS 1998) for optimal bull trout rearing. Dotted vertical gray lines delineate water year boundaries. 45 Figure 12. ROMN SEI field crew collect data on stream physical habitat in the Clark Fork in 2008. The area in the photo shows good examples of woody debris, complex off channel habitats, overhanging banks and riparian cover. 53

Executive Summary vii Figures (continued) Page

Figure 13. Clarification on arrangement of finning and armoring reference and non-reference thresholds for relative bed stability (log RBS; Stoddard et al. 2005). Unlabeled orange regions are intermediate between reference and non-reference states. Black arrow indicates approximate location of 2009 GRKO RBS value. 56 Figure 14. Leafy spurge (Euphorbia esula) on banks of the Clark Fork at GRKO in 2007. 62 Figure 15. Hydrographs for 2007-2010 water years at the Clark Fork River Deer Lodge USGS gauge (station 12324200). Figures include current water years (heavy blue or red lines), mean daily values for 1980-2010 (faint green lines) and the midpoint of current and historical discharge. Note different scale for 2007-2008 water year. Additional statistics are given in Table 12. Graphics were produced in the Climate Data Summarizer, version 3.2 courtesy of Mike Tercek. 64 Figure 16. Hydropsyche spp. caddisfly, common in the Clark Fork at GRKO. 70 Figure 17. Scanning electron micrograph image of the most common algae at GRKO, Diatoma moniliformis...... 74 Figure 18. Point locations of the New Zealand mudsnail (Potamopyrgus antipodarum) in the Montana/Idaho/Wyoming area. Red dots indicate established populations. Yellow dots indicate confirmed collections. Approximate location of GRKO is shown with the red star. Data are from several (uncoordinated) sample designs and therefore are not a valid sample of this region as a whole. Data span 2004-2009. Data and map courtesy of the USGS Nonindigenous Aquatic Species program. 75 Figure 19. Point locations of confirmed presence of didymo Didymosphenia( geminata) in the Montana region. Data are from several (uncoordinated) sample designs and therefore are not a valid sample of this region as a whole. Data span 2004-2009. Records outside of Glacier National Park are based on data from USGS National Water Quality Assessment (NAWQA), EPA Environmental Monitoring and Assessment (EMAP), and samples from other studies and are courtesy Karl Hermann, Sarah Spaulding, and Tera Keller. Data in Glacier National Park are from ROMN SEI monitoring. 75 Figure A1. Jason Smith (NPS, GRKO) collects macroinvertebrates on the Clark Fork in 2009 using the standardized methods of the SEI protocol. 101

viii Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Tables Page

Table 1. Summary of regulatory status of waters in GRKO as of 2012; (a) NPS (2012) compilation of extent of impaired status and (b) MT DEQ summary of impairment causes and TMDL status (MT DEQ 2012b). Extent numbers are restricted to within the park’s boundaries or to adjacent* waterbodies. ORW and 303(d) designation applies to all flowing water types. 9 Table 2. Summary of SEI field data. 16 Table 3. Summary of 21 core SEI bioassessment metrics and rubrics for their interpretation. As discussed in text, both current and historic models are used. 24 Table 4. SEI sample events at GRKO, 2008-2010. 33 Table 5. Summary condition table excerpt for select water and sediment measures at GRKO from 2008 to 2010. We include example vital signs and indicators, a brief description of results and patterns, and symbolize the status, trend, and our confidence in those summaries. See the Executive Summary for the complete Summary Condition Table. 33 Table 6. Summary of water and sediment physiochemistry for the Clark Fork River in Grant-Kohrs Ranch National Historic Site, 2008-2010. Except as indicated, assessment points are state of MT chronic aquatic-life criteria with human-health values in parentheses. All assessment points define non-reference except those in green4. The non-reference state is above all assessment points except for those with an inequality indicating that it is below the value. Concentrations in bold red indicate results in non-reference at least once during 2008-2010. Values in bold are either intermediate or non-reference (depending on the existence of a non- reference assessment point). For additional clarifications on Table content see Notes below. 34 Table 7. Mean August stream water temperature estimated from daily mean values for water years 2000-2009 at the Deer Lodge USGS gauge. Data were not available for the 2006 water year. Sample sizes for other years was about 200. 43 Table 8. Results of the seasonal Kendall test for trends in discharge and select constituent concentrations on the Clark Fork River using data from the Deer Lodge USGS gauge. Bold type indicates statistically significant trend models (at p<0.10 following USGS Schertz et al. 1991, Slack and Lorenz 2003) in constituent concentration (or loading if flow adjustment is significant). 47 Table 9. Summary condition table excerpt for select riparian and stream habitat metrics at GRKO in 2009. We include example vital signs and indicators, a brief description of results and patterns, and symbolize the status, trend, and our confidence in those summaries. See the Executive Summary for the complete Summary Condition Table. 48 Table 10. Summary of core physical habitat metrics for the Clark Fork River in Grant-Kohrs Ranch National Historic Site, 2009. Assessment points are for reference and non-reference condition classes with the direction of inequality indicating whether the state is above or below a value. Metric values in bold red are in non- reference. Metric values in bold green are in reference. Metric values in bold are either intermediate or non-reference (depending on the existence of a non-reference threshold). 49 Table 11. Frequency of 15 target invasive taxa at GRKO during 2009 sample event. Values range from 0 to 1, with 1 describing a species that occurred at all 11 riparian plots. 61 Table 12. Select stream flow statistics from the Deer Lodge USGS station on the Clark Fork River near GRKO. All discharge results are in cubic feet per second (cfs). 65 Table 13. Flood frequency analysis for the extended Deer Lodge USGS gauge

Contents ix Tables (continued) Page period of record (1978 to 2013). Two year and ten year return intervals are highlighted in red. All discharge results are in cubic feet per second (cfs). 65 Table 14. Summary condition table excerpt for select biological metrics at GRKO in 2008 and 2009. We include example vital signs and assessment points, a brief description of results and patterns, and symbolize the status, trend, and our confidence in those summaries. See the Executive Summary for the complete Summary Condition Table. 66 Table 15. Summary of core macroinvertebrate metrics for the Clark Fork River in Grant-Kohrs Ranch National Historic Site, 2008-2009. Assessment points are derived from various sources and are for reference or non-reference condition classes with the direction of the inequality indicating whether the state is above or below a value. Select ecoregion assessment points are given within parentheses. Metric values in bold red are in non-reference. Metric values in bold green are in reference. Metric values in bold are either intermediate or non-reference (depending on the existence of a non-reference assessment point). 67 Table 16. Summary of core diatom metrics for the Clark Fork River in Grant-Kohrs Ranch National Historic Site, 2008-2009. Assessment points are derived from various sources and are for reference or non-reference condition classes with the direction of the inequality indicating whether the state is above or below a value. Metric values in bold red are in non-reference. Metric values in bold green are in reference. Metric values in bold are either intermediate or non-reference (depending on the existence of a non-reference assessment points). 72 Table 17. Summary condition table excerpt for the overall assessment of the condition of the Clark Fork at GRKO in 2008-2010. We include example vital signs and indicators, a brief description of results and patterns, and symbolize the status, trend, and our confidence in those summaries. See the Executive Summary for the complete Summary Condition Table. 77 Table A1. Categories of ROMN SEI physical habitat measures. 100 Table C1. Macroinvertebrate taxa list and relative abundances from GRKO SEI sample events in 2008 and 2009. Counts are followed by the relative abundances (in parentheses) within each sample. Bold values indicate dominant (>10% relative abundance) within a sample. Benthos taxa were identified by multiple taxonomist so we use Operational Taxonomic Units (OTUs) to collapse identifications to comparable values. A (blank) family indicates the level of identification was above family. 107 Table C2. Diatom species list and abundance from long-term stations from GRKO SEI sample events in 2008 and 2009. Counts are followed by the relative abundances (in parentheses) within each sample. Bold values indicate dominant (>10% RA) within a sample. Nearly all diatoms were identified to the species level by a single taxonomist so we do not require OTUs. P indicates the taxon was present in a whole slide search (P is arbitrarily given a count of 0.5). 109 Table D1. ROMN SEI quality control and quality assurance measures. 111 Table D2. Water chemistry detection limits for GRKO 2008-2010 data. 113

x Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Appendices Page

Appendix A: Field Method Overview and Justification 97 Appendix B: Macroinvertebrate and Diatom Bioassessment Metrics 103 Appendix C: Taxa Lists 107 Appendix D: Quality Assurance and Quality Control Overview 111 Appendix E: Glossary 115

Contents xi

Executive Summary

The following report presents results from pilot monitoring of Stream Ecological Integrity (SEI) in the Clark Fork River at Grant-Kohrs Ranch National Historic Site (GRKO) from 2008-2010. Overall, the ecological integrity of the river during this time was largely intact and of generally higher quality, a somewhat surprising result given the history of intensive land uses in the watershed. Metals from historical mining in the watershed continued to occur in elevated concentrations, but may not have been having a strong effect on key biological communities in the river. A summary of key results is given in a Summary Condition Table that follows.

The general objectives of SEI monitoring are to document the status and long-term trends in ecological condition of streams and rivers in a park, and to use this information to help understand why stream conditions may have changed. Resource managers may, in turn, use information from SEI monitoring to make resource management decisions within their parks. The SEI protocol is an integrated and comprehensive approach to monitoring stream resources in parks within the Rocky Mountain Inventory & Monitoring Network (ROMN). The ROMN selected indicators of stream ecosystem condition as high priority Vital Signs in its planning process because streams are fundamental components of every ROMN park and their ecology is intimately linked and reflective of conditions in the watersheds they drain. Streams support a broad spectrum of ecological services including hydrologic cycling, nutrient processing, and wildlife habitat as well as socioeconomic functions including recreation, fisheries, and water sources. Finally, streams are sensitive to stresses such as excessive sediment and nutrient inputs, withdrawal for municipal and agricultural use, and changing climate making them ideal for long-term ecological health monitoring. At GRKO, SEI monitoring is also motivated by conditions that were created by past mining activities in the area and by the associated Superfund restoration. However, SEI monitoring is not designed for nor meant to quantify the effectiveness of ongoing remediation and restoration of the Clark Fork channel and riparian habitats in GRKO, or other work being conducted under the auspices of the Superfund Record of Decision or the state monitoring plan.

Our assessment has largely been conducted for informational purposes, outside of any regulatory context, to provide the park and its stakeholders with baseline information regarding the current condition of the Clark Fork. The NPS does not have regulatory authority over waters in a park or the authority to evaluate beneficial designated uses. Formal interpretation of the State of Montana’s water quality standards and impairment decisions are beyond the scope of these analyses and are not provided. Therefore, any comparisons we make using SEI or auxiliary data to water quality standards, criteria or assessment point do not officially include any statement as to whether a beneficial designated use was attained or not. However, NPS can participate with states in collecting and (informally) evaluating data used in the protection of water bodies in parks but that lie under state jurisdiction. Therefore, parks like GRKO can benefit from the nonbinding or informal data and interpretations in this report.

The ROMN defines the ecological integrity of a stream as the capacity to support and maintain a balanced, integrated, and adaptive community of organisms; and having a species composition, diversity, and functional organization comparable to minimally- disturbed natural streams in the ecoregion. The SEI protocol includes a broad spectrum of indicators to help evaluate the ecological integrity of the Clark Fork River at GRKO. Core indicators include multiple measures of water and sediment physiochemistry (i.e., nutrients, metals, temperature); physical habitat (i.e., substrate composition, riparian cover, channel geomorphology); and community-level assays of two important biological assemblages (i.e., macroinvertebrates and algae). The latter provide an integrated aspect for assessing ecological integrity because these organisms respond to environmental conditions over time

Executive Summary xiii and space. We include measures of drivers (or stressors) and ecological responses, so that we may better understand the linkages between these and help parks apply these results to resource management. We are able to do this in large parks where we have an extensive sample size or, for single sites like at GRKO, after we have a time series of data.

SEI methods are documented in a draft protocol and are largely derived from well- established and existing protocols developed by ROMN partners, including the U.S. Environmental Protection Agency (EPA), the U.S. Geological Survey (USGS), and the Montana Department of Environmental Quality (MT DEQ). The application of a standardized monitoring protocol across sites facilitates comparison of streams and rivers within an ecoregion.

SEI data were analyzed following well-established methods. Biological response was emphasized because the presence and health of populations of key taxa are indicative of overall stream health. Bioassessment tools include Multimetric Indices, which combine multiple elements of an assemblage into a synthetic assay of community structure; a River Invertebrate Prediction and Classification model, which compares an expected species list at each site with what is actually collected, and models that describe the prevalence of algae taxa that increase when specific stressors occur in a stream. Water quality monitoring data were interpreted using existing or proposed standards (also called criteria). For biology and habitat response, thresholds from state and federal agencies were used as relevant to the Clark Fork in GRKO. Finally, for select responses, additional or alternate assessment points were derived from reference sites in partner monitoring stream networks within the Middle Rockies ecoregion to facilitate comparison to a broader-scale reference state.

Throughout this document, we use the terms “reference” and “non-reference” to describe how an indicator relates to criteria or other assessment points. An indicator can be in a reference state if it lies within a range of values that represent an intact state (i.e., “good condition”) from an ecological or other perspective (e.g., human health). In contrast, the indicator may also be in a non-reference or degraded state (“poor condition”). Or, in many cases, an indicator may be intermediate, somewhere between a reference and non-reference state.

This effort has been, and will continue to be, a cooperative undertaking between ROMN staff, aquatic scientists, especially from USGS, EPA, and MT DEQ, and most importantly, GRKO management and staff. Our results represent the efforts of many dedicated scientists and resource managers over several years and are valid, representative, and address our objectives. They should be useful for understanding and, within the constraints of the protocol, managing the Clark Fork. However, our data are incomplete—we are at the beginning of an effort that will take many years before there is sufficient data to understand long-term trends. As we accrue data we will also continue to improve our understanding of the mechanisms and drivers behind aquatic system health and specifically, SEI responses. Future SEI monitoring will permit continued analysis of SEI data relative to the criteria and assessment points for the state and Middle Rockies ecoregion. These analyses will help scientists continue to monitor the most cost-effective and management-relevant suite of indicators of ecological integrity of the Clark Fork in GRKO.

xiv Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site GRKO SEI Summary Condition Table SEI results for GRKO are summarized in the following resource condition summary table. We include example vital sign and indicators, a brief description of results and patterns and symbolize the status, trend, and our confidence in those summaries. This table is a considerable simplification of the detail in the full report and should be used with caution. For example, some indicators within a category have divergent patterns (i.e., one may be in a reference state and another is not). We qualitatively weigh this variance using our best professional judgment to derive an overall assessment for each category.

The following provides a key for the symbols used in the Summary Condition Table.

Status Trend Confidence

Significant Condition is Concern/ Improving/ High Non-reference Improving trend

Condition is Caution/ Unchanging/ Medium Intermediate Stable trend

Condition is Good Condition/ Deteriorating/ Low Reference Decreasing trend

No symbol Unknown trend NOTES: The status of a vital sign or indicator can be in a “reference state,” that is, based on our knowledge of the vital sign within a range of values considered “healthy” from an ecological or other perspective (e.g., human health). The status may also be “non- reference” or unhealthy; or it may be in a state that is “intermediate” between reference and non-reference. These designations are be simplified into green circles for “good condition,” yellow circles for “intermediate” or “caution,” or red circles for “significant concern.” The trend in condition over time is also symbolized; improving conditions are represented by up-arrows, deteriorating conditions are represented by down-arrows, and stable conditions are represented by horizontal or flat arrows. In some cases, we do not have sufficient data over time to evaluate trends, for those vital signs or indicators there is no arrow symbol within the circle. Confidence in our evaluation of status and trend is symbolized by the thickness or character of the outside line of the symbol. “High” confidence in the data and our interpretation is represented by a thick line, “medium” confidence is represented by a thin line, and “low” confidence is symbolized by a dashed line.

Executive Summary xv Summary Condition Table Clark Fork River at GRKO, 2008 to 2010

Vital Sign Summary Symbol (Example Indicators) Overall ecological The ecological integrity of the Clark Fork from 2008 to 2010 integrity was largely intact and of medium to higher quality— somewhat surprising result given the history of intensive land uses in the watershed. However, there are lingering issues with metals in water and sediment from historical mining in the watershed, and select aspects of in channel and riparian habitat are in a non-reference condition. In general the condition of the system may be slowly improving. We have medium confidence in our interpretation.

Water Most major ion concentrations were in an acceptable range yet physiochemistry chloride and sulfate concentrations were often higher than in (major ions, nutrients, ecoregion reference sites. Arsenic, cadmium, copper and lead metals) exceeded State of Montana water quality criteria. Nutrient concentrations were occasionally high but we currently lack sufficient data to apply the complete MT DEQ nutrient assessment protocol. There is evidence that the long-term trend in several parameters is improving. We have medium confidence in our assessment of major ions, nutrients, metals at GRKO.

Water in-situ All of the core NPS parameters were in an acceptable range. The chemistry long-term trend in stream temperature suggests rising temperatures (pH, conductivity, DO, —or a deteriorating condition (but the period of record is short temperature) at only around nine years). We have medium confidence in our assessment of in situ parameters at GRKO.

Sediment chemistry Arsenic , copper, and zinc in sediment concentrations were often (metals) above their non-reference assessment points. Total mercury concentrations were intermediate but because the primary toxic form of mercury is methylmercury, these data are somewhat inconclusive. We currently lack data to assess trends, but work by partners suggests there is an improving trend in many metals found in sediments. We have medium confidence in our of sediment chemistry at GRKO.

Habitat, sediment Fine substrates in the Clark Fork were more prevalent and had (size, stability) higher levels of embeddedness than select species of sediment intolerant macroinvertebrates and bull trout prefer. Bed load was also relatively substantial and mobile, but we need additional investigation to link sediment load/source to biotic condition at the site. We have lower confidence in our assessment of sediments at GRKO (due to a small sample size).

In stream and ripar- Riparian cover was patchy on the Clark Fork and there were ian habitat more adjacent potential stressors in the GRKO floodplain than at (complexity, cover, ecoregion reference sites. Invasive plants were fairly common, and disturbance) some occurred with higher frequency than in ecoregion reference sites. The park’s best management practices likely mitigate most local habitat stressors, but upstream disturbance at GRKO likely has a complex impact on habitat in the park. We need more data to confirm these issues and assess trend in these responses. We have lower confidence in our assessment of in stream and riparian habitat at GRKO (due to a small sample size).

xvi Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Vital Sign Summary Symbol (Example Indicators) Habitat, stream Stream flow during 2008 to 2010 relative to the period of record flow suggest SEI monitoring occurred in somewhat wet years. Seasonal (amount and timing) patterns were fairly typical for mountain valley streams, but there was a suggestion of higher and earlier spring runoff. Over the long term, total annual stream flow may be decreasing, but more data are needed. We have medium confidence in our assessment of stream flow at GRKO.

Biological Although there are several stressors impacting the Clark Fork communities, at GRKO, they may not be having a strong effect on the overall macroinvertebrates integrity of the macroinvertebrate community as most synthetic (MMI and RIVPACS metrics were in a reference state relative to MT DEQ and reference metrics) assessment points from across the ecoregion. Some nutrient metrics may suggest a small biological response to increased nutrients. We lack data to assess trends, although longer-term monitoring by USGS and MT DEQ suggests there have been marginal improvements in macroinvertebrate communities in the Clark Fork. We have medium confidence in our assessment of benthos communities at GRKO.

Biological Although there are several stressors impacting the Clark Fork communities, at GRKO, they may not be having a strong effect on the overall diatoms integrity of the macroinvertebrate community as most synthetic (increaser metrics, metrics were in a reference state relative to MT DEQ criteria and MMI) ecoregion assessment points. Some nutrient metrics may suggest a small biological response to increased nutrients. We lack data to assess trends, although longer-term monitoring by USGS and MT DEQ suggests there have been marginal improvements in macroinvertebrate communities in the Clark Fork. We have medium confidence in our assessment of benthos communities at GRKO.

Like macroinvertebrates, most diatom metrics were above MT DEQ criteria (or in a reference state). Diatoms are the base of the food chain and the intact diatom community at GRKO may be one of the reasons why we also see fairly high-quality macroinvertebrate assemblages. There were subtle suggestions in some sediment metrics that the high level of fine sediments in the channel might be impacting diatom communities. Likewise, nutrient metrics were variable with some suggestion of issues, especially in 2008. We lack data to assess trends in diatom communities. We have medium confidence in our assessment of diatom communities at GRKO.

Aquatic invasives No aquatic invasive species were found, although Didymosphenia (presence) geminata (rock snot) and New Zealand mud snails (Potamopyrgus antipodarum) are in the Clark Fork watershed and spreading. SEI monitoring will watch closely for these and other invasive species over the coming years. We have medium confidence in our assessment of aquatic invasives at GRKO.

Executive Summary xvii

Acknowledgements This report was prepared by ROMN staff based on input and information from many people; any errors are the responsibility of ROMN staff. The ROMN Technical Committee laid the groundwork for this protocol and continues to provide vision and guidance for our long-term inventory and monitoring program. Thanks to the current and past members of the committee: Mark Biel, Ben Bobowski, Fred Bunch, Jeff Connor, Ken Czarnowski, Chris Ford, Jack Potter, Jason Smith, Melana Stichman, Michael Stops, Terry Terrell, Kathy Tonnessen, Judy Visty, Leigh Welling, Phil Wilson, Rick Wilson, and Chris Ziegler.

Thanks are due to park staff members who shared their field labor, experience and knowledge of streams and other park resources. Special thanks to Jason Smith and Eric Mason, Laura Rotegard, Ben Bobowski, and Chris Ford.

We would like to thank the many outside scientists and experts who have generously shared their knowledge of the Clark Fork in GRKO with us, including: Don Campbell, Dave Clow, and Alisa Mast from the USGS; Tina Laidlaw, Michael Bade, Paul Ringold, John Van Sickle, Phil Larsen, John Stoddard, and Dave Peck from the EPA; Chuck Hawkins from Utah State University; Rosie Sada, Mike Suplee, and Dave Feldman from MT DEQ; Diane Winter, Jennifer Bowman, and Wease Bollman from Rhithron, Inc.; Chris Seibold from the Kiowa Lab at Colorado University-Boulder; Ric Hauer from the Flathead Lake Biological Laboratory; and Loren Bahls from Hannea Inc. We especially thank Karl Hermann, Michael Bade, and all other staff from the EPA Region 8 lab for their assistance and for the pro bono analyses of our sediment and select water chemistry samples.

We also extend our thanks to NPS Natural Resource Stewardship and Science Directorate (NRSS) employees who provided support, advice and information, including Steve Fancy, John Gross, Roy Irwin, Pete Penoyer, Jeff Albright, and Bill Hansen. In similar ways, the staff of other NPS monitoring networks, especially the Northern Colorado Plateau Klamath Networks, benefited our efforts and we are grateful to them. Thanks also to Bruce Bingham, as Intermountain Region Inventory and Monitoring Coordinator, for providing sound advice, encouragement, and support.

Our work was made possible by the administrative support and advice provided by the NPS Intermountain Region, especially Gay Shockley, Leslie Aills, and Bruce Bingham. Kathy Tonnessen, in her role as the Rocky Mountain Cooperative Ecosystem Studies Unit NPS representative, has been extremely helpful in facilitating cooperative task agreements with academic partners. Thanks to CarolAnn Moorhead and Nina Chambers for assistance with editing.

Comments and suggestions from Rosie Sada, Darrin Kron and Dave Feldman from MT DEQ, and Michael Bozek from NPS, who all graciously agreed to peer review this report, were very helpful and used to improve report content. Finally, little of this would have been possible without the SEI field crew including: Jason Smith, Billy Schweiger, Donna Shorrock, Isabel Ashton, Kyle Motley, Alex Cahlander-Mooers, Rosie Fewings, Jennifer Burke, Mike Britten, Sue Novak, and Eric Schmidt.

Acknowledgements xix

Introduction

The National Park Service By monitoring these vital signs in national Inventory and Monitoring Program parks across the network, the ROMN aims The purpose of the National Park Service to efficiently provide comparable long-term (NPS) Inventory & Monitoring (I&M) ecological health information in national Program is to develop and provide scientifi- parks across broad latitudinal and elevation cally sound information on the current status gradients from the southern to the northern and long-term trends in the composition, Rocky Mountains. structure, and function of park ecosystems. Stream Ecological Integrity (SEI) monitoring As part of the NPS’s effort to improve park in GRKO addresses five of ROMN’s 12 high- management through greater reliance on priority vital signs: water chemistry, surface scientific knowledge, a primary role of the water dynamics, freshwater communities, I&M Program is to collect, organize, and invasive/exotic aquatic biota, and invasive/ make available natural resource data and exotic plants. Three other high priority to contribute to the Service’s institutional vital signs are indirectly linked with the SEI knowledge by facilitating the transformation protocol including: landscape dynamics, of data into information through analysis, wet and dry deposition, and weather and synthesis, and modeling of specific key “vital climate. The SEI protocol in GRKO is a signs.” The I&M Program defines the term multi-faceted ecological health approach to vital sign as a subset of physical, chemical, monitoring streams as guided by a detailed and biological elements and processes of Stream Ecological Integrity (SEI) protocol park ecosystems that is selected to repre- (Schweiger et al. In Review). The SEI sent the overall health or condition of park protocol describes an integrated approach resources, known or hypothesized effects of to understanding status and trend in stream stressors, or elements that have important ecological condition, capturing the strengths human values (Fancy et al. 2009). of both fixed-site water quality-based The Rocky Mountain Inventory and approaches and probability surveys. Monitoring Network (ROMN) is comprised of six NPS units: Glacier National Park Purpose of Report (GLAC), Grant-Kohrs Ranch National This report, Stream Ecological Integrity Historic Site (GRKO), and Little Bighorn at Grant-Kohrs Ranch National Historic Battlefield National Monument (LIBI), Site, presents summaries and general Montana; and Florissant Fossil Beds interpretation of SEI work conducted during National Monument (FLFO), Great Sand 2008-2010 at GRKO. We provide key data Dunes National Park and Preserve (GRSA), and results in the main body of the report. and Rocky Mountain National Park Other data or results are in the appendices or (ROMO), Colorado (Figure 1). Through a are available upon request from the ROMN. comprehensive three-year scoping process, the ROMN and its partner parks and Because this is the first report on SEI scientific collaborators, identified 12 high- monitoring at GRKO, we include priority vital signs for focused long-term justification and some detail for key elements monitoring. The vital signs include: Wet of the protocol; readers experienced in and Dry Deposition; Weather and Climate; stream monitoring and assessment may only Water Chemistry; Surface Water Dynamics; need to skim the Introduction and Methods Freshwater Communities; Invasive/Exotic sections. The SEI protocol (Schweiger et al. Aquatic Biota; Groundwater Dynamics; In Review) provides additional rationale for Wetland Communities; Invasive/Exotic the design and field elements, summaries Plants; Vegetation Composition, Structure, of analytical methods, overviews of data and Soils; Focal Species (Beaver, Elk, Grizzly management, and an administrative plan for Bear, and GRSA Endemic Insects); and long-term implementation. The SEI protocol Landscape Dynamics (Britten et al. 2007). also includes a series of standard operating procedures (SOPs) that provide detailed

Introduction 1 Figure 1. Grant-Kohrs Ranch National Historic Site in relation to other parks in the ROMN.

instructions for executing field, analytical, multidimensional concept and usually a and select administrative components of the single indicator or vital sign is insufficient protocol. to characterize it. Therefore, we monitor an integrated set of response measures Stream Ecological Integrity: such as community-level composition, Synthetic Index to General stream habitat (including hydrology), and Condition water chemistry. The ROMN recognizes The SEI protocol focuses on the “ecological that ecological integrity may not be the integrity” of streams. Ecological integrity primary goal of park resource management, is the capacity to support and maintain particularly at historical parks like GRKO a balanced, integrated, and adaptive where cultural resource management may community of organisms having a take precedence. Moreover, monitoring species composition, diversity, and ecological integrity may not provide project- functional organization comparable to specific results for a single, focused issue that of natural habitats of the region (Karr in a park like “effectiveness monitoring” 1991). Ecological integrity is a complex, would. The ROMN works with network

2 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site parks to develop monitoring approaches the 7th order Clark Fork River downstream that are specific to each park and the key of the town of Deer Lodge between issues impacting stream condition in a given Cottonwood Creek and Fred Burr Creek system, integrating the broader, longer-term (Figure 2 and Figure 3). The Clark Fork concepts within ecological integrity with River (Figure 4) begins at the confluence of more case-specific needs. Silver Bow Creek and Warm Springs Creek and flows northward through the broad and The ecology of streams is both intimately meandering floodplain of the Deer Lodge linked with and reflective of the watersheds Valley with an average gradient of 3.8 m/ they drain (Hynes 1972). A defining feature km. From its headwaters, the Clark Fork of streams and rivers is their dependence flows north for approximately 43 river miles on the landscape in which they reside for past the towns of Galen, Deer Lodge, and inputs of energy and nutrients (Naiman Garrison. The river then runs northwest 1992, Hunsaker and Levine 1995). Streams for 77 river miles to the headwaters of the support a broad spectrum of ecological Milltown Reservoir near Bonner. The river services including nutrient processing, has four west-side perennial tributaries hydrologic cycling, critical habitat for which flow from the Flint Creek Range and facultative (e.g., beavers) and obligate several small east-side tributaries, which (e.g., stoneflies) species, and multiple only flow during spring snowmelt. During socioeconomic functions for humans the summer months, water in the Clark Fork (e.g., water sources, fisheries). Moreover, River and its west-side tributary streams is because streams are typically highly sensitive diverted for irrigation (Smith et al. 1998), to stressors at both local and landscape some of which returns to the river channel scales, they are one of the most useful types as surface- or groundwater return flow. The of ecosystems for long-term ecological Clark Fork River generally divides GRKO monitoring. in half, with all the domestic buildings and Despite the many services they provide, buildings associated with cattle ranching streams are among the most significantly al- operations located to its east. The land on tered ecosystems in North America. Streams the west side of the river consists of pastures face numerous and varied threats, including and hayfields. The prominent bench along impacts from climate change, atmospheric the western edge of the ranch is an important deposition, altered hydrology, acid mine element of the parks viewshed. drainage, agriculture, pollution from boats, invasive species, erosion, improper sewage Why Monitor the Clark Fork River? plant or drain field operations, and storm There are several reasons behind the choice water runoff. Many of the day-to-day and to conduct long-term stream monitoring in long-term management decisions at GRKO GRKO. In general, streams and rivers are are at least partially connected to the Clark fundamental components of the ecological Fork. Fisheries, water chemistry, surface and cultural context of ROMN parks like water quantity, riparian composition and GRKO (Seastedt et al. 2004, Hauer et al. function are all often pressing matters for 2000, Hauer et al. 2007, Mast et al. 2005, park resource managers. Proper function- Stottlemyer et al. 1997). They are often what ing of the Clark Fork River and associated visitors come to see and what they remember riparian habitat is of great importance to when they leave, or especially in the case of NPS management since cultural and natural GRKO, form an important backdrop to the resources associated with the ranch can be historical events preserved by the park. The threatened by floods and erosion and the NPS recognizes that aquatic resources are River and riparian community are important some of the most critical and biologically components of the cultural landscape. important resources in the national park system and that they are vulnerable to The Stream Resource at Grant- degradation from activities both within and Kohrs Ranch National Historic Site external to parks. GRKO is situated along a 5.5 km reach of

Introduction 3 Figure 2. Grant-Kohrs Ranch National Historic Site showing the SEI sample site and USGS stream gauge location in Deer Lodge. The green squares show transect locations sampled during a full sample event, the red star gives the location of sentinel sample events and the blue triangle shows the USGS gauge. Inset shows the upper Clark Fork River watershed from the Blackfoot River near Missoula.

4 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Figure 3. Oblique view (looking North) of the Clark Fork River at Grant-Kohrs Ranch National Historic Site (red border) and the town of Deer Lodge, MT. Background imagery is early fall 2010 (courtesy Google Earth). Note the park boundary shown is an unofficial ‘fence line’ boundary. The green squares show transect locations sampled during a full sample event, the red star gives the location of sentinel sample events and the blue triangle shows the USGS gauge. The Deer Lodge water treatment ponds are visible in the upper portion of the ranch. Stream flow is from the foreground of the image. The inset shows the upper Clark Fork region with Anaconda, Butte, and the Milltown ponds (source of contaminated sediments) in the foreground. NPS/BILLY SCHWEIGER

Figure 4. The Clark Fork River in GRKO during a 2008 SEI sentinel water and sediment sampling event. Stream flow is from the foreground of the image.

Introduction 5 Current Importance of the Clark nitrogen might be leached to the shallow Fork to GRKO and the Deer Lodge groundwater system that may drain into the Valley Clark Fork. The Clark Fork River is also a The Clark Fork River is important at GRKO; key lifeline of the vast Deer Lodge Valley. It is the river and its associated hydrogeologic an important source of water for irrigation in system is a primary resource at GRKO the valley as well as a water supply for many and the park protects a small portion of towns, including Deer Lodge. It is also a the upper Clark Fork River drainage. It is valued source of recreational activities such integral to many of the ecological processes as fishing and rafting. occurring within the ranch, and it plays a Finally, as discussed below, the Clark Fork central role in current ranching operations. and its floodplain are also part of an ongoing The park has multiple water-right claims Superfund project to restore and mitigate to the use of surface water from the river impacts of acid mine drainage and heavy metal and its tributaries for flood irrigation of deposits (EPA 2004). This creates a complex park lands. GRKO obtains drinking water array ecological processes that SEI monitoring from the town of Deer Lodge municipal must attempt to quantify and a diverse regulatory system (supplied from the Clark Fork). arena with which the ROMN assists the park. Its wastewater is handled by the Deer Lodge wastewater treatment facility which Historical Role of the Clark Fork is located just inside the northern park boundary. From 2000 to 2008, the ranch River used secondary treated sewage effluent It was a beautiful stream, the water clear and from the wastewater treatment facility to sparkling and alive with the finest trout, and irrigate some of its hayfields during summer the same was true of every spring we crossed. months. For example, in June, July, and The valley was full of antelope and many head August of 2000 and 2001, 0.48 km2 of park of fat cattle belonging to the mountaineers who land was irrigated with 310 million and lived there. 650 million liters of effluent, respectively —Conrad Kohrs in 1862 upon arrival in (Johnson 2002). In 2000, the NPS funded Deer Lodge Valley a study to determine the fate of sewage The Clark Fork River played an important effluent nutrients (primarily nitrogen) role in the western ranch lifestyle that that were applied to soils in GRKO by a the Grant-Kohrs National Historic Site sprinkler irrigation system (Johnson 2002). commemorates as part of a living museum The study determined that although most of (Figure 5). The historic ranch headquarters the applied nitrogen was retained by plants are nestled in the Deer Lodge Valley along and in the soil, as much as 40 percent of the the Clark Fork River. There are six small

Figure 5. Robert Warren on Clark Fork River Bridge, ca. 1936. GRKO PHOTO #16054, Grant-Kohrs Ranch NHS Historic Archives.

6 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site tributary creeks and nine natural springs as complete hydrologic systems within a located on the ranch that feed into the watershed where the natural processes that Clark Fork River. This natural abundance of deliver water, sediment and woody debris fresh water, coupled with ideal topography are maintained and human disturbances and deep soils, provided the Grant, Kohrs, are minimized. Important natural processes and Warren families the natural resources for healthy stream system functioning needed to support their domestic and include natural processes of flooding, ranching activities for over 150 years. The stream migration and associated erosion expansive views and wide-open spaces of and deposition that results in natural habitat the Deer Lodge Valley played an important features such as floodplains, cutbanks and role in establishing the character of the river bluffs, riparian habitats, accumulations original Grant-Kohrs Ranch and continue to of woody debris, terraces, gravel bars, do so for the current historic site. The Flint riffles, pools, etc. ROMN monitoring aims Creek Range and the distant peaks of Mt. to provide data and understanding of the Powell and Deer Lodge Mountain provide Clark Fork River hydrologic system and a stunning backdrop to the rolling foothills these natural processes (Schweiger et al. In below. These mountain ranges contribute Review). While the regulatory considerations to the rugged feel and sense of isolation of summarized below also provide an important the ranch, and the snow-capped peaks of context for SEI monitoring, the emphasis the highest mountains easily remind valley of long-term ecological monitoring within inhabitants of the harsh and unforgiving the protocol is a key distinction from other conditions characteristic of Montana monitoring conducted by the State of Montana winters. The openness of the western hills or other federal agencies like the EPA. and pastures also contrast vividly with the human-scaled building clusters, corrals, and Regulatory Considerations fields that comprise the Home Ranch and Streams and rivers are one of the more Warren Hereford Ranch complexes. These regulated natural resources and multiple hills also contrast with the flat river benches, federal and state programs must be rich colors, and fine textures aligning the implemented by NPS and its partners Clark Fork River and provide the ranch with to protect aquatic habitat and its many a sense of enclosure. All these views are functions. The National Park Service essential components of the open, rugged, is required to manage all park units in and isolated character that defines the accordance with the Organic Act and other physical and cultural context of the Grant- laws so as not to be “in derogation of the Kohrs Ranch. values and purposes for which these various areas have been established...” (General Ecological Monitoring Authorities Act, NPS 1970). Under the Considerations General Authorities Act, all resources, The mission of the NPS I&M program and including water resources of a park, the ROMN is to provide long-term ecologi- are protected by the Department of the cal monitoring data and information for a Interior/National Park Service. Only an act suite of “Vital Signs” to park managers to of Congress can change this fundamental assist them in preserving park resources responsibility of the NPS. unimpaired for the enjoyment of current and future generations. The ROMN has imple- NPS Policies mented Stream Ecological Integrity monitor- The NPS Organic Act, the Code of Federal ing at GRKO to monitor several important Regulations, and NPS Management Policies Vital Signs including water quality and chem- (2006) broadly require park management istry, surface water dynamics, invasive/exotic to maintain, rehabilitate, and perpetuate aquatic biota and freshwater communities the inherent integrity of aquatic resources (Britten et al. 2007). and processes. NPS policies direct park managers to work with the appropriate Understanding and managing streams and partners (e.g., MT DEQ, and the EPA) to rivers is best accomplished by treating them obtain the highest possible standards for

Introduction 7 park waters under the Clean Water Act Interpreting ROMN Stream Ecological (CWA) and to maintain or restore water Integrity Monitoring Data quality. For smaller “flow through” parks like As emphasized in several places in this GRKO, this is complex given the small part report, the primary motivation for SEI of the river’s watershed within GRKO. monitoring is to understand and document the long-term ecological condition of Clean Water Act GRKO’s aquatic resources in order to better The CWA, first promulgated in 1972, protect and manage them. However, the is designed to maintain and restore the NPS does not have regulatory authority over chemical, physical, and biological integrity waters in GRKO—any regulatory action of the nation’s waters including waters in would be the responsibility of the State of national parks. As part of the act, Congress Montana. recognized the primary role of the states and tribal nations in managing and regulating For our purposes, we use criteria that water quality. Section 313 of the CWA we feel are most useful in the ecological requires that all federal agencies comply interpretation of SEI and associated with the requirements of state law for water auxiliary data (see below and Appendix A). quality management, regardless of other We use current State of Montana criteria, jurisdictional status or landownership. older criteria no longer in use by the The MT DEQ implements the protection State, criteria still in development by the of water quality for the waters of the state state, Federal water quality standards, and under the authority granted by the CWA novel reference points we develop from through best management practices and ecoregional reference sites. In most cases, we through water quality standards. Standards emphasize current standards from the State are based on the designated uses of a water of Montana as these are often conservative body or segment of water, the water quality and rigorous and have local applicability. criteria necessary to protect that use or Importantly, the ongoing Superfund project uses, and an anti-degradation provision to (EPA 2004) creates a complex array of protect the existing water quality. Water regulatory processes that SEI monitoring Quality standards are the foundation of must interact with. While the ROMN is not a water quality-based pollution control responsible for the implementation of the program as mandated by the Clean Water remediation at GRKO and the SEI protocol Act. Water quality standards define the is not optimized for understanding all of goals for a waterbody by designating its the consequences of the work, SEI data will uses, setting criteria to protect those uses, assist the park as this project unfolds. and establishing provisions such as anti- degradation policies to protect waterbodies Legal/Regulatory Status of the Clark from pollutants. The state’s anti-degradation Fork at GRKO policy is a tiered approach to maintaining Water use classifications for all stream seg- and protecting various levels of water quality. ments in Montana were established by the Minimally, the existing uses of a water State of Montana (Administrative Rules of segment and the quality level necessary to MT 17.30.611 to 17.30.614 as in MT DEQ protect the uses must be maintained. The 2002). The Water Resource Division of the second tier provides protection of existing NPS compiles data on protected uses and water quality in segments where quality any impairment status of waters in all NPS exceeds the fishable/swimmable goals of units. Table 1 presents this summary data for the Clean Water Act. The third tier provides GRKO as of 2012 (NPS 2012). protection of the state’s highest quality waters where ordinary use classifications The mainstem of the Clark Fork River may not suffice; these are classified as at GRKO (the reach from Cottonwood Outstanding Resources Waters (ORW). Creek to the Little Blackfoot River) has a beneficial use classification of C-1, defined as suitable for the following uses: primary

8 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site YES 303(d) in Watershed? NO NO NO NO NO NO NO NO NO - TMDL Complete? NO Waters? Outstand ing Resource ing Resource - C1 C1 C1 C1 Ben eficial Use Classification - - 0 ------line Miles Adjacent* 303(d) Im paired Shore paired ------Aquatic Life, Cold Water Fishery, Primary Contact Fishery, Aquatic Life, Cold Water Recreation Associated Uses Fishery Aquatic Life, Cold Water Primary Contact Fishery, Aquatic Life, Cold Water Recreation Primary Contact Fishery, Aquatic Life, Cold Water Recreation Primary Contact Fishery, Aquatic Life, Cold Water Recreation Fishery Aquatic Life, Cold Water Fishery Aquatic Life, Cold Water Fishery Aquatic Life, Cold Water Fishery Aquatic Life, Cold Water 0.97 Shore line Miles Adjacent* 0 ------303(d) Impaired Impaired Adjacent* Lake Acres Lake Acres ------19.45 Adjacent* Lake Acres Lake Acres 0 0 -- 0.043 303(d) Impaired Impaired Adjacent* Stream Miles Stream 0 -- Mill Tailings Probable Sources Probable Channelization Agriculture, Mill Tailings Mill Tailings Agriculture Discharges Municipal Point Source Agriculture, Discharges Municipal Point Source Agriculture, Channelization Agriculture, Agriculture 0.059 0.101 303(d) Adjacent* Stream Miles Stream . See NPS whenever possible to verify adjacent features utilized were staff, land segment maps and contacting park as reviewing of information, such . Other sources 0 0 -- 3.45 303(d) Impaired Impaired Stream Miles Stream -- 4.85 3.22 3.63 Stream Miles Stream DEQ summary of impairment of impairment summary (b) MT DEQ status and impaired extent of of compilation NPS (2012) of 2012; (a) GRKO as waters in status of of regulatory Summary Type Waterbody Waterbody Phosphorus (Total) Physical substrate habitat alterations Sedimentation/Siltation Zinc Copper Lead Low flow alterations (Total) Nitrogen Probable Impairment Causes Probable or littoral vegetative covers Alteration in stream-side Intermittent Stream/River Canal Lake/Reservoir Perennial Perennial Stream/River (2012) for details. (b) It is important to note that identifying adjacent hydrographic features may include some level of subjectivity and require a level of subjectivity and require may include some features identifying adjacent hydrographic a boundary with a park. It is important to note that that shares water feature defined as any *Adjacencies were processor of judgment by the certain degree Table 1. Table designation ORW and 303(d) waterbodies. boundaries or to adjacent* park’s to within the numbers are restricted DEQ 2012b). Extent TMDL status (MT causes and types. all flowing water applies to (a)

Introduction 9 contact recreational use (bathing, swimming, (copper, cadmium, lead, and zinc) related and recreation); growth and propagation to historic mining, milling, and smelting of salmonid fishes and their associated activities in the headwaters of the Clark aquatic life, waterfowl, and furbearers; and Fork River (Figure 6). Tailings from these agricultural and industrial water supply. operations were transported downstream This reach was on the 2012 303(d) list for and deposited along the floodplain of the the State of Montana (MT DEQ 2012b) river by at least six major floods during for not supporting aquatic life, cold water the late 1800s and early 1900s (Smith et fisheries, and primary contact recreational al. 1998). Sedimentation ponds were later use. Sources of impairment included built to prevent movement of tailings from mill tailings, which have caused elevated the headwaters; however, the Clark Fork copper, lead, and zinc concentrations, and continues to be contaminated in two ways: agriculture and municipal point-source by erosion of previously deposited tailings, discharge, which have resulted in elevated and by direct leaching of metals from tailings nitrogen, phosphorus, and sediment into surface and ground water. Copper is concentrations. In addition, flow in the the primary metal of concern because it is river during August and September is the most toxic to aquatic life and plants in often less than the minimum necessary the river floodplain. Many of the probable for maintenance of a salmonid fishery. sources of impairment listed in Table 1 are Contaminated sediments and siltation also likely due to these events. have caused partial impairments for primary contact recreational use. As of 2012, the In April of 2004, a Record of Decision status for many TMDLs on tributaries to (ROD) was issued by the U.S. Environmental the Clark Fork as required to address these Protection Agency (EPA) for the remediation impairments are complete with the main of contaminated areas within Reach A stem in progress and a high priority (MT of the Clark Fork River, which includes DEQ 2012c). GRKO (EPA 2004). General remediation efforts will include removal of larger tailing 2 Clark Fork Superfund Project deposits (>37 m in area and 60 cm in depth) GRKO is within the Clark Fork River followed by landscaping/top soil addition Superfund Site, an area which is and finally re-vegetation of the excavated contaminated with arsenic and heavy metals area. Smaller tailing deposits and impacted

Figure 6. Clark Fork Operable Unit. GRKO is located next to the town of Deer Lodge. Map courtesy of EPA (2009b).

10 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site soil and vegetation will be remediated in monitoring to demonstrate that over time the situ using lime addition, soil mixing, and remediation activities are achieving the goals re-vegetation. Consultants for DEQ will that were set forth in the Clark Fork Opera- begin a draft remediation design for GRKO ble Unit Record of Decision (EPA 2004) and in 2012. The NPS has specific cleanup in the Monitoring and Maintenance Plan needs and responsibilities under the laws for the Clark Fork Operable Unit (MT DEQ that govern national historic sites. The ROD 2008b). ROMN SEI monitoring is not meant was expanded and modified specifically to monitor the effectiveness of ongoing re- for GRKO. There are many comprehensive mediation and restoration of the Clark Fork sources for information on this project and channel and riparian habitats within GRKO. this document is not intended as a complete SEI monitoring of the Clark Fork River does treatment. not replace any research remediation and restoration effectiveness monitoring or other The state of Montana has conducted work work being undertaken by the state or other in the basin focused on the Superfund federal agencies under the auspices of the project since at least 1984. The USGS has Record of Decision or the state monitoring also conducted focused work in the basin, plan. above and beyond the usual background monitoring done by USGS district offices Climate Change (i.e., Dodge et al. 2009). Finally, of special NPS Vital Signs monitoring is a long-term interest to SEI monitoring, McGuire (2010, effort. Therefore, it should effectively 1993) details long-term monitoring of elucidate long-term changes in stream macroinvertebrates in the Clark Fork Basin condition given shifting climate regimes. conducted for MT DEQ. Multiple sites were Changes in stream hydrology in the West monitored for macroinvertebrates over over the past fifty years due to climate varying periods of record including one site changes are well documented (Barnett et just upstream from GRKO. Methods for this al. 2008) and these changes are expected to study were comparable to the SEI protocol continue. Physical, chemical, and biological (although targeted macroinvertebrate characteristics of aquatic ecosystems are well sampling was used, specific to select documented to act as indicators of impacts macrohabitats in a sample reach, likely due to climate change (McKnight et al. resulting in less representative samples). 1997). Even modest temperature increases Ten metrics were selected for use following in the western United States may cause guidance from the EPA (Plafkin et al. 1989). significant changes to the hydrologic cycle, Although there is some overlap in the as manifested in earlier snowmelt, earlier metrics used in SEI, it is unclear why current ice-out on lakes, reduced summer base flows MT DEQ-specified metrics and thresholds (Dettinger et al. 2004), a lower snowpack were not used within the USGS work. volume at lower to mid elevations (Knowles Nevertheless, general conclusions from et al. 2006), and increased flooding due to McGuire (2010, 1993) are likely comparable rain-on-snow events in winter (Heard 2005). to SEI results and provide a useful long Hydrographs (i.e., the magnitude and timing term context for interpreting trends in the of spring run-off) will likely shift to earlier bioassessment of the Clark Fork. floods. These changes will, in turn, likely MT DEQ (2010b) presents an interim affect the seasonal dynamics of stream and comprehensive long-term monitoring riparian biota (Palmer and Bernhardt 2006). plan for remediation actions implemented While more proximate, classic stressors within the Clark Fork River Operable Unit impact many ROMN streams (especially of the Milltown Reservoir/Clark Fork River at GRKO), climate change may be the most Superfund Site. The monitoring plan de- important direct and indirect stressor over scribes the requirements for: (1) annual the long term. Many of the effects of cli- monitoring to ensure that remediation and mate change may take some time to appear, restoration projects are meeting established however, and most will be confounded with performance standards, and (2) long-term other sources of stress. Therefore, the SEI

Introduction 11 protocol is designed to measure a broad impacts on community structure and can spectrum of indicators across multiple time alter soil and water chemistry. Moreover, scales such that we can attempt to tease apart invasion into riparian areas is often a useful patterns of ecological response to multiple indicator of general disturbance (often stressors. anthropogenic at least in part). Target invasive plants were developed in concert Aquatic and Riparian Invasive with staff from all ROMN parks. These lists Species include taxa that are ecologically varied, easy Climate induced changes in the hydrological to identify, truly invasive, ecologically or cycle and multiple proximate direct distur- economically intrusive, and some of which bances often lead to changes in species dis- are also included in regional surveys like tribution including invasion by taxa that can EMAP (Ringold et al. 2008). threaten native or endemic species and the stability of aquatic ecosystems (Rahel and Monitoring Objectives Olden 2008). We include simple methods The general goals for long-term ecological that quantify the presence and abundance of monitoring of the Clark Fork at GRKO a select list of invasive taxa developed with focus on documenting the status and trend park staff. These methods should not replace in condition, understanding the causes more specific protocols that may be imple- of change in condition, and assisting mented by a park or other collaborators for in the application of SEI results and other or more specific purposes. relevant auxiliary information to resource management. Statistically rigorous estimates The Montana Department of Fish, Wildlife of trend are likely possible after five cycles of and Parks (MT FWP) maintain a list of sentinel data collection have been completed species of concern in the state (MT FWP at a sentinel site. Simpler estimates of change 2012). Most of these taxa are vertebrates are conducted after the first two to four (that we do not collect) or plants (for which cycles. we use alternate citations, see below). Invasive aquatic invertebrates on the list that 1. Determine the condition (status) on five- could occur at GRKO include the zebra and year intervals of the Clark Fork at the quagga mussel (Dreissena polymorpha and GRKO sentinel site using bioassessment Dreissena rostriformis), the New Zealand (macroinvertebrate and diatom mudsnail (Potamopyrgus antipodarum), assemblages), physical habitat, and and the rusty crayfish Orconectes( rusticus). water/sediment physiochemistry. The mudsnail has established populations in south west Montana. Mudsnails tolerate 2. Determine the percent change in the siltation, thrive in disturbed watersheds, condition of select responses detailed nd th and benefit from high nutrient flows. They in objective 1 after the 2 -4 cycle of can impact trophic dynamics of native trout sampling. fisheries and alter the physical characteristics 3. Determine the long-term trend in the of the streams themselves. While not on condition of select responses detailed the list, Didymosphenia geminata (didymo), in objective 1 after the 5th and each a freshwater alga native to North America subsequent cycle of the sampling. may be of concern for GRKO. Didymo has long been considered a cold-water species, 4. Determine long-term trends in select but has recently spread to lower latitudes water physiochemistry at the nearby and warmer waters, and increasingly forms upstream USGS gauge in Deer Lodge, large blooms that cover streambeds and may Montana. change species composition. 5. Relate any spatial or temporal patterns Riparian corridors are often ideal habitat (including trend when possible) in select for invasive plants. These species may have responses detailed in objective 1 to

12 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site important ecological and anthropogenic c. Compare to baselines derived from drivers. partner or auxiliary data or, after 5 repeated sample cycles, SEI data. These models depend on scale- appropriate These are developed for a subset covariate SEI and auxiliary data, with clear of SEI responses based on data (and often causal) connections to SEI availability and quality and the responses and are therefore done on an as relevance of the response to Clark possible basis. Fork management.

6. Assess these responses as follows: Several monitoring objectives specific a. Compare to existing and published to management needs at GRKO may be assessment points, including any included, to varying degrees, within ROMN regulatory criteria, ecological SEI monitoring as the protocol matures. The thresholds or ecoregion assessment relevance of these to SEI monitoring and points that are relevant to the Clark the capacity of the Network to successfully Fork at GRKO and its management. include them in the core SEI protocol will require careful consideration. The ROMN b. Compare to ecoregion assessment will work with park management to help points derived from distributions interpret and apply SEI data to GRKO of reference site data from the management needs. Likely applications surrounding ecoregion. These include: ongoing Superfund restoration, are developed for a subset of SEI mine reclamation work with the U.S. Forest responses based on data availability Service on Cottonwood Creek, irrigation, and quality and the relevance of the bull trout (Salvelinus confluentus), and other response to Clark Fork management. salmonid management, beaver and osprey management, grazing management, and understanding fire impacts in the riparian corridor.

Introduction 13

Methods

The following sections present brief Temporal scales of inference may be more summaries of how and why sites were limited for many responses (especially chem- selected (sample design), why particular istry) and may only represent conditions at responses were selected and how data the time of sampling. In addition, rare and were collected in the field. For a complete short-term events such as floods typically are description see Appendix A and the SEI not captured in our water physiochemistry protocol narrative and associated SOPs measures. However, water and sediment (Schweiger et al. In Review). chemistry samples are routinely collected several times per year in order to build a Sample Design database that represents the range of condi- SEI data collection at GRKO occurs within tions that occur at the sampled site(s). Other an approximately 500-meter sample reach parameters may be more readily inferred defined by series of subsample locations. The to a longer timeframe. Specifically, while reach extends from near the main bridge in biological responses are collected at base the middle of the park to the park’s southern flow (as this is when most species are at a boundary (see Figure 2 and Figure 3). The more identifiable life-stage and potentially sample design used in GRKO SEI monitoring more stressed by stream flow levels) these was a hand-picked (not randomly selected), responses often integrate longer time peri- ROMN sentinel site approach (see Britten ods (i.e., lifespans or the time required for an et al. 2007) for a complete overview of assemblage to develop). ROMN sample design strategies). This site was selected both for access and because it Sample Events includes the uppermost reach of the river in Events at sample sites are of two types. First, the park where surface water first flows into full sample events typically include all meth- the system. The ROMN does not sample any ods (water/sediment chemistry, habitat, and of the tributaries or small ditches that have a biology) and are done at base flow, once per confluence with the Clark Fork in the ranch year. Second, sentinel sample events include or any ground water resources. only water/sediment chemistry. They are conducted on key limbs of the hydrograph Because the sample reach was not selected (rising, falling, etc.). We generally conduct at random, in a strict sense the inference three sentinel events at each sentinel site of SEI monitoring is limited to this reach per year (the main event at base flow also or even select points within it. However, includes water/sediment chemistry). we are sampling a sizable portion of the river in the park (the total extent of the Due to high flow and unsafe conditions in mainstem of the Clark Fork within GRKO 2008, biological samples (macroinvertebrates is only approximately 5.5 km) and we feel and periphyton) were collected using an ad our sample reach is largely representative of hoc arrangement of 11 subsample locations this extent. Therefore, we generally interpret spread across a variety of macrohabitats both results at the scale of the park. We do upstream and downstream of the GRKO include data on select drivers and covariates bridge. In 2009, safe conditions permitted across larger scales as applicable to help us to conduct a complete sample event. understand SEI data in a larger context. Because many of the same methods (net This is accomplished largely through (1) mesh size, time of each sample, periphyton analyses that include landscape-level and collection details, etc.) were used in these climatic drivers of stream condition and (2) sample events, we feel that the samples are our use of auxiliary data from other sources sufficiently similar and treat the two data (especially the nearby USGS stream gauge in sets equivalently in most interpretation. Deer Lodge) that have a broader temporal Because many of the same methods (net and spatial extent and/or inference. mesh size, time of each sample, periphyton

Methods 15 collection details, etc.) were used in these Lake Biological Station (FLBS; R. Hauer, sample events, we feel that the samples are pers. comm. 2007), U.S. Forest Service sufficiently similar and treat the two data sets (Heitke et al. 2011), NPS Water Resource equivalently in most interpretation. Division (WRD; Irwin 2006), several USGS approaches (e.g., Fitzpatrick et al. 1998), and Field Data the MT DEQ (MT DEQ 2012a). However, Stream monitoring methods are relatively SEI methods are not perfect duplicates of well established and we draw upon this any of these other protocols and there may wealth of knowledge for the ROMN SEI be subtle or non-quantified artifacts from protocol used in GRKO. All SEI laboratory methodological differences in our data. methods follow standard and accepted Where we feel this might have a meaningful protocols implemented by each individual impact we qualify results as necessary. lab and include detailed quality assurance The SEI protocol includes a broad spectrum and control methods. The application of a of indicators to help evaluate the ecological standardized monitoring protocol across integrity of the Clark Fork (Table 2). We sites facilitates comparison of streams and include measures of drivers (or stressors) rivers within an ecoregion. Sources of SEI and ecological responses, such that we may methods include the EPA Environmental better understand the linkages between Monitoring and Assessment Program these and help parks apply these results to (EMAP; Stoddard et al. 2005), Flathead resource management. Core indicators

Table 2. Summary of SEI field data. Indicator Class Summary In Situ temperature (synoptic and continuous), water; oxygen, dissolved; ph; specific conductance.

Major ions alkalinity, dissolved; calcium, dissolved; chloride, dissolved; fluoride, dissolved; hardness , dissolved; magnesium, dissolved; potassium, dissolved; silica, dissolved; sodium, dissolved; sulfate, dissolved; total suspended solids.

Nutrients ammonia, dissolved; nitrite + nitrate, dissolved; nitrite, dissolved; nitrate, dissolved; nitrogen, total; orthophosphate, Water and Sediment Physiochemistry dissolved; phosphorous, total; carbon, organic, dissolved; carbon, organic, total; chlorophyll-a in periphyton; ash free dry mass in periphyton.

Metals in water aluminum, arsenic, barium, beryllium, cadmium, chromium, copper, iron, lead, manganese, selenium, zinc.

Metals in aluminum, total; arsenic, total; barium, total; beryllium, total; sediment cadmium, total; chromium, total; copper, total; lead, total; iron, total; mercury, total; selenium, total; silver, total; zinc, total.

instantaneous q continuous q (from nearby USGS gauge) thalweg profile channel geomorphology woody debris tally Physical Habitat channel dimensions substrate quantification fish cover riparian vegetation human influence assessment of channel constraint, debris torrents, and major floods

Benthic Macroinvertebrates reach-wide habitat (composited across systematic transect array) sample

Periphyton reach-wide habitat (composited across systematic transect array) sample

16 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site sediment physiochemistry (i.e., nutrients, derivatives) is used by several other long- metals, temperature); physical habitat (i.e., term monitoring programs (state and substrate composition, riparian cover, federal). channel geomorphology); and community level assays of two important biological Auxiliary Data assemblages, macroinvertebrates and algae. We use several data sets not generated The latter provide an integrated aspect within the SEI protocol to compare to for assessing ecological integrity because GRKO data and enhance our ability to these organisms respond to environmental characterize the Clark Fork at GRKO. We conditions over time and space. also develop assessment points used to informally interpret SEI results from some We present summaries of field of these auxiliary data sets (see below). data collection methods, details on Most of these data come from monitoring instrumentation and brief justifications for efforts conducted by state and other federal the major responses we collect in Appendix agencies as described below. Future SEI A. The full protocol is in Schweiger et al. (In reports will also include data from the Review). We do include some detail below National Rivers and Stream Assessment regarding auxiliary data harvested from (NRSA) sampled in 2008 and 2009 (EPA other monitoring programs and compared 2007). Analytical methods, including to GRKO SEI data to assist in interpretation. assumptions within comparisons between Readers unfamiliar with the SEI protocol or SEI and auxiliary data, are summarized in a stream monitoring in general should review subsequent section. Appendix A. Subsequent narrative in this report assumes that the reader is familiar Water Physiochemistry and with basic stream monitoring methodology. Streamflow Given the proximity of the USGS gauge in Site Layout Deer Lodge to GRKO (USGS 2012a), and the The design of the sample reach in GRKO amount of high-quality data available from follows the EPA EMAP site layout (Peck USGS, we elected to harvest, analyze, and et al. 2006) as modified by the ROMN interpret a large amount of USGS data in this SEI protocol (Schweiger et al. In Review). report to assess trends. Data were retrieved Eleven cross-section transects, used for from the National Water Information physical habitat characterization and System (NWIS) database for the Deer Lodge biological subsampling, are placed at station over variable periods of record from equal intervals along the reach (50 m apart 1969 through 2010 for 14 nutrient, trace along the thalweg; Figure 7). Point samples element, and major ion parameters. Daily for sediment, benthos, and periphyton measurements of stream discharge and collection are set up at middle, left, or stream water temperature were also acquired right locations on each of these transects. from May 2001 through April 2010. Only Riparian plots measuring 10x10 m are data publically available through NWIS established on the banks at each transect for and thus passing all USGS QAQC checks collection of riparian vegetation attributes. were used. However, following Mast (2007) Finally, a thalweg profile is established in we elected to retain trace-element results, the deepest part of the channel along the although it is documented that dissolved entire sample reach. Water and sediment concentrations in USGS samples collected chemistry is collected at the downstream prior to 1992 may have been contaminated end of the sample reach or at the nearby during sample collection and processing bridge. This sample reach length captures (USGS Office of Water Quality Technical repeating patterns of local habitat structure Memorandum 91.10; USGS 2012b). and biological variability associated with riffle-pool structure and meander- We also used grab sample water chemistry bend morphology in the Clark Fork (see data collected from 113 sites (12 degraded, Kaufmann et al. 1999 and references 35 reference, and 66 with unknown or within). This general site layout (or similar

Methods 17 Figure 7. ROMN SEI sample reach in oblique profile. Stream flow is from the top to bottom of the figure. Cross- sections, sub-sample locations, riparian plot, and thalweg profile all shown. Inset pictures show select details for substrate, channel and riparian sampling.

intermediate condition) sampled as part of the state of Montana because assessment the EMAP Western Pilot Project (Stoddard points we develop from these data are et al. 2005) from 2000 to 2004 within the informal, are not used by MT DEQ to assess Middle Rockies ecoregion (Figure 8, in part). stream condition and source data was Note that the specific sample size available available from a larger region (unlike some from these data sets for a given parameter of the biological data sets described above). varied and are listed with results given Rather, we used sites from the complete below. Chemistry methods used by EMAP ecoregion, regardless of what state the site were nearly identical to those within the SEI was in. This included sites in Montana, protocol. We did not restrict habitat data to Idaho, and Wyoming.

18 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Biological For biological data (in this report, limited to macroinvertebrates) this includes (Figure 8): 42 sites (1 degraded, 16 reference, and 25 with unknown or intermediate condition) sampled as part of the EMAP Western Pilot Project (Stoddard et al. 2005) from 2000 Figure 8. MT DEQ, EPA, and STAR monitoring sites in the to 2004, 17 sites in larger rivers similar to Middle Rockies ecoregion within Montana as used in the Clark Fork (2 degraded, 2 reference, comparison to select ROMN SEI results for GRKO. Sites and 13 with unknown or intermediate are classified as reference, degraded or unknown (by the programs responsible for each site) with green showing condition) sampled by MT DEQ from 2001 reference sites, red degraded sites, and orange unknown to 2005 (Bollman 2006), 64 (all reference) (“test”) sites. Inset shows all sites across the state with sampled by the state as part of an on-going symbology based on the source program of each site. reference project (Suplee et al. 2005), and Note that additional sites in the ecoregion but outside of 74 sites (31 degraded, 25 reference, and 18 Montana are not shown here were used for select habitat with unknown or intermediate condition) and chemistry parameters. sampled from 2000 to 2003 by Utah State University as part of the EPA Science to external source data had to be from the Achieve Results (STAR) program (Hawkins Middle Rockies ecoregion in Montana, et al. 2003). Three of the sites are treated have a multi-habitat sampling method, (for in multiple sources (i.e., a few EMAP sites benthos) a minimum 300 organism lab are also MT DEQ reference sites) reducing count (for SEI we count at least 600) and a the total unique sample size compared to minimum of genus-level identification of GRKO SEI data to 194. Note that the specific insects, (including Chironomids). Jessup et sample size available from these data sets al. (2006) and MT DEQ (2012a) show that, for a given metric varied and are listed with for at least the current bioassessment tools results given below. Field and analytical used in this report, models using MT DEQ methods for these programs were largely or EPA methods (on which the SEI protocol comparable to SEI protocols. To be included, is based) result in similar interpretations.

Methods 19 We restricted auxiliary biological data to the quality of SEI indicators. The ROMN sites in the state of Montana because much is implementing this strategy in our large of the source data were only from Montana parks (currently, Glacier NP for streams and streams and rivers. Rocky Mountain NP for wetlands), but in GRKO we are not able to sample multiple Habitat independent sites due to the restricted We used habitat data (Figure 8, in part) extent of the stream resource in the park from 133 sites within the Middle Rockies and ROMN budget constraints. A similar ecoregion (12 degraded, 35 reference, and approach might be used with replicates at 66 with unknown or intermediate condition) a single or few sites, although this presents sampled from 2000 to 2004 as part of the some issues with pseudoreplication and EMAP Western Pilot Project (Stoddard et temporal autocorrelation. al. 2005). Note that the specific sample size available from these data sets for a given Signal to Noise Ratio parameter varied and are listed with results One way we are overcoming the issue of given below. Habitat data collected by the limited independent replicates is to borrow EMAP program used a near identical field relevant results from similar monitoring and analytical protocol as the SEI protocol. across a broader region. We use a relatively We did not restrict habitat data to the state simple statistic generated for EMAP data (see of Montana because assessment points we above; Stoddard et al. 2005; P. Kaufmann, develop from these data are informal, are not U.S. EPA, Pers. Comm, summer, 2012) called used by MT DEQ to assess stream condition “signal to noise” (S:N). Signal to noise is and our source data were not restricted to the ratio of variance between multiple sites Montana (like some of the biological data (signal) and the variance of repeated visits sets described above are). Rather, we used to the same site (noise). Loosely speaking, sites from the complete ecoregion, regardless it is a measure of the repeatability of a of what state the site was in. This included measure. A low S:N value indicates that a sites in Montana, Idaho and Wyoming. metric has nearly as much variability within a site (over time) as it does across different Quality Assurance and Quality sites, and thus indicates a measure that does Control not distinguish well between sites. A high Two of the more important requirements S:N value suggests a response that has lots of long-term monitoring program are to of signal or real information. S:N is derived assure: (1) that data are of high quality, and using a general linear model to compare (2) that response measures and metrics are the within-year variance among streams valid estimators of the endpoints of interest (signal) with the variance between repeat that facilitate meeting monitoring objectives. visits within the same year (noise). The noise The ROMN SEI protocol (Schweiger et al. In variance includes the combined effects of Review) and associated SOPs including the within-season habitat variation, differences ROMN Quality Assurance Performance Plan in estimates obtained by separate field crews, specify various field and laboratory quality and uncertainty in the precise relocation of assurance and quality control (QAQC) plots. procedures (summarized in Appendix D) We report S:N values from data sampled that address the first of these. during 2000-2004 by EMAP at 1524 sites The implementation of the SEI protocol across the western U.S.. We do this for at GRKO presents some challenges for select SEI biological and habitat responses meeting the second QAQC requirement. collected at GRKO where we use the same Ideally, we would have a large sample size methods as EMAP. Because this approach of independent sites on the Clark Fork assumes that SEI crews perform identical to in GRKO with these well replicated data EMAP crews and that there is no systemic allowing estimation of sources of variation difference in the years used (2000-2004 vs. in stream condition to statistically determine 2008-2010), we only use the EMAP S:N as rough guides. Note that one of the EMAP

20 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site sites in this data set is on the Clark Fork, near we report a minimum, maximum, and GRKO (see Figure 8). median value summarized across all events during 2008-2010. We use a median (rather Analyses than a mean) because of lower bounds The following sections present summaries of zero, censored or imputed non-detect of how GRKO SEI data were analyzed and values, outliers and the often very skewed interpreted (for complete descriptions distributions typical of these data (Mast see Schweiger et al. In Review). We first 2007). describe the statistics and models we use to Analyses of water chemistry samples by estimate the status (single point in time or laboratories often include “non-detects” summaries over short time periods) in the or values below the ability of a given lab’s condition of the Clark Fork during 2008- capacity to accurately resolve (see Appendix 2010. We then summarize how we estimate G for laboratory detection limits). These trend in condition (using auxiliary data present challenges for analyzing and from the USGS). Next, we summarize how interpreting data. We use a regression on these results are interpreted or assessed. order statistics (ROS) imputation method Importantly, we use multiple approaches from Helsel (2005) for all censored lab data. to interpreting SEI data including a broad The ROS method resolves nondetects on spectrum of current and historic metrics, the basis of a probability plot of results from informal comparison to regulatory criteria a larger dataset (we used all data within the (because NPS is not the ultimate regulatory ROMN SEI database including data from authority for water bodies in LIBI or any GLAC and LIBI) to better estimate the NPS unit), and derivation of and then nondetect values (but summary statistics comparison to unique or proprietary presented here are only based on GRKO assessment points. Our approach to samples). interpreting results is different from what regulatory agencies like MT DEQ or EPA might employ. It is in keeping with the NPS Stream Temperature mandate to manage ecosystems within parks Stream temperature data are available as following NPS guidance and the founding a continuous time series from the nearby purpose of GRKO. We are careful to qualify USGS gauge in Deer Lodge from 2001 these methods and to distinguish them to 2010 and from SEI loggers in the park from more formal treatments by regulatory from 2008 to 2010. Various intervals of agencies. Finally, we present how we assess temperature data from these loggers are the quality of SEI data and how we apply used to 1) estimate potential baselines for information from similar monitoring comparison to future stream temperatures, (Stoddard et al. 2005) that may roughly 2) explore short term patterns in the approximate the precision of some of the phenology of freeze-thaw cycles, and 3) indicators used at GRKO. estimate seasonally-adjusted models of trends in water temperatures (further Status described in the section below on trend analyses). We present a subset of these data We estimate status using simple descriptive using time series plots by water years from statistics of core SEI responses or various 2008 to 2010. To help understand simple modeled metrics derived from raw data. patterns in these data relevant in their Results are presented within tables and a interpretation as a baseline, we compare variety of graphics. For many responses, temperature and USGS discharge from we do not have either spatial or temporal the Deer Lodge gauge from 2001 to 2009 replicates and so we cannot yet report using a simple no-parametric correlation. estimates of variance. We restrict data in these models to spring and summer (March 1 through August 31) Water and Sediment Physiochemistry to focus on patterns in ecologically more For synoptic or instantaneous in situ and relevant runoff through base flow portion of grab samples of water physiochemistry the hydrograph.

Methods 21 We estimate mean August stream bed stability. Over 100 habitat responses temperature as possible baseline statistics collected in the field are reduced to dozens from daily values at the USGS gauge from of metrics. Using guidance in Kaufmann the 2001 to 2009 water year. We use August et al. (1999), Stoddard et al. (2005), and temperatures because this is generally the USFWS (1998) we further reduced the set hottest month of the year and may be when interpreted here to 33. warming trends are most evident as climate regimes shifts over time. However, the At the time of this reports creation we only choice of a specific temperature statistic had a single full sample event (in 2009) that to use as a baseline is complex and we are included habitat. Therefore, we can only exploring alternatives (Dettinger et al. 2004, report single values for habitat metrics and Isaak et al. 2011, Arismendi et al. 2012) until we accrue more habitat data we have including mean or maximum mean spring no estimates of the spatial or temporal water temperatures (as a seven-day rolling variability in habitat structure at GRKO average of daily data; Isaak et al. 2011) which (some metrics are expressions of within may be useful given shifts in runoff and reach spatial variance in a habitat response as other aspects of a streams hydrograph with measured during a sample event). However, changing climate. Moreover, as discussed as discussed below, we include estimates of below, thresholds for bull trout based on signal to noise (S:N; Kaufmann et al. 1999) temperature do exist (USFWS 1998, 2002) in each metric using 2000 – 2004 EMAP data and to better understand the biological from across the Middle Rockies ecoregion context of stream temperatures at GRKO we (Stoddard et al. 2005). These provide some qualitatively compare mean daily values and context for how variable each metric is over mean seven-day maxima from 2008 to 2010 space and time. This improves our ability to to these thresholds. interpret habitat condition as well as offer insight into which metrics perform the best. We summarize the phenology of freeze-thaw Given the near identical habitat methods cycles by estimating the last (or a thaw) and within the SEI and EMAP protocol these first day of freezing in each water year. We values do provide a sense of how variable define a thaw or freeze as any period of 2 each metric is across the ecoregion and days or more with temperatures above or provide some context for the GRKO SEI below 0.5ºC. We compare these dates across results. ad hoc historic (2001 to 2006 water years) and current time periods (current begins Hydrology: Stream Flow with SEI monitoring in 2007) using non- Using the Climate Data Summarizer (version parametric Kruskal-Wallis tests. 3.4; Walking Shadow Ecology 2012) we create hydrographs by water year with data Physical Habitat from the USGS gauge in Deer Lodge from Habitat Metrics 2007-2010. These figures include discharge The large number of raw physical habitat during the water year, the median discharge measures collected during a full SEI sample over the historic period of record (1980- event are summarized into synthetic metrics 2010) from the USGS Deer Lodge gauge, and expressed at the sample reach scale. and the midpoint of the annual discharge SEI habitat sampling within a reach is for these two time intervals. We summarize systematic and this feature lends itself to the total discharge across each water year, calculating representative summaries of the including maximum flow, the date of habitat characteristics of the sample reach. maximum flow, annual median flow, and Habitat metrics include simple summaries its standard deviation. Finally, a statistical of raw data, areal cover estimates from cover analysis of annual peak flow data from the class data, proximity weighted estimates, period of record for the Deer Lodge gauge and more detailed specific approaches was performed using Hydraulic Engineering for calculating woody debris abundance, Center-Statistical Software Package (HEC- residual pool characteristics, sinuosity, and SSP 2.0; USACE 2010). This tool generates recurrence intervals (from 1 to 500 years,

22 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site also known, for example, as a “100 year DEQ River Invertebrate Prediction and flood event”), and the range in probabilities Classification (RIVPACS) modeling (from 0.2 to 99) of exceeding the annual system. RIVPACS models are calibrated peak streamflow. with observations made at reference- quality sites across a region and the models Biology relate taxon occurrences to multiple environmental gradients. In effect, the Bioassessment The primary way in which we analyze approach simultaneously models the niche biological data is via assemblage or relationships of many taxa and predicts the community-level metrics generated from specific taxa that should occur at a site given bioassessment models. This follows a its natural (i.e., reference) environmental long tradition in stream monitoring and characteristics. We estimate and interpret assessment (e.g., Kolkwitz and Marsson two statistics from RIVPACS, the simplest 1908) known as bioassessment (Barbour being the ratio of Observed to Expected taxa et al. 2000). Bioassessment assumes that and the second a Bray-Curtis dissimilarity the composition of biological communities score. Note that MT DEQ uses the term reflects the overall ecological integrity of “Observed/Expected models” to describe a system. Evidence suggests it may detect their use of RIVPACS but we elected to stressors that other approaches fail to reveal retain the more general name “RIVPACS” (Karr and Dudley 1981, Karr and Chu 1997). given our derivation of multiple statistics from the model. MT DEQ and other ROMN partners have developed a variety of bioassessment For diatoms, we use two approaches: models to help monitor stream and river increaser models and a suite of individual condition over the last two decades and metrics that describe specific aspects of we use a subset of these (Table 3). We community structure. Diatom increaser emphasize models used by MT DEQ (as of models summarize the relative abundance 2011) given their higher levels of precision. of diatom taxa that, as a group, exist in However, we feel that many older metrics detectable amounts and demonstrate a or those that may be more regional in scale meaningful, measurable, and significant have general interpretative value for our response to specific stressors (sediment, purposes, and we include select examples nutrients, or metals). We employ both of these in this report. Where useful, we do Middle Rockies ecoregion specific increaser include interpretation of individual taxa (i.e., models (for sediment; Teply and Bahls 2010a, invasive species or taxa that are indicative of b) and Mountains/Foothills ecoregion a particular aspect of stream condition). increaser models (for nutrients and metals; Teply and Bahls, 2005). For additional details For benthos, we focus on two approaches: see the references above, Appendix B and First, we derive scores using various Schweiger et al. (In Review). iterations of a Multimetric Model (MMI; also known as an Index of Biotic Trend Integrity). MMI models combine multiple Given the short time the SEI protocol has characteristics of a macroinvertebrate been implemented at GRKO, we cannot assemblage that change in some predictable perform any meaningful trend analyses with way with increased human influence that SEI data. When we have a sufficient record alters environmental conditions. They (another 3-5 years), many of the response have a long history of successful use in measures and metrics detailed above will bioassessment (however, beginning in be analyzed using trend models. USGS data 2012 Montana DEQ no longer employed from the nearby gauge in Deer Lodge do a macroinvertebrate MMI in biological allow select trend models as summarized assessments). We include component below. metrics of the most current MT DEQ MMI to aid in its interpretation. Second, we apply the 2011 version of the MT

Methods 23 Table 3. Summary of 21 core SEI bioassessment metrics and rubrics for their interpretation. As discussed in text, both current and historic models are used. Metrics Range and Interpretation Macroinvertebrates Low Valley Multimetric Index 0 - 100; low values suggest lower stream integrity RIVPACS (O:E) 0 - 1+; low values suggest lower stream integrity RIVPACS (Bray Curtis dissimilarity) 1- 0; high values suggest lower stream integrity Valley/Foothill Prairies Multimetric Index 0 - 100; low values suggest lower stream integrity Karr Benthic Index of Biotic Integrity 0 - 100; low values suggest lower stream integrity Hilsenhoff Biotic Index 0 - ~10; low values suggest higher nutrient conc. Fine Sediment Biotic Index 0 - ~10; low values suggest more fine sediment Temperature Index 0 - ~20; low values suggest colder stream temp. Metal Tolerance Index 0 - ~10; low values suggest higher metal concentrations Component metrics in current MMI Percent EPT* 0 - 100; low values suggest lower % Percent Chironomidae 0 - 100; low values suggest lower % Percent Crustacea and Mollusca 0 - 100; low values suggest lower % Shredder Taxa Richness 0 - ~15; low values suggest fewer shredder taxa Percent Predator 0 - 100; low values suggest lower % Diatoms Sediment Increasers, Middle Rockies 0 - 100; low values suggest a lower probability of sediment problems Nutrient Increasers, Mountains/Foothills 0 - 100; low values suggest a lower probability of nutrient problems Metal Increasers, Mountains/Foothills 0 - 100; low values suggest a lower probability of metal problems Shannon Diversity 0 - ~5; low values suggest lower diatom diversity Siltation Index 0 - 100; low values suggest less sediment Pollution Index 0 - ~15; low values suggest lower nutrient concentrations Disturbance Index 0 - 100; low values suggest less disturbance Montana Diatom Multimetric Index 1 - 4; low values suggest lower stream integrity

*EPT = Ephemeroptera, Plecoptera, and Trichopetera (families of macroinvertebrates)

Water Physiochemistry also use daily mean discharge to adjust for We use the Estimate Trend (ESTREND; flow-related variability if the concentration- Schertz et al. 1991, Slack and Lorenz discharge relationship is significant (at the 2003) model to estimate long-term trend 0.10 probability level and if less than five in select water physiochemistry at the percent of the data were censored). A model Deer Lodge USGS gauge. ESTREND is a selection procedure was used to select the regression model that computes a trend specific form of the flow adjustment model (slope), which represents the median rate used. Incorporating flow in ESTREND not of change in concentration or discharge for only improves the power of the statistical the selected period of record. Following test, but decreases the possibility that an USGS convention, trends were considered observed trend is an artifact of discharge statistically significant at the 0.1 probability (Hirsch et al. 1982, Schertz et al. 1991). level. The model uses nonparametric The statistical methods used in ESTREND seasonal Kendall tests or a parametric Tobit overcome non-normal and seasonally model to account for variation across time varying data with missing, censored values, due to season. The specific number of and outliers, all of which adversely affect “seasons” for a parameter is determined by the performance of conventional statistical pattern in and the density of data (typically techniques. We include simple scatterplots there are 4 to 6 seasons in most water of select responses over the period of record chemistry parameters). The models may at each gauge to help visualize any pattern.

24 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Assessment: Reference Condition, et al. (2010) as these methods have become Assessment Points, and well established in other federal monitoring Interpretation programs; are used, in part, by the State A key step in the analysis and reporting of Montana (Suplee et al. 2005); and are of SEI data is interpreting the meaning beginning to coalesce within the NPS. behind an estimated result or set of results. For example: Is a result suggestive of high- Reference Condition quality condition and why? Does a trend Ecological assessments, directly or indirectly, in a response suggest that condition is compare measures of indicators of the status moving toward a non-reference state, and of condition or trend in condition to some why? Clearly, this is a critical aspect of a comparative condition state or rate of change mature long-term monitoring program in this state. Usually this is a condition in and something the ROMN stresses. We do the absence of human disturbance, or the not want to merely collect data and report “natural state” (Steedman 1994, Hughes numbers; rather, we seek to, in concert 1995, Jackson and Davis 1995, Davies with ROMN park staff and other partners, and Jackson 2006) and is described as the include management-relevant interpretation reference condition (Karr and Chu 1999). in the reporting of our monitoring results. Importantly, the reference condition is a gradient or distribution of values as a broad The ultimate goal of any ecological spectrum of natural ecological drivers monitoring effort is to understand the such as climate, geography, or successional condition or quality of an ecosystem or dynamics that have led to natural variation in some resource of interest at a defined spatial condition. (e.g., a single site or an entire park) and temporal (e.g., once or over time) scale. Reference condition has been used in Ecological condition is influenced by both ecological assessment in at least four anthropogenic and natural drivers, and different ways (Stoddard et al. 2006): thus occurs as a distribution or gradient of 1. the condition of ecosystems at some states. For most resources, the distribution point in the past; of condition levels ranges from some dysfunctional, aberrant or “non-natural” 2. the condition that might be achieved state to a functional, intact or “natural” if resource management was more state. If assessed over a constrained time effective; period (e.g., once), condition can be expressed as the status of the ecological 3. the best existing condition; or resource of interest. If assessed over time, it 4. the condition of systems in the absence may be expressed as a trend (which might of significant human disturbance. have a value of zero, indicating no trend). Indicators or vital signs are components The current NPS I&M treatment (S. Fancy, or elements of an ecological resource of pers. comm. November, 2013) defines interest that are particularly information a reference condition as a quantifiable rich and that represent or indicate or otherwise objective value or range of condition. Measures are specific, often values for an indicator of condition that is field-based, measurements used to quantify intended to provide context for comparison or qualitatively evaluate the condition with current condition. Furthermore, of an indicator at a particular place and NPS considers the reference condition to time. Measures are never perfect and the represent an acceptable resource condition inability to measure or quantify condition supported by appropriate information exactly adds additional variance to the and scientific or scholarly consensus. This characterization of condition. is a general definition and allows any of the four forms defined by Stoddard et al. We generally follow approaches and use (2006) listed above. However, a specific terminology as presented in Stoddard et al. and operational definition of reference (2006), Bennetts et al. (2007), and Hawkins

Methods 25 condition is important for application in estimate of ecological integrity. However, ROMN SEI monitoring at GRKO. For reference sites truly or completely unaffected example, a reference condition estimated as by the global influence of human activity do the condition of ecosystems at some point not exist. Therefore, minimal disturbances in the past is very different from a reference are allowed (e.g., atmospheric contaminants condition estimated as the condition that at levels below known effects). Once today’s sites might achieve if they were established, the distribution created by a better managed. Moreover, these two types group of reference sites in MDC will vary of reference conditions (historic and future little over time. Long-term climatic, geologic, or desired with ideal management) may be and ecological fluctuations will inevitably difficult to numerically estimate. change the characteristics of individual sites within this distribution, but the range The ROMN has elected to use two forms of MDC should be nearly invariant, and its or definitions of reference condition, distribution can serve as an anchor by which depending on which ROMN park is being to judge current condition. evaluated. First, for our wilderness parks (Glacier, Rocky and Great Sand Dunes An LDC describes the best available National Parks), we use a minimally condition within a landscape. It is based disturbed condition (MDC) or the best on a set of explicit rules that allow a degree existing condition. Second, for LIBI, nested of human activity (Bailey et al. 2004, within a more anthropogenic landscape, Hughes et al. 1986, Hughes 1995). These we use a least disturbed condition (LDC) rules vary from region to region, given the or a condition defined by the absence of characteristics of the landscape and human significant human disturbance. We define use of the landscape and are developed these further below. iteratively with the goal of establishing the least amount of ambient human disturbance A MDC or LDC as a distribution of numeric in the region under study. Because the estimates of condition can be empirically degree of human disturbance changes over derived from reference sites or stream sites time (i.e., as either degradation or restoration independently determined to be in a MDC occurs), the LDC may also vary with time— or LDC state. The process for evaluating this must be carefully considered in assessing candidate sites to determine their condition monitoring data. state can be complex (Stoddard et al. 2006). In summary, samples of candidate Assessment Points reference sites are generated from large, While a reference condition is important, unbiased surveys of a parks streams or or even required for some approaches to from representative samples of streams in assessment (see below), most interpretations the ecoregion(s) in which a park is located. about status or trend in an ecological Reference sites are derived from these response are made relative to single samples using a series of filters or models parameters (or a discrete range of values) that cull sites that are above relevant criteria that define boundaries between condition or other assessment points (see below). states. In the simplest case, these boundary The assessment points used for this filtering values are compared to monitoring data are non-biological and are usually water and an assessment is made as to whether chemistry criteria, the density or simple the monitoring data is in reference, presence of anthropogenic disturbances intermediate, or non-reference (more (i.e., a coal mine) in a streams watershed or complex methods may use the complete general large scale land use patterns (i.e., reference condition distribution for the proportion of a sites watershed that is comparison and/or statistical methods to logged). attach a probability to these statements). The MDC describes a state in absence The type, form, and estimation methods of significant human disturbance, or a behind these boundaries have been variously condition that is the best approximation or presented in the literature. Bennetts et al. (2007) provide a useful summary relevant to

26 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site NPS I&M monitoring and proposes that the Many monitoring programs derive term assessment point be used as a catch-all assessment points from distributions of an term and as a way to more explicitly connect indicator created from a sample of LDC monitoring, assessment, and park resource or MDC reference sites (Stoddard et al. management. Assessment points represent 2005, Suplee et al. 2005, CO DPHE 2010). points along a continuum of condition We use the term ecoregion assessment states where scientists and park managers points for these boundaries, as in many have agreed that assessment of the status or cases the reference distributions used are trend of a resource relative to program goals, from a sample of reference sites from an natural variation, or potential concerns ecoregion (although in ROMO or GLAC is warranted. These points may take the we might also or instead use data from our following form: extensive unbiased surveys of the parks streams). Most current programs simply Assessment points may be true ecological choose a percentile from the MDC or LDC thresholds where abrupt change occurs in reference distribution as the assessment ecosystem condition (Groffman et al. 2006). boundary (Stoddard et al. 2005). In most Ecological thresholds can be particularly cases, these percentiles are less conservative important as there are sometimes in LDC, allowing a more disturbed irrecoverable consequences to crossing ecoregion assessment point that is in a them (Groffman et al. 2006). Examples in sense more realistic in these landscapes. stream ecosystems relevant to ROMN parks While these percentiles are arbitrary, might include heavy metal concentrations because they occur within a well-defined driven by mine drainage that can often reference condition distribution and are dramatically and potentially permanently explicit, they have value, especially when alter communities of macroinvertebrates and used over time. Other methods for locating diatoms (Clements et al. 2000, Kashian et al. these boundaries may also be employed 2007). Many examples exist in the terrestrial that model the relationships between ecosystems, such as fire-type systems being response and disturbance at reference sites invaded by annual grasses (Bestelmeyer (Schoolmaster et al. 2012, 2013a,b), and thus 2006, Briske et al. 2006, 2008). However, might estimate more meaningful ecological despite widespread agreement among patterns, perhaps even a true ecological scientists that ecological thresholds are real threshold. Ecoregion assessment points and can be extremely important, they are figure prominently in ROMN interpretation often difficult to estimate as they can be of monitoring data at GRKO. influenced by multiple complicating factors (Lindenmayer and Luck 2005, Groffman et A similar type of assessment point can al. 2006). Ecological thresholds often require be generated from time-series data and complex modeling and rich data sets to function as baseline assessment points estimate and describe causal relationships or more traditionally just baselines. These among ecological response and system may be simple summary statistics derived drivers. Assessment points set at an actual from a sufficient record of data or more ecological threshold may run the risk of not complex trend model parameters. The type allowing sufficient time for management of data used in estimating the baseline value actions to avoid a park resource becoming determines much about the quality and form degraded beyond repair. However, if the of the assessment point. If the source data system dynamics around a threshold are are from a well-designed survey of an entire understood, assessment points at ecological resource sampled over time, they likely have thresholds may afford management clear more value than if from a single site that direction in where or how to manage a may not be representative of the resource system because of the possible strong of interest. Similarly, if the source data span behavior when systems trend near these a long time period that captures relevant proverbial cliffs. temporal dynamics they are likely more representative than if the data came from a single sample event. Baselines may be true

Methods 27 ecological thresholds if they are generated but this, of course, adds to the complexity from models that include sufficient and may be part of the reason management covariates to estimate causal patterns in the assessment points are somewhat rare in response of interest. Alternatively, they may practice (Bennetts et al. 2007). be somewhat arbitrary (but if derived from a sufficient period of record, likely represent a Several other forms or definitions of meaningful estimate of the state or trend in a assessment points exist in the literature response). and in practice within monitoring and assessment programs (e.g., critical loads, Criteria (also called standards) are desired future conditions, threshold of assessment points defined for decision that potential concern; Bennetts et al. 2007). have an explicit connection to a regulatory However, as of this documents publication, policy. These are usually based on human we focus on ecological thresholds (where health or environmental effects and possible), ecoregion assessment points, generally represent the lower limits of the baselines, and criteria. Management acceptable range in a condition gradient assessment points are also important, (or the boundary of non-reference) and but these should be set by park resource are used by regulatory agencies to ensure managers (ideally in consultation with that a resource does not become impaired. the ROMN and other collaborators) and Water quality criteria are well-known this was under development at the time of examples. Criteria are often estimated writing. using experimental methods (e.g., finding the lethal dose of some chemical), but can ROMN Assessment Process at GRKO come from modeling as well. The NPS In broad stroke, we compare our monitoring defines and uses criteria internally as an results to assessment points with the approach to facilitating decisions regarding following priority: established numeric water the management of public use of parks quality criteria (as defined by the States of (NPS 2005). The ROMN’s use of criteria as Montana or Colorado or the EPA), relevant assessment points is informal (i.e., as stated published assessment points (that may not several times in this document, not legally be used by a State) including ecological binding), but in some cases, these values assessment points (published or derived connect our assessment to a rich history of by the ROMN), or to baselines from SEI monitoring by regulatory agencies such as or other data. We use numeric (versus MT DEQ or the EPA. narrative) assessment points to remove as much subjectivity from the process as Finally, management assessment points possible. We interpret these comparisons represent a point or zone that triggers using ecological theory, NPS I&M Division management action within a given context. guidance, NPS resource management Management assessment points are often policies, and collaborative work with set to facilitate a priori consideration ROMN partners, especially park staff and of undesirable ecosystem changes (e.g., management. As noted above, we use the ecological thresholds) and enable more terms reference, intermediate and non- proactive management responses. This is reference to label condition states or how a complex as it requires knowledge of the SEI response relates to an assessment point. rate of change in a system (so the tipping point at which a system first begins to In GRKO, given a non-pseudoreplicated change is knowable), the response time or temporally auto-correlated sample size required for management to change system of one and various statistical issues with behavior, and monitoring keyed into these assessing single sites (Hawkins et al. 2010), temporal dynamics to inform management we do not perform statistical tests that of success (or not). Ideally, management compare SEI data to assessment points—all assessment points include all the factors that comparisons are qualitative and are simply management decisions must be based on, whether a SEI value during a defined time such as social, economic, and political values, period, is above or below a assessment point,

28 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site with no estimate of the statistical probability we do emphasize biological and habitat of this occurrence. As we collect additional responses as these are often more synthetic SEI data at GRKO we will explore more or integrative vital signs. Disagreement in rigorous methods. In GLAC and ROMO, two or more SEI responses may reflect error with well-designed surveys and large sample in the methods or models used to create the sizes, a much broader array of statistical tests metrics or collect the data. Alternatively, are available. The appropriate hypothesis it may reflect the true nature of complex when statistically assessing individual sites ecological systems; it is not unusual for one or a population of streams is whether the aspect of an ecological system to be intact condition observed is outside the range of and one to be in a non-reference state. values expected among an appropriate set of Note that the “Summary Condition Table” reference sites (Smith et al. 2005, Bowman presented in the Executive Summary of and Somers 2006, Hawkins et al. 2010). this report represents a weight of evidence approach to derive a general statement of Note that in most cases MT DEQ, when condition, trend, and confidence in these exercising their regulatory authority, does characterizations within major classes not simply designate a stream as impaired if of response as used to characterize the a parameter is above a criterion once (with condition of a ROMN park stream resource. an exception for human health criteria). For example, in nutrient assessment a number of We do not restrict the indicators or metrics exceedances are allowed over a well-defined we use (and any assessment points created time period and spatial extent before the for these) to those currently employed state would designate a site as impaired. by MT DEQ or the US EPA, nor do we Suplee and Suplee (2011) present details on purposefully restrict the number we nutrient assessment in Montana, including use. Rather, we use a broad spectrum statistical tests used to establish an actual of models, indices and metrics. We feel impairment. this is appropriate as it enables a more comprehensive perspective and connects SEI We occasionally apply multiple assessment data to a larger array of historic and ongoing points to a single response indicator or its stream monitoring by ROMN partners. That measure. This is appropriate when alternate said, we do emphasize approaches that states assessment points are derived from data and other partners have demonstrated are at different scales or when one describes more precise and predictive. We will also an ecological response and one a strictly simplify the set of models used as we learn regulatory perspective. When results using what works best over time. different assessment points disagree or when general conclusions differ across indicators Criteria (e.g., water chemistry suggests a site is in a We harvest criteria from the most current reference state, but its biology does not), we sources available from the State of Montana use a weight of evidence approach to help (MT DEQ 2012e) or the US EPA (EPA resolve overall stream condition; however, 2009a.). Montana numeric water quality

Importantly, the NPS does not have regulatory authority over water quality in GRKO under the Clean Water Act including the ability to evaluate beneficial designated uses. As noted above, the Organic Act and various NPS Management Policies do require park management to maintain, rehabilitate, and perpetuate the inherent integrity of aquatic resources and processes and NPS will work with appropriate State and other Federal partners to do so. SEI assessment at GRKO is conducted outside of any Clean Water Act motivated regulatory context. Any comparisons we make to state or federal criteria do not include any official statement as to whether a beneficial designated use is attained. NPS can participate in collecting data used in the protection of water bodies in parks under state jurisdiction through the Clean Water Act.

Methods 29 criteria are for potential effects of chronic methodological differences in the protocols exposure over an extended period of time employed by the various programs that (months) or for acute exposure over a short generate reference site data that would period of time (hours or days). Permitted increase measurement error. We feel that levels are lower for chronic aquatic-life this is not a prohibitive assumption for our exposure than for acute exposure. Standards purposes as in most cases the methods for chronic exposure may not be directly used are derivatives of the US EPA EMAP comparable to SEI results obtained from protocol and are quite similar. Jessup (2006) episodic or one-time grab samples and shows that stream monitoring data collected from a compliance standpoint, acute by several different programs in Montana, instantaneous criteria afford the only typically give comparable results, especially direct comparison for such data. However, when using more derived models/metrics from a resource-conservation standpoint, from a MMI or RIVPACS model. instantaneous grab-sample data, when compared against more stringent aquatic- Second, we assume ecoregion reference life chronic criteria, can provide a means sites used to derive the MDC are similar of early warning and an indication of a enough to the Clark Fork at GRKO such that problem that may require more attention. they can be used as essentially replicates of For these reasons, we use aquatic-life fixed controls (Hawkins et al. 2010). The chronic standards with the goal of providing spatial variation among reference sites in a advance warning of an impending problem ecoregion (an outcome of multidimensional before it becomes severe. This mirrors USGS environmental heterogeneity that exists protocols also applied to data from ROMN naturally across any landscape and its parks (Mast 2007). waterways) represents natural variation associated with water body types that Ecoregion Assessment Points occurs within the ecoregion–this is the For GRKO, following park management primary reason why the suite of reference guidance, we develop a MDC from reference sites used are not restricted to identical or stream sites identified by MT DEQ (Suplee et near identical stream types as the site(s) al. 2005), US EPA (Stoddard et al. 2005), and being tested. The variation expressed by other partners within the Middle Rockies these distributions is constrained to a ecoregion and estimate novel ecoregion degree by the shared landscape context assessment point values within these and geophysical drivers of stream sites distributions following methods in Hughes within a common ecoregion (Hynes 1975). et al. (1986) and Stoddard et al. (2005). We Hughes et al. (1986) suggests that the accept the reference sites designation and proximate environmental features common methods used to reach these conclusions to ecoregions should strongly influence the by Suplee et al. (2005) and Stoddard et al. biotic character of streams. Thus, ecoregions (2005). For responses that increase with might partition (control for) the collective decreasing condition we use the >95th or effects of the most important natural <75th percentile values for non-reference factors that influence the distribution and and reference (respectively). For responses abundance of aquatic biota. that increase with increasing condition we The assumption that ecoregion reference th th use the <5 or >25 percentile as assessment sites create a valid and useful context for points for non-reference and reference comparison and interpretation of SEI (respectively). monitoring data is fairly restrictive. However, Assumptions the approach has been used for at least two The estimation of ecoregion assessment decades by multiple agencies, has enabled points from ecoregion reference site data effective ecological monitoring and resource requires some important assumptions management, and we feel it is useful for (Hughes et al. 1986, Hawkins et al. 2010). understanding SEI data. In a sense to cover First, it assumes that there are no meaningful all the bases, we also use models (i.e.,

30 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site RIVPACS) that take an alternate approach Summary and do not a priori assume reference sites In summary, the key steps of our assessment are grouped regionally. Rather, given the methods at GRKO are as follows: well documented (Ostermiller and Hawkins 2004, Hawkins 2006, Ode et al. 2008) 1. Compare SEI data to existing association of reference-site biota (and the published State of Montana criteria indices derived from them) with natural and interpret these comparisons. gradients that transcend ecoregions or This is done for most water chemistry vary markedly within them RIVPACS uses constituents and several MT DEQ biota–environment relationships derived biological metrics. If regulatory criteria from reference sites to predict the most exist (from MT DEQ or EPA) we apply likely ecological reference condition at any and interpret these values before other individual assessed site (Hawkins et al. 2010). assessment points as required by the Of note, in larger ROMN parks, with larger Clean Water Act. samples of independent sites we have been For water physiochemistry, we use able to generate proprietary and novel MMI aquatic-life chronic criteria with the models that account for these gradients goal of providing advance warning of an (Schoolmaster et al. 2012, 2013a,b), perhaps impending problem before it becomes blending the best of these two approaches. severe.

Baselines 2. Compare SEI data to existing and For some indicators we statistically relevant ecological thresholds. In some summarize data collected at the beginning cases the true nature of the threshold of SEI monitoring or use partner data from may not be fully understood and the USGS stream gauge in Deer Lodge to therefore our interpretations of these estimate a baseline condition against which boundaries are cautious. We apply this future or current status and trend might be to a few key water or sediment chemistry compared. SEI derived baseline assessment constituents, and several biological and points will require several monitoring habitat metrics. cycles to have any confidence that they are estimates of a meaningful condition state. In 3. Compare SEI data to novel ecoregion most cases, determining whether a baseline assessment point values derived from assessment point occurs in the reference distributions of reference site data range of the condition gradient will require from the surrounding Middle Rockies modeling. Until this is known, interpretation ecoregion. We apply this to a few key of comparisons of current condition to water chemistry constituents, and baselines or of any trend developed with several biological and habitat metrics. the baseline as a starting point may only be For setting assessment points in relative in nature. distributions from ecoregion reference sites we follow EPA guidance for the percentile values used to establish non- reference/reference assessment points.

4. Compare SEI data to baselines derived from USGS gauge data or other auxiliary data. Once we have accrued sufficient SEI data we may use our own data to generate baseline assessment points as well.

Methods 31

Results and Discussion Sample Sites and Events few meaningfully elevated concentrations, From 2008 to 2010 a total of eight sample although a few select parameters did exceed events were conducted at the GRKO State of MT criteria. Table 5 presents a SEI sentinel site (Table 4). These include simple summary of these results with more “sentinel” samples where only water and details below. sediment were collected and “full” or Of particular note, several metals of concern “base” sample events with a complete related to the mine tailings released into implementation of the SEI protocol. the valley in the late 1800s and early 1900s occasionally exceeded chronic aquatic life Water and Sediment criteria. Similarly, sediment criteria specific Physiochemistry to the Clark Fork were exceeded by arsenic, Summary copper, and lead at least once during 2008- In general, water and sediment 2010. Total mercury in sediment was present physiochemistry was in an intermediate in low concentrations in all sediment to reference state from 2008-2010 with samples but was not above its criteria.

Table 4. SEI sample events at GRKO, 2008-2010. Event Date Event Description Hydrograph Limb Notes Non-standard SEI site layout for Targeted collection of benthos and 10/06/2008 Base biology; no sediment or habitat periphyton, water physiochemistry measures

05/18/2009 Water and sediment physiochemistry Rising/Peak Sentinel sample event at or near bridge

Water and sediment physiochemis- Full sample event across complete 08/13/2009 try, benthos, periphyton, all habitat Base sample reach responses, instantaneous discharge

11/16/2009 Water and sediment physiochemistry Base/Winter Sentinel sample event at or near bridge

5/7/2010 Water and sediment physiochemistry Rising Sentinel sample event at or near bridge

6/11/2010 Water and sediment physiochemistry Peak/Falling Sentinel sample event at or near bridge

8/20/2010 Water and sediment physiochemistry Base Sentinel sample event at or near bridge

10/29/2010 Water and sediment physiochemistry Base/Winter Sentinel sample event at or near bridge

Table 5. Summary condition table excerpt for select water and sediment measures at GRKO from 2008 to 2010. We include example vital signs and indicators, a brief description of results and patterns, and symbolize the status, trend, and our confidence in those summaries. See the Executive Summary for the complete Summary Condition Table. Vital Sign Summary Symbol (Example Indicators)

Most major ion concentrations were in an acceptable range yet chloride and sulfate concentrations were often higher than in ecoregion reference sites. Arsenic, cadmium, copper and lead exceeded State of Montana water quality Water physiochemistry criteria. Nutrient concentrations were occasionally high but we currently lack (major ions, nutrients, metals) sufficient data to apply the complete MT DEQ nutrient assessment protocol. There is evidence that the long-term trend in several parameters is improving. We have medium confidence in our assessment of major ions, nutrients, metals at GRKO.

All of the core NPS parameters were in an acceptable range. The long-term trend Water in-situ chemistry in stream temperature suggests rising temperatures—or a deteriorating condition (pH, conductivity, DO, (but the period of record is short at only around nine years). We have medium temperature) confidence in our assessment of in situ parameters at GRKO.

Arsenic , copper, and zinc in sediment concentrations were often above their non-reference assessment points. Total mercury concentrations were intermediate Sediment chemistry but because the primary toxic form of mercury is methylmercury, these data are

(metals) somewhat inconclusive. We currently lack data to assess trends, but work by partners suggests there is an improving trend in many metals found in sediments. We have medium confidence in our of sediment chemistry at GRKO.

Results and Discussion 33 Nutrient concentrations did occasionally Many constituents have formal State of and marginally exceed Clark Fork River Montana water quality criteria (as well as specific criteria but we lack data to full assessment points from other sources used assess nutrients. Chloride and sulfate (often by the State). As noted above, when using used as general indicators of anthropogenic MT DEQ criteria NPS cannot formally disturbance), were higher at GRKO than interpret data in the context of meeting reference assessment points from Middle designated uses. Given the large number of Rockies ecoregion reference sites (Stoddard existing MT state water quality criteria we et al. 2005). However, using USGS data from only compare two parameters, sulfate and the nearby gauge, significant decreasing chloride, to ecoregion assessment points trends were seen in conductivity, total we derive. Both of these constituents are suspended sediment, calcium, cadmium, known to be useful indicators of general copper, lead, and zinc. This suggests anthropogenic disturbance (Stoddard improving water quality over the last decade et al. 2005), and we feel that qualitative and a half or so. comparisons to ecoregional assessment points provide useful context for Measures, Criteria, and Assessment interpretation at GRKO. Points Table 6 presents a summary of the status All imputation models to deal with non- of SEI water physiochemistry results for detects in SEI physiochemistry data were 2008-2010. We interpret results following significant and statistics in the table reflect the general ROMN assessment approach these adjustments. Results in Table 6 are as outlined above. All interpretations are restricted to the SEI sentinel site in GRKO done in the context of the ecology of the sampled from 2008 to 2010. For a similar Clark Fork River and management needs for analysis of a larger number of sites over a GRKO (i.e., maintaining the historic context longer period of time see Mast (2007). of the Grant Kohrs ranch). Table 6. Summary of water and sediment physiochemistry for the Clark Fork River in Grant-Kohrs Ranch National Historic Site, 2008-2010. Except as indicated, assessment points are state of MT chronic aquatic-life criteria with human-health values in parentheses. All assessment points define non-reference except those in green4. The non-reference state is above all assessment points except for those with an inequality indicating that it is below the value. Concentrations in bold red indicate results in non-reference at least once during 2008-2010. Values in bold are either intermediate or non-reference (depending on the existence of a non-reference assessment point). For additional clarifications on Table content see Notes below. Criteria or No. Minimum Median Maximum Constituent or property Assessment analyses0 value value value Point Field properties Temperature, water (°C) 7x 4.8 5.5 19.1 (see below)1 Oxygen, dissolved (mg/L) 7x 8.2 9.3 12.2 <8, <42 pH (standard units) 7x 8.1 8.5 8.6 (see below)3 Specific conductance (μS/cm) 7x 266 428.3 480.9 500,5004 Major constituents, dissolved

Alkalinity, dissolved (mg/L as CaCO3) 7 88.1 148.9 235.5 -- Calcium, dissolved (mg/L) 7 36.8 61.3 67.9 -- Chloride, dissolved (mg/L) 7 6 7.9 23.2 0.4, 3.85 Fluoride, dissolved (mg/L) 3 0.4 0.5 0.5 (4)

Hardness, dissolved (mg/L as CaCO3) 7 77.8 186 216 -- Magnesium, dissolved (mg/L) 8 6.5 13.1 13.4 -- Potassium, dissolved (mg/L) 7 1.9 2.2 2.6 -- Silica, dissolved (mg/L as Si) 8 16.9 18.6 22.1 -- Sodium, dissolved (mg/L) 7 9.6 13.4 15.5 --

34 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Table 6. Summary of water and sediment physiochemistry for the Clark Fork River in Grant-Kohrs Ranch National Historic Site, 2008-2010 (continued). Criteria or No. Minimum Median Maximum Constituent or property Assessment analyses0 value value value Point

5 Sulfate, dissolved (mg/L as S04) 8 41.7 72.9 86.7 4.8, 77.4 Total Suspended Solids (mg/l) 4 6.4 13.1 42.4 -- Nutrients in water or periphyton, dissolved and/or total recoverable Ammonia, dissolved (mg/L as N) 8(6) 0.002* 0.003* 0.01* 3.156 Nitrite + Nitrate, dissolved (mg/L as N) 7(3) 0.004** 0.02** 0.11 0.1(10)8 Nitrite, dissolved (mg/L as N) 7(3) 0.001** 0.002** 0.005 (1) Nitrate, dissolved (mg/L as N) 7 0.002 0.07 0.19 (10) Nitrogen, total (mg/L as N) 8 0.16 0.26 0.31 0.39 Orthophosphate, dissolved (mg/L as P) 8(1) 0.0008** 0.004** 0.02 -- Phosphorous, total (mg/L as P) 8 0.007 0.02 0.03 0.029 Carbon, organic, dissolved (mg/L as C) 8 2.1 2.7 9.8 -- Carbon, organic, total (mg/L as C) 4 1.9 2.6 4.2 -- Chlorophyll-a in periphyton (mg/m2) 2 4.0 7.0 10.1 10010, 15010 Ash Free Dry Mass in periphyton (g/m2) 2 8.9 9.2 9.5 3510 Metals in water, total recoverable Aluminum, total (μg/L) 7(2) 52.8** 210** 865 -- Arsenic, total (μg/L) 8(2) 9.3 14.5** 22 150(18)12 Barium, total (μg/L) 4 43 49 52.7 (1000) Beryllium, total (ug/L)^ 4(4) ------(4) Cadmium, total (μg/L) 3(2) 0.3** 0.3** 0.3** 0.52(5)12 Chromium, total (μg/L)^ 4(4) ------(100) Copper, total (μg/L) 8 6 21.8 74 3.8(1300)12 Iron, total (μg/L) 8(1) 50.9 283.5** 1190 1000(300)11 Lead, total (μg/L) 7(4) 0.7** 2.6** 13 0.55(15)12 Manganese, total (μg/L) 8(1) 35** 75.4** 120 5011 Selenium, total (μg/L)^ 8(8) ------5(50) Zinc, total (μg/L)^ 3(3) ------37(2000)12 Metals in water, dissolved Aluminum, dissolved (μg/L)^ 8(8) ------87 Arsenic, dissolved (μg/L) 7(3) 5.6* 10.3* 19.9 150(10)7 Barium, dissolved (μg/L) 3 30.9 46.1 46.3 -- Beryllium, dissolved ug/L)^ 4(4) ------Cadmium, dissolved (μg/L)^ 3(3) ------0.46(4.41)7 Chromium, dissolved (μg/L)^ 3(3) ------(86)7 Copper, dissolved (μg/L) 7(1) 5 9* 12 3.6(1248)7 Iron, dissolved (μg/L)^ 7(7) ------Lead, dissolved (μg/L) 7 77.8 186 216 0.9(10.5)7 Manganese, dissolved (μg/L) 7(4) 0.008* 0.30* 25.2 -- Selenium, dissolved (μg/L)^ 7(7) ------5(50)7 Zinc, dissolved (μg/L)^ 4(4) ------36.5(1972)7 Metals in sediment Aluminum, total (mg/kg) 7 3590 6180 10500 -- Arsenic, total (mg/kg) 7 17.4 50.9 118 9.79,3313 Barium, total (mg/kg) 7 80.3 123 206 -- Beryllium, total (mg/kg) 7 0.3 0.5 0.9 --

Results and Discussion 35 Table 6. Summary of water and sediment physiochemistry for the Clark Fork River in Grant-Kohrs Ranch National Historic Site, 2008-2010 (continued). Criteria or No. Minimum Median Maximum Constituent or property Assessment analyses0 value value value Point Metals in sediment (continued) Cadmium, total (mg/kg) 4 0.3 0.5 0.9 0.99,4.9813 Chromium, total (mg/kg) 7 6.2 10.6 16.4 -- Copper, total (mg/kg) 7 212 371 780 31.6,14913 Lead, total (mg/kg) 7 38.4 61.6 108 35.8,12813 Iron, total (mg/kg) 7 8390 13200 19500 -- Mercury, total (mg/kg) 7 0.146 0.337 0.594 0.180,1.0613 Selenium, total (mg/kg)^ 7(7) ------Silver, total (mg/kg) 7(1) 0.61* 1.4* 3.2 -- Zinc, total (μg/L) 7 260 403 731 121,45913

Criteria/Assessment Point Notes All water quality assessment points are State of Montana chronic aquatic life standards (MT DEQ 2012e) with human health standards in parentheses except as follows: 1 There are no relevant State of Montana temperature criteria for SEI instantaneous in situ temperature data. 2 Freshwater Aquatic Life Standards for DO are cold water 1 day minima for (1) early life stages followed and (2) other life stages. They are instantaneous water column concentrations to be achieved at all times (MT DEQ 2002) 3 For C-1 waters, induced variation of hydrogen ion concentration (pH) within the range of 6.5 to 8.5 must be less than 0.5 pH unit. Natural pH outside this range must be maintained without change. Natural pH above 7.0 must be maintained above 7.0 (MT DEQ 2002) 4 Conductivity assessment point (EPA 2007a) is a growing season instantaneous maximum followed by a monthly average as developed for Tongue River tributaries in eastern Montana and is applied informally and with caution to GRKO 5 Assessment points for sulfate and chloride are generated from reference sites in the Middle Rockies ecoregion at the <75th or >95th percentile values for reference/non-reference, respectively (Stoddard et al. 2005, 2006) 6 MT DEQ (2010a) ammonia criteria is a table value lookup for an acute value for total recoverable ammonia nitrogen at a pH of 8.3 assuming early life stages of salmonids (i.e., bull trout) are present or are a resource objective 7 Values for dissolved trace elements are derived from the State of Montana water quality numeric criteria using formulas

from U.S. EPA (2009) and a median hardness of 186 mg/L as CaCO3 8 Nitrite + Nitrate is informal and is for a base flow in the Middle Rockies Ecoregion (Suplee et al. 2008), the human health standard is from MT DEQ (2012e) 9 Total nutrient criteria are proposed for base flow in the Clark Fork between Warm Springs Creek and Bitterroot River (Su- plee and Sada de Suplee 2011, Suplee et al. 2008) 10 Chlorophyll-a and AFDM are “nutrient assessment support criteria” for base flow in the Clark Fork between Warm Springs Creek and the Bitterroot River (Suplee and Sada de Suplee 2011, MT DEQ 2011a): 100 mg Chl-a /m2 is a summer mean; 150 mg Chl-a /m2 is a summer maximum); Note the different units (mg/m2 vs. g/m2) 11 Human health value is a secondary standard based on aesthetic properties such as taste, odor, and staining and is more conservative than chronic standards. MT DEQ does not assess iron secondary standards that apply to taste and odor to water-dwelling organisms are expected, whereas PECs are the concentrations at which negative effects on sediment dwelling organisms are judged more likely to occur than not 12 Derived from criteria in Record of Decision (EPA 2004) using formulas from U.S. EPA (2009) and a median hardness of 186

mg/L as CaCO3; may differ from statewide standards 13 Sediment criteria are Threshold of Effect Concentrations (TECs) followed by Probable Effect Concentrations (PEC) values from MacDonald et al. (2000). TECs predict the absence of sediment toxicity while PECs predict its presence. Non-detect Notes 0 Values for sample size are total N with the number used in ROS models for non-detects in parentheses * Value or median contains predicted results from ROS model ** >80% of results at Detection Limit(s), results tenuous ^ All results at Detection Limit(s), no model possible Other Notes: x N for field properties is the number of events (the number of discrete data values within each event ranges from 1 to ~100) --, not detected; No., number; mg/L, milligrams per liter; μg/L, micrograms per liter; °C, degrees Celsius; μS/cm, microsiemens

per centimeter at 25°C; CaCO3, calcium carbonate; N, nitrogen; P, phosphorous.

36 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Quality Assurance and Quality Control surface water due to longer rock-water Signal to noise values are not readily interaction time. Minimum values of specific available from EMAP for water chemistry conductance occurred during higher-flow data; therefore, QAQC for SEI water conditions probably due to the influx of physiochemistry is currently restricted dilute snowmelt water. to field and lab procedures as discussed in Appendix E. In summary, data are not The State of Montana has established a included in the results presented here minimum instantaneous dissolved-oxygen unless they passed all field and lab QAQC concentration of 8 mg/L for early life stages procedures. When we have enough of fish and 4 mg/L for other life stages in replicates, additional QAQC will be included class C-1 Rivers like the Clark-Fork (MT in future reports. Appendix E presents DEQ 2002). Dissolved oxygen at GRKO mode details on ROMN QAQC procedures. was above the criterion (or in a reference Appendix G presents laboratory detection condition) during all SEI sample events. limits for water chemistry parameters. Dissolved-oxygen concentrations were greater during winter and spring and In situ Parameters decreased in late summer when water In situ water chemistry during SEI sentinel temperature and biological respiration in the events showed a fairly high degree of river is increased (although our sample size variation. There were few exceedances of by season is very small). As our sample size the available assessment points for in situ of DO in Clark Fork increases, these data parameters and none that met all of the will be more and more useful for comparison requirements for applying State of Montana to proposed State of Montana criteria. water quality criteria. Finally, the State of Montana established Episodic water temperatures measured in classification standards for C-1 waters that 2008 to 2010 (in spring summer and into state any induced variation of pH within the the fall) ranged from around 5º to almost range of 6.5 to 8.5 must be less than 0.5 pH 20ºC. Qualitatively speaking, this range is not units, that natural pH outside this range must excessively warm or cold for the Clark Fork be maintained without change and natural in GRKO. State of Montana classification pH above 7.0 must be maintained above 7.0 standards for a class C-1 river like the Clark (MT DEQ 2002). There is no manipulation Fork are focused on point-source discharges of pH by the NPS within GRKO and while and are not easily applied to SEI data (MT there are complex interactions with other DEQ 2002). Data from SEI continuous water physiochemistry parameters that temperature loggers and the nearby USGS can influence pH, some of which do have gauge, including informal comparisons to anthropogenic sources upstream and assessment points used in bull trout habitat within GRKO, there is no evidence that pH assessment presented below are better assays within the Clark Fork in 2008-2010 was not of water temperature at GRKO. “natural.” Our episodic measurements of in situ pH in the Clark Fork were all above 7.0, Specific conductance ranged from 266 to with most somewhat basic probably due to 481 µS/cm. There are no State of Montana source material in the surficial geology of the numeric standards for specific conductance; watershed. Heavy metals tend to be more however, values at GRKO in 2008-2010 were toxic in acidic waters where they are more reasonable for a river like the Clark Fork. soluble and more bioavailable, so the more Specific conductance varied inversely with basic pH at GRKO is perhaps a good thing. discharge (r=-0.94, p<0.001), with maximum values during low-flow conditions. In general, our in situ results from GRKO Groundwater is a large component of do not suggest any problems in water streamflow during low-flow conditions and quality. However, these parameters can vary often has higher concentrations of dissolved strongly over the course of 24 hours (i.e., solids (and hence higher conductance) than by temperature or from sorption of metals

Results and Discussion 37 between water, sediments, and aquatic plants and paper or textile production—many (Nimick 2003). Diel fluctuation, beyond that of these occur in the watershed above measured over the two to four hours of most GRKO. Sulfate is also probably released continuous data collected during SEI sample from oxidation of sulfide minerals in mine events, is not included in the values in Table wastes located in waste piles and tailings 6 and so these results should be treated with ponds in the headwater valleys of the Clark caution. Fork and in flood-plain deposits along the river - there is some research that links Major Constituents the tailings and contaminated flood plain Major dissolved constituents at GRKO were sediments to ground water quality at GRKO dominated by calcium and sulfate. Alkalinity (e.g., Woesner and Johnson 2001). Sulfate ranged from 88 to 236 mg/L, indicating that is also probably released from oxidation the Clark Fork in GRKO was relatively well of sulfide minerals in mine wastes located buffered. There are few existing state or in waste piles and tailings ponds in the federal criteria for anions and cations and headwater valleys of the Clark Fork and in the one defined by the State of Montana (a flood-plain deposits along the river—there human health standard for fluoride) was not is some research that links the tailings exceeded by SEI data. and contaminated flood plain sediments to ground water quality at GRKO (e.g., As of the time of this reports publication MT Woesner and Johnson 2001). DEQ and EPA were developing guidance for a chloride criterion (Darrin Kron, Pers. However, SEI monitoring cannot Comm., 2013). When these are available conclusively link concentrations of these we will informally apply them to SEI data. two anions to any of these sources. Sulfate However, because chloride (and sulfate) can be derived from natural sources such are important indicators of possible general as weathering of carbonate rocks common anthropogenic disturbance (e.g., Biggs et al. in the Clark Fork basin and can vary greatly 2004) we developed ecoregional assessment depending upon flow. Chloride is also widely points to aid in the interpretation of the distributed in nature (it constitutes about measured concentration of these chemicals 0.05% of the earth’s outer crust), generally at GRKO. Based on surrounding ecoregion in the form of sodium and potassium salts. reference sites, a reference concentration Bright and Addison (2002) show that natural was below 0.4 and 4.8 mg/l for chloride and background concentrations of chloride are sulfate, respectively. A non-reference state on the order of 1 to 100 mg/L. Moreover, was above 3.8 and 77.4 mg/l for chloride and concentrations of these two chemical were sulfate, respectively. very low at reference sites in the ecoregion (i.e., older EPA criteria (EPA 1988) suggest The median concentration of sulfate (72.9 reference chloride concentrations for mg/L) from 2008 to 2010 was between the streams and rivers based on chronic toxicity non-reference and reference assessment to plant, fish, and invertebrate species are points (or in an intermediate state). below 230 mg/L, not the very low 0.4 mg/l Maximum sulfate levels (86.7 mg/L) did value seen in ecoregion reference sites) and exceed the ecoregion non-reference we are hesitant to conclude that there were assessment point value. All chloride problems at GRKO stemming from chloride concentrations from 2008-2010 were or sulfate during 2008-2010. non-reference when compared to the ecoregion assessment point. This suggests Mast (2007) showed that during 1980-2004 GRKO had concentrations of chloride and concentrations of most major constituents perhaps sulfate above the ideal as measured were lowest during high flow in May and at reference sites across the ecoregion. June when there were large contributions Anthropogenic sources of chloride and of dilute snowmelt to the river and then sulfate include fertilizers, a suite of industrial varied over a relatively narrow range from chemicals, sewage, irrigation drainage, September through April when streamflow combustion of fossil fuels, mining activities, is sustained primarily by groundwater inputs. We do not yet have sufficient SEI

38 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site data to confirm that these patterns still (in addition to N or P concentrations). exist, but similar to specific conductance, Proposed nutrient criteria are also specified concentrations of major ions (i.e., calcium, (MT DEQ 2011a) that differ from EPA magnesium, and sulfate) do generally have or previous state criteria in several useful an inverse relation with discharge measured ways. For example, they are specific to at the Deer Lodge USGS gauge. Most major summer months (July to September) when ions are derived primarily from geologic stream ecosystems are often most stressed weathering in the Clark Fork Basin and due to low-flow conditions and elevated are present in higher concentrations in temperature, they vary by ecoregion or even groundwater due to longer rock-water by stream (criteria for total N and P are interaction time. Groundwater is a more specific to the Clark Fork), and they were important contributor to the Clark Fork derived from a modeling exercise comparing during low flow periods (Mast 2007). During nutrient data to biological response and high-flow conditions, the influx of dilute human-perception of eutrophication. snowmelt water likely drives the overall decline in concentrations of major ions. We are fortunate to have the MT DEQ nutrient protocol available to apply within Total suspended sediment (TSS) SEI monitoring at GRKO. The SEI protocol concentrations increased from around 6 provides many of the data elements required mg/L during low-flow conditions to over to implement the MT DEQ methodology. 42 mg/L during high-flow conditions, However, some specific data collection and reflecting greater kinetic energy for erosion analytical requirements are not currently and transport. While these are relatively supported by SEI methods. For example, low concentrations of suspended sediment, a 12-sample minimum during a base we saw suggestions (significant, but based flow time period is needed as there is an on very small sample sizes) of positive allowed exceedance rate of 20% with a correlations between TSS and several specific statistical test required to ascertain metals and major ions. Currently, a likely significance of a possible impairment. limiting factor on loading of suspended Therefore, at this time we cannot fully or sediment to streams is the rate of supply appropriately apply the MT DEQ approach of material to the stream channel from to assessing nutrients and our current efforts natural or anthropogenic disturbance. As should be considered to only be laying the the restoration of the Clark Fork proceeds groundwork for possible future application concentrations may increase due to of the complete method. construction activity. The Clark Fork has long been recognized Nutrients as having higher nutrient concentrations Nutrients collected at SEI GRKO events (Lohman and Priscu 1992, Dodds et included several forms of nitrogen al. 1997). Work by MT DEQ (Tri-State (ammonium, nitrite + nitrate, total Water Quality Council 2009; Suplee et nitrogen, etc.), phosphorous (total and al. 2012) suggested that there were high orthophosphate), and organic carbon. The N and P levels on the Clark Fork in the State of Montana has developed a rigorous 1970s and 1980s with exceedance rates assessment methodology for determining of 50% for total phosphorous and 77% wadeable stream impairment due to for total nitrogen. Coupled with below excess N and P. Suplee and Sada de Suplee average precipitation and river flow this (2011), or supporting research conducted contributed to nuisance blooms of algae by MT DEQ and cited within, details (most noticeably, filamentousCladophora ). field, laboratory, QAQC, and analytical Interestingly, prior to the 1970s, toxic approaches for determining if a stream reach metals may have prevented development has a nutrient problem. The methods include of blooms. In 1998 nutrient and algae a tiered or decision pathway based approach standards were established for the river and and allows (or requires) supporting data a ten-year plan developed to lower loads. from productivity and biological metrics Suplee et al. (2012) present results collected over the ensuing 12 years. In summary,

Results and Discussion 39 nutrient reductions occurred in the basin Dissolved carbon (DOC) is not part of the but were less substantial in the upper Clark MT DEQ nutrient methodology; however, it Fork (including GRKO) than downstream is a useful constituent to include in long term of Missoula. There were no significant monitoring (Hauer et al. 2007). In Montana reductions in algal biomass. These patterns headwater streams DOC generally ranges largely match SEI N and P concentrations from 1 to 5 mg/L of C, whereas in larger riv- in 2007-2010: low and near detection limits ers or streams associated with wetlands val- (note that the laboratory used by the ROMN ues may approach 50 mg/L of C (Hauer and for nutrients has very low detection limits) Lamberti 2006). DOC values at GRKO were with only a few individual exceedances in within the range expected for larger rivers. select N and P fractions. Total nitrogen and nitrite + nitrate marginally exceeded a State In conclusion, while our methods do not of Montana provisional criterion on May 5th, meet the suite of requirements that would 2009. Likewise, total phosphorous exceeded allow MT DEQ to conclude that there was its proposed criterion on June 11th, 2010. any type of nutrient impairment on the However, SEI data on chlorophyll-a and Clark Fork at GRKO, our data largely match AFDM indicate that algal biomass was below similar work done on the river and indicate the proposed MT DEQ (2011a) criteria at that while nutrients are still present in the GRKO (suggesting there were no meaningful Clark Fork at GRKO they are probably not algal blooms during this time). This may problematic. The nutrient exceedances in be a reflection of spatial structure in algae SEI data are minor and infrequent, there are distribution along the river as qualitatively complex and unresolved patterns between the channel did appear to have significant nutrient concentrations and biological algae cover. Finally, Mast (2007) showed response and we cannot connect any point strong natural seasonal patterns in N and P or nonpoint source to nutrients at GRKO. at GRKO, with concentrations during early spring runoff as much as five times higher Metals in Water than baseflow summer concentrations. We focus on total metals given State of These patterns appeared in SEI data Montana criteria for total, rather than as well—we observed higher nutrient dissolved concentrations (except aluminum, concentrations during or soon after spring which has a dissolved fraction criteria runoff. from the State). Total metals represent all metal in a sample, whether they are bound In most samples, the highest proportion to sediment, are in particulate form or are of total N at GRKO was dissolved and dissolved. A dissolved metal fraction is highly labile nitrate. Similarly, dissolved the fraction that is in solution only. Note orthophosphate (the bioavailable form that data on metals only found bound to of phosphorus) was usually the highest sediments are presented below. While most proportion of total P. Labile nutrients are State of Montana criteria are for total metals, typically transported short distances before because it is primarily the dissolved form they are taken up by periphyton before that causes aquatic toxicity the EPA has being re-released or semi-permanently recommended that criteria for metals be bound in algal biomass (Hauer et al. 2007, re-expressed as dissolved concentrations Mulholland et al. 2000). As presented (EPA 1996). Therefore, we also include and below, bioassessment metrics in 2008 present dissolved metals including, where and 2009 suggest macroinvertebrates available, criteria following methods in EPA and (inconsistently) periphyton might be (2009a). Note that many of the aquatic- responding to excess nutrients, suggesting life water quality standards for metals vary that while nutrients and associated algal according to the hardness of the water and biomass were not excessive, the Clark we use adjusted values as needed. Fork was relatively productive during this time period and the structure of these Metals detected in SEI samples at GRKO communities may reflect this. included arsenic, cadmium, copper, iron, lead, manganese, and zinc. These

40 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site are all metals associated with Superfund periods. Sediment moves much more slowly, mine tailings mobilized and distributed and contamination in sediment can have throughout the basin by the floods during lasting effects on water quality and aquatic the late 1800s and early 1900s (Smith et al. life. Extensive data have been collected from 1998). Arsenic (total), one of the primary Clark Fork River bed sediments since 1985 concerns in the Clark Fork, ranged from by the USGS (e.g., Dodge et al. 2005, 2008) around 9 to 22 µg/L and was often above and other partners (especially Moore and its human health criterion. Manganese also Langner 2012) in support of the Superfund often exceeded its human health based process. Concurrent and complementary criterion. Lead and copper, two metals also monitoring of sediments in the Clark directly due to the mine tailing were always Fork will be an important part of the SEI above their aquatic life criteria in our data. protocol over time, especially as remediation However, we measured no exceedingly large progresses. There are no published State flushes of any metals (such as a 231µ g/L zinc of Montana numeric standards for metals level measured by Lambing (1991) during in sediments. However, MacDonald et an intense rainstorm on July 12, 1989 that al. (2000) developed Threshold Effect caused a fishkill). Concentrations (TEC) and Probable Effect Concentrations (PEC) for select metals in Patterns in metals in water and discharge the Clark Fork and we informally apply or season were difficult to resolve given these to our data. TECs and PECs are small sample sizes. However, using a consensus values derived from multiple larger data set, Mast (2007) showed that studies of the effects of metals in sediments metal concentrations at GRKO have a and are expressed as a geometric mean strong seasonal pattern with the highest of the individual thresholds these studies concentrations during or soon after peak indicated were potentially thresholds in the flow, perhaps due to greater amounts of impacts of concentrations of a given metal suspended sediment in the river during in sediment. They have been tested for their high-flow conditions (Lambing 1991). Metal reliability in predicting toxicity in sediments concentrations also show substantial diel by using matching sediment chemistry and cycles in the Clark Fork River (Brick and toxicity data. TECs are a lower (or reference) Moore 1996) with highest concentrations threshold, below which there are likely no during the night. This is likely driven by a diel impacts and PECs are an upper (or non- pattern in the sorption of metals between reference) threshold, above which impacts stream water and streambed sediments and are likely aquatic plants (Nimick 2003) with cycles in stream water temperature and pH appearing SEI data show that median and maximum to control the amount of each metal that concentrations of arsenic and copper in is adsorbed or desorbed. We will have to surficial sediments were several times the wait for a few more seasons of SEI data to PEC or non-reference threshold. Maximum confirm these patterns in ROMN monitoring levels of zinc also exceeded the PEC, with data, although we suspect that they have median and minima in an intermediate state. persisted. Median and maximum concentrations of lead were intermediate. Other metals with Metals in Sediment higher concentrations in SEI data included Metal contaminated sediments scoured aluminum, barium, and iron (however, there from upstream tailings ponds are the are no recognized assessment points for primary reason for the ongoing Superfund these metals and there elevated levels may be remediation of the Clark Fork and its due to natural sources). These results mirror floodplain (Smith et al. 1998, EPA 2004). patterns in both dissolved and total fractions Sediments are unlike surface water in that of water samples (see above) and are similar they typically have a longer residence time. to results in Moore and Langner (2012) and Water moves into and out of the Clark Fork Dodge et al. (2005, 2008). Patterns among River fairly rapidly. This can cause wide sediment metal concentrations, discharge, variations in water quality over short time and seasons were difficult to resolve given

Results and Discussion 41 small sample sizes, and will be addressed at 0.337 mg/kg, but well below the highest with more SEI data in the future. values seen in Scudder et al. (2009) for heavily mined watersheds. Moore and Langner (2012) have over 20 years of monthly data at three sites in the While the results for arsenic, copper, basin (including Deer Lodge) and less lead, and total mercury concentrations in frequent data at over 20 sites basin-wide the sediments of the Clark Fork are not that provide an excellent record of metal surprising given the history of the basin, in sediment dynamics in the Clark Fork. several factors suggest caution when Statistical models using these empirical interpreting these data. We have a small data suggest a decreasing trend in the sample size and our samples are episodic levels of many metals, including at Deer and not connected to events like storms Lodge, with complex interactions among that can mobilize metals in sediments. SEI scale, spatial structure, and season and sediment samples are composited across metal concentration. For example, the multiple depositional areas along the sample models suggest that the time for metal reach (where fine grained sediments collect), concentrations to fall below a TEC or PEC is which should reduce small-scale spatial highly dependent on starting concentrations variability. However, we cannot collect and background levels, with the degree samples from riffles where MacDonald et al. or success of the remediation playing a (2000) show concentrations are often 30 to relatively minor role in metal concentrations 40 percent lower than in depositional areas. over time. Riffles are often key habitat for benthos that support fish populations. As with other metals, there are no formal sediment criteria for mercury from the State Special Case: Water Temperature of Montana or from EPA (all criteria are In the following section we present for water column samples). MacDonald additional analyses of water temperature et al. (2000) provide TEC and PEC for data from SEI loggers collected in 2008 to total mercury in the Clark Fork of 0.180 2010 and the USGS Deer Lodge gauge from and 1.060 mg/kg, respectively. Median and 2001 to 2010. Our focus is on developing maximum total mercury concentrations baselines from these contemporary water were intermediate or higher than PEC on temperature data for comparison into the sediment dwelling organisms, but below future. We also use these data to conduct concentrations at which negative impacts a seasonally adjusted model of trend in are likely or expected to occur. However, water temperatures (presented in the trend because the primary toxic form of mercury is results section below). Daily mean stream methylmercury, total mercury-based toxicity temperatures from the USGS gauge and estimates are not expected to be highly the SEI loggers were highly correlated accurate. MacDonald et al. (2000) note (during their overlap in 2008-2010; r= that TEC correctly predicted total mercury 0.991, p<0.001). This suggests that stream toxicity only 34% of the time, whereas the temperatures at the nearby USGS gauge and consensus-based PEC correctly predicted at the SEI site are somewhat redundant (or toxicity 100% of the time. While the success that it is probably appropriate to use the of using total mercury and the TEC value is USGS gauge data to characterize patterns somewhat low, the consequences of actual in GRKO). Nevertheless, we feel that mercury toxicity are severe enough that maintaining the low cost SEI logger at the we feel it is still a useful threshold. Finally, GRKO bridge is warranted given the often Scudder et al. (2009) present data on stream unavoidable gaps in these types of data sets sediment total mercury samples collected and the NPS ownership of these important from streams in mined mountainous areas data. across the country between 1998 and 2005 that suggest the average concentration is Baselines around 0.175 mg/kg of mercury (total). We develop a baseline for future The median at GRKO was higher than this comparisons of stream temperatures using

42 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Table 7. Mean August stream water temperature estimated from daily mean values for water years 2000-2009 at the Deer Lodge USGS gauge. Data were not available for the 2006 water year. Sample sizes for other years was about 200. Water Year Mean August Temperature (ºC ) Standard Deviation 2000 17.2 0.96 2001 15.8 0.97 2002 17.9 1.05 2003 16.2 1.72 2004 13.6 1.97 2005 16.4 1.33 2006 -- -- 2007 16.7 1.65 2008 16.7 1.54 mean August temperature (again from flow is expected to be strongly affected daily values at the gauge; Table 7) suggests (Arismendi et al. 2012). From 2001-2009, a possible baseline of around 16.3º C spring through summer discharge and daily (the average across all water years). We mean temperature were negatively correlated use August temperatures because this is (r= -0.52, p<0.05). Low discharge tends be generally (July may also be) the hottest correlated with both slower water velocity month of the year and may be when warming and shallower areas in the channel which can trends are most evident as climate regimes create warmer stream water temperatures. shifts over time. August is also generally As we present below, at least the more recent when base flow is first reached during the years (2007 to 2010) in this time period water year and when stress from reduced were wet years with higher discharge than flows (and elevated water temperatures) the 30-year normal. This may suggest that may occur. However, the choice of a specific elevated spring and summer temperatures temperature statistic to use as a baseline with reduced flows are not as dramatic as is complex and we are also exploring they could be (i.e., if flows were lower, or alternatives (Dettinger et al. 2004, Isaak et al. more typical, water temperatures may have 2011, Arismendi et al. 2012). been even higher). Arismendi et al. (2012) demonstrated trends toward a shorter time We will continue to assess the usefulness lag between maximum stream temperature of this baseline, including comparisons to and minima in discharge at over 20 other approaches or baseline statistics, and undammed sites (like the Clark Fork) across use them in future reports as warranted. the west with a strong negative association However, several complex factors control between their magnitudes. Aquatic biota stream temperatures (i.e., latitude, elevation, in these systems may be increasingly groundwater, turbulence, direct solar experiencing narrower time windows to radiation, ambient temperatures, aspect, recover or adapt between extreme events of etc.) and these should also be considered in low flow and high temperature. evaluating the relevance of these baselines and in interpreting any observed changes Recent water temperatures and bull trout in stream temperatures relative to these assessment points values. The trend models presented below With the probable exception of the upper accomplish this to a degree by statistically end of the drainage (including GRKO), accounting for discharge and season. If we which has been degraded by contamination use the value given above we will need to by heavy metals, bull trout continue to be think at least about stream discharge and present (albeit sometimes in low numbers) potentially ambient air temperature in our in nearly all of the Clark Fork River interpretations. The discharge-temperature watershed (USFWS 2002). With successful relationship has important implications remediation of metal contamination, the under a shifting climate regime where stream GRKO reach may recover as a viable bull

Results and Discussion 43 Figure 9. Daily stream temperature by water year for Clark Fork River at GRKO bridge crossing: (top) 2008-2009, (middle) 2009-2010, and (lower) 2010-2011. Green dots show mean daily water temperatures based on the SEI temperature logger at the GRKO bridge. The grey line in 2008-2009 shows mean daily temperatures from the USGS gauge in Deer Lodge, MT. Blue lines indicate discharge (cfs) from the USGS gauge in Deer Lodge, MT Boxes and red line indicate temperature thresholds for functional bull trout populations (USFWS 1998): the red line is the summer daily thermal maximum for adult bull trout, the blue box is the fall spawning temperature range, and the pink box and is the range for winter/spring egg incubation. Note that box sizes vary slightly given the different temperature scales in each panel. Note that some of the USFWS (1998) thresholds may not be ideal for the Clark Fork at GRKO (especially for later summer spawning). MT DEQ is working on proposed temperature thresholds that account for the difference between optimal growth and long term toxicity testing. Moreover, while the Clark Fork at GRKO is a historic bull trout fishery, several other aspects of in stream and riparian habitat are not likely to be suitable anytime soon. Note that box size varies slightly given the different temperature scales in each panel.

44 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site trout fishery. Therefore, we informally apply temperature assessment points developed for bull trout Endangered Species Act assessments (USFWS 1998) to recent SEI or USGS temperature data. The bull trout assessment points define seven-day average temperature temperatures for functional bull trout populations during incubation (1º to 6ºC, October-February), rearing (15ºC as a maximum, June-August), and spawning (4º to 10ºC, mid-August to late October). Values that are outside these ranges or above the summer maxima will not likely support these important life stages/behaviors and therefore viable populations (USFWS 1998).

Stream water temperatures at GRKO during 2008-2010 were occasionally outside of or higher than the assessment points (Figure Figure 10. Mean seven-day maximum water temperature for water 9 and Figure 10). The daily values in Figure years, 2001-2010 at Deer Lodge USGS gauge. Horizontal red line depicts 9 must be compared with caution to the maximum summer temperature threshold (USFWS 1998) for optimal bull trout rearing. Dotted vertical gray lines delineate water year boundaries. assessment points for seven-day mean values, but qualitatively there are several periods longer than seven days that exceed an assessment point. Perhaps more relevant, over the 416 weeks in which seven-day mean data (Figure 10) were available from 2001 to 2010, 86% of the summer data were above the maximum rearing temperature assessment point.

Based on stream temperatures alone, the mainstem of the Clark Fork in and around GRKO may not be currently viable bull trout habitat. However, as some of the USFWS (1998) assessment points may not be ideal for the Clark Fork at GRKO (especially for later summer spawning), MT DEQ is working on proposed temperature assessment points that account for the difference between optimal growth and long term toxicity testing (Darrin Kron, Pers. Figure 11. Mean date of last (left panel) or first freeze (right panel) for water years, 2001-2010 at Deer Lodge USGS gauge. Historic and current periods are largely Comm. 2013). Moreover, arbitrary. Current is defined by the beginning of ROMN SEI monitoring in 2007 and while the Clark Fork at historic is considered to be 2001 to 2006. GRKO is a historic bull trout

Results and Discussion 45 fishery, several other aspects of in stream and relationship is significant. A model selection riparian habitat are not likely to be suitable procedure was used to select the specific anytime soon, as discussed below. form of the flow adjustment model used.

Freeze-thaw phenology We tested 17 parameters over variable Although our sample size is small, ten periods of record from 1978 to 2010 years of data from the USGS gauge suggest including discharge, temperature, field that the phenology of freeze-thaw cycles parameters, and select dissolved major and on the Clark Fork may be shifting (Figure trace elements (Table 8). All parameters 11). The last day of freezing in the spring with a significant trend included meaningful occurred earlier (around February 24th with seasonal adjustment, but the role of flow a standard deviation of six days) during varied with parameter. We did not test total 2007-2010 vs. March 6th (SD=14 days) recoverable trace-element concentrations during 2001-2006 (however, these two as Mast (2007) found no trends using these periods are arbitrary and this was not a data (for a shorter time period, but using significant result). The difference in timing similar models). of the first day of freezing in the winter was No significant trends were detected for marginally significant (p=0.06), occurring discharge, temperature, suspended sediment, st later (around November 21 with a standard and several metals, including arsenic. deviation of six days) during 2007-2010 vs. The trend in discharge was negative and th November 11 (SD=nine days) during 2001- marginally significant (p=0.13), in line with 2006. Importantly, more data and rigorous patterns seen across the region (Pederson modeling are needed to confirm these patterns et al. 2010). SEI data in 2008-2010 (see in stream temperature phenology. above) showed that arsenic almost always exceeded its criteria for all fractions in water Select Trends in Water and always in sediment—it appears this has Physiochemistry not changed much over the last 10 years. SEI data in GRKO proper are not yet of Flow adjustment (i.e., controlling for the sufficient duration to estimate any true trend important role of stream flow in temperature models. Therefore, we used data from the dynamics) did result in a positive slope USGS station in Deer Lodge to test for for minimum, mean and maximum water longer-term trends in select physiochemistry temperature (marginally significant at p=0.14 measures. We assume that given the close for minimum temperature), in line with proximity of the gauge these data are the weak suggestion of increasing stream relevant to GRKO. water temperatures in our simple analyses The ESTREND tool we use (Schertz of phenology (see above and Figure 11). The et al. 1991, Slack and Lorenz 2003) is ESTREND models treat stream temperature a regression model that computes an more carefully than our simple analyses. annual trend (slope), which represents the There were significant (but often small) median rate of change in concentration trends (both with and without flow or discharge for the selected period of adjustment) in pH (+), conductivity (-), record. Following USGS convention, trends dissolved copper (-) and dissolved zinc (-). were considered statistically significant at The magnitude and significance of these the 0.1 probability level. The model uses results were similar between unadjusted nonparametric seasonal Kendall tests or and flow-adjusted concentrations indicating a parametric Tobit model to account for patterns in discharge were not driving the variation across time due to season. The trends. When flow was accounted for, there specific number of “seasons” or form of the were trends in total suspended sediment (-), flow model adjustment for a parameter is dissolved calcium (-), dissolved cadmium determined by pattern in and the density (-), and dissolved lead (-). Flow adjustment of data. ESTREND may also use daily often strongly improved these models, mean discharge to adjust for flow-related especially for TSS. Keeping a sediment variability if the concentration-discharge

46 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Table 8. Results of the seasonal Kendall test for trends in discharge and select constituent concentrations on the Clark Fork River using data from the Deer Lodge USGS gauge. Bold type indicates statistically significant trend models (at p<0.10 following USGS Schertz et al. 1991, Slack and Lorenz 2003) in constituent concentration (or loading if flow adjustment is significant).

Unadjusted Flow-adjusted Flow Model # of Period of Parameter Trend p-value Trend p-value p-value Seasons Record Discharge (cfs/yr) -1.72 0.13 ------4 1978-2011 pH (standard units/yr) 0.005 0.08 0.009 0.05 4; <0.001 6 1985-2010 Temperature, maximum (ºC/yr) -0.15 0.48 0.20 0.37 8; <0.001 4 2001-2010 Temperature, minimum (ºC/yr) 0 1.0 0.19 0.14 8; <0.001 4 2001-2010 Temperature, mean (ºC/yr) 0 0.82 0.20 0.31 11; <0.001 4 2001-2010 Conductivity (umhos/yr) -3.5 0.003 -2.8 0.002 5; <0.001 6 1985-2010 Total Suspended Sediment -0.05 0.59 -1.26 0.03 1; <0.001 6 1986-2010 (mg/L/yr) Suspended Sediment <0.0625mm 0 0.84 0.10 0.45 4; <0.001 6 1985-2010 (mg/L/Yr) Calcium, dissolved (mg/L/yr) -0.2 0.23 -0.32 0.01 5; <0.001 3 1986-2010 Arsenic, dissolved (μg/L/yr) 0.06 0.14 0.02 0.57 1; <0.001 6 1986-2010 Cadmium, dissolved (μg/L/yr) / / -0.03 <0.001 TBR; <0.01 4 1985-2010 Copper, dissolved (µg/L/yr) -0.14 <0.001 -0.18 <0.001 1; <0.001 6 1985-2010 Iron, dissolved (µg/L/yr) 0 0.51 -- -- 1*; <0.001 6 1985-2010 Magnesium, dissolved (µg/L/yr) -0.01 0.49 -0.06 0.15 4; <0.001 3 1986-2010 Manganese, dissolved (µg/L/yr) -0.12 0.51 -- -- 1; ns 6 1985-2010 Lead, dissolved (µg/L/yr) / / -0.12 <0.001 TBR; <0.001 6 1985-2010 Zinc, dissolved (µg/L/yr) -0.37 <0.001 -0.37 <0.0001 1; <0.001 6 1985-2010 Notes: Data are adjusted for flow when there was a significant relationship between concentration and flow; Trends are significant using an alpha<0.10; “—“: not calculated; cfs/yr: cubic feet per second per year; µg or mg /L/yr: micrograms or milligrams per liter per year; µmhos/yr: microseimens per liter per year; ºC/yr: degrees Celsius per year; “*”: % censored>5 but assumed to be acceptable and a seasonal Kendall test for censored data used with no flow adjustment allowed (even if flow is significant); TBR: Tobit model, slopes, and p-value reported are from a Chi- square test of the model assuming a Gaussian distribution; “/” when using the Tobit model, flow adjustment must be included. load in a water column is dependent on due to mining cleanup efforts and other the energy contained with a high discharge changes or improvements due to generally so this is an intuitive result. Likewise, flow improved waste management (i.e., due to adjustment was key for cadmium and lead implementation of the Clean Air and Clean suggesting these parameters were perhaps Water Acts in the 1970s). Cadmium, copper, mobilized under higher flows. and zinc are associated with mine wastes. It may be that the decreasing trends in these The results from the trend models, although parameters reflect remediation efforts in the limited, to the periods of record, suggest headwater areas of the Clark Fork drainage water quality for several parameters has during the 1980s and early 1990s. However, improved over the last two decades or so at trends in other metals (arsenic, see above) Deer Lodge (and likely GRKO). Moreover, are less clear and Moore and Langner long-term trend analyses offer insight into (2012) suggest that the degree or success potential future changes in water quality. of remediation may play a relatively minor In particular the fairly large magnitude of role in metal concentrations over time, with the conductivity and TSS (at -3.5 µmhos/ spatial variation in initial concentrations yr and -1.26 mg/L/yr respectively) trends and background levels also being important. suggests they reflect real improving trends Short term pulses of contaminants may also in water physiochemistry (the reference occur with ongoing remediation efforts. state for the Clark Fork is likely fairly Continued monitoring will be important for dilute). These two parameters are useful assessing future Superfund cleanup efforts at general indicators of overall water quality. GRKO. The small increase in pH may also be

Results and Discussion 47 data (1998-2009) on nutrient dynamics in density of large wood, but less total wood. the Clark Fork basin that suggests significant Cover from woody riparian canopy was low declines in TP but not TN occurred along and there was a high level of anthropogenic the entire river. Most reductions occurred impacts, with some likely meaningfully downstream of Missoula. Nearer GRKO, large patches of invasive plants. Coupled nutrient reductions were less substantial with the fairly simple riparian buffer, this and there were no significant reductions in suggests that there are or could be localized algal biomass (see above for comparisons to disturbances to the river. Patterns in fine SEI data). The large P-load reduction was sediments are also seen (in subtle ways) in likely from a basin-wide phosphate laundry macroinvertebrate and (using older models) detergent ban set in 1989. diatom metrics.

Physical Habitat The SEI data presented in this report was Summary collected in somewhat wet years and given In general, much of the in channel and the importance of hydrology as a controlling riparian habitat at GRKO was in a non- variable in streams and rivers, this should reference state in 2009. Table 9 presents a be considered when assessing our results. simple summary with more detailed results However, over the long term, total annual below. stream flow may be decreasing in the Clark Fork, but more data is needed. We do not In particular there was some indication of have enough SEI habitat data to test for excess small sediments (sand and fines), trend—when we do this will be a focus of possibly related to a mildly incised channel. SEI analyses. However, while sediments were more mobile than in a reference stream, they were not so Habitat Metrics and Assessment Points much so that they would place the system Thirty-three core SEI physical habitat in a non-reference state in terms of bed metrics were generated for the Clark Fork stability. In stream habitat was characterized River at GRKO from data collected during by a fairly complex channel with a variable the 2009 sample event (Table 10). We chose thalweg profile and a high density of residual the final metrics from the many available pool. There appeared to be a relatively high Table 9. Summary condition table excerpt for select riparian and stream habitat metrics at GRKO in 2009. We include example vital signs and indicators, a brief description of results and patterns, and symbolize the status, trend, and our confidence in those summaries. See the Executive Summary for the complete Summary Condition Table. Vital Sign Summary Symbol (Example Indicators)

Fine substrates in the Clark Fork were more prevalent and had higher levels of embeddedness than select species of sediment intolerant macroinvertebrates and Habitat, sediment bull trout prefer. Bed load was also relatively substantial and mobile, but we need

(size, stability) additional investigation to link sediment load/source to biotic condition at the site. We have lower confidence in our assessment of sediments at GRKO (due to a small sample size).

Riparian cover was patchy on the Clark Fork and there were more adjacent potential stressors in the GRKO floodplain than at ecoregion reference sites. Invasive plants were fairly common, and some occurred with higher frequency than in In stream and riparian ecoregion reference sites. The park’s best management practices likely mitigate habitat most local habitat stressors, but upstream disturbance at GRKO likely has a com- (complexity, cover, disturbance) plex impact on habitat in the park. We need more data to confirm these issues and assess trend in these responses. We have lower confidence in our assessment of in stream and riparian habitat at GRKO (due to a small sample size).

Stream flow during 2008 to 2010 relative to the period of record suggest SEI monitoring occurred in somewhat wet years. Seasonal patterns were fairly typical Habitat, stream flow for mountain valley streams, but there was a suggestion of higher and earlier

(amount and timing) spring runoff. Over the long term, total annual stream flow may be decreasing, but more data is needed. We have medium confidence in our assessment of stream flow at GRKO.

48 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site x and -- 9.2 9.0 8.7 4.3 3.9 5.5 7.4 2.4 4.6 2.1 6.0 6.9 4.3 0.8 4.7 3.8 18.3 13.8 12.8 10.9 22.2 0.8** 5.5** reference EMAP Signal:Noise are in non-reference. are in non-reference. 2* >0.1 1b 3 1a bold red ------*; A*: <18 2 <34.8 <1.9 >0.43 Non-reference <-1.8 F*: * 2 <-0.4 5 4 2 2 3 1a 4 1b ------<6 <0 *; A*: <20 >34 <9.7 >1.2 2 >4.1 >0.04 Reference >-1.1 F*: 75 27 0.5 1.1 1.3 124 0.31 1.82 0.02 40.0 0.10 0.92 20.4 19.1 1.02 51.8 19.3 -4.30 -5.94 -1.29 23.87 52.45 0.016 0.00005 GRKO 2009 are either intermediate or non-reference (depending on the existence of a non-reference threshold). of a non-reference on the existence (depending intermediate or non-reference are either bold are in reference. Metric values in Metric values are in reference. bold green condition classes with the direction of inequality indicating whether the state is above or below a value. Metric values in or below a value. the state is above indicating whether of inequality classes with the direction condition points are for points 2009. Assessment Site, Historic National Ranch in Grant-Kohrs Fork River the Clark metrics for habitat physical of core Summary Channel sinuosity terrace to bankfull ht (m) Incision from Deviation of % streambed sand & finer from exp.(excess sand+fines”, %) sand & finer from Deviation of % streambed (fining index, %) expected Deviation of log mean substrate diameter from at bankfull (log) Mean bed particle diameter/critical (mobile) diameter Pool density (m of residual pool / m of sampled stream) Pool density (m of residual silt & finer (“fines”; <0.6 mm in diameter) (%) Streambed sand & finer (<2 mm) (%) Streambed Substrate embedded by sand and fines (%) exp.(excess fines, %) silt & finer from Deviation of % streambed Number of large woody debris (pieces /100m of channel) Number of large woody (%) cover of aquatic macrophytes Areal (detectable by the unaided eye, %) cover of filamentous algae Areal ledges (%) banks, overhanging veg, boulders and rock undercut cover of woody debris, brush, Areal Mean residual depth (cm) Mean residual (cm) expected from depth Deviation in mean residual (dimensionless) of variation Thalweg depth coefficient in the bankfull channel (m3/m2) of woody debris Volume Bankfull depth/wetted depth (morphometric index of “flashiness”) Bankfull depth/wetted depth (morphometric index at the bank, %) Canopy density (measured layer vegetation (%) Riparian canopy+mid+ground Bankfull width/depth potential) Bankfull width/wetted width (flood inundation (inverse index of “droughtiness”) Low flow/annual mean runoff Relative Bed Stability Floodplain Interaction Substrate Cover for Stream Biota Habitat Volume Habitat Complexity SEI Core Physical Habitat Metrics Physical Habitat SEI Core Riparian Vegetation Cover Riparian Vegetation Hydrologic Regime (Habitat Based) Table 10. 10. Table non-reference in Metric values

Results and Discussion 49 - x -- -- 8.0 2.8 6.0 1.6 4.2 11.2 16.6 EMAP Signal:Noise percentile values for percentile th /<5 th 3 1a ------<0.23 <0.04 Non-reference 3 1a ------>0.67 >0.07 Reference 41 0.0 1.72 0.18 2.11 0.67 0.61 0.002 GRKO 2009 See Table 10 See Table percentile values for reference/non-reference, respectively respectively values for reference/non-reference, percentile th />95 th . 2009 (continued) Site, Historic National Ranch in Grant-Kohrs Fork River the Clark metrics for habitat physical of core Summary Riparian canopy+mid+ground layer woody veg (%) layer woody veg Riparian canopy+mid+ground Frequency of individual invasive plants of individual invasive plants Frequency of all target invasive plant Frequency Riparian/near-stream agriculture – all types (proximity weighted index) weighted – all types (proximity agriculture Riparian/near-stream index) . Weighted (prox roads Riparian/near-stream index) . Weighted (prox agriculture rowcrop Riparian & near-stream index) . Weighted (prox dikes, revetment walls, Riparian/near-stream inverse index) . Weighted all types (prox Human disturbances of inverse index) . Weighted (prox Riparian vegetation alteration Assessment points generated from reference sites in the Middle Rockies ecoregion for indicators where increasing value is associated with a declining condition and thresholds in ecoregion reference site reference in ecoregion condition and thresholds value is associated with a declining increasing for indicators where sites in the Middle Rockies ecoregion reference Assessment points generated from Assessmentpoints generated from reference sites in the Middle Rockies ecoregion for indicators where decreasing value is associated with a declining condition at the >25 value is decreasing for indicators where in the Middle Rockies ecoregion sites reference from Assessmentpoints generated SEI Core Physical Habitat Metrics Physical Habitat SEI Core Human Disturbances Assessment points from Stoddard et al. (2005) Stoddard Assessment points from (2008, 2010); value for percent fines is for bull trout and we use this over the USFWS (1998) value as it is more conservative, value for percent sand is for sediment sensi conservative, value for percent this over the USFWS (1998) value as it is more and we use fines is for bull trout Bryce et al. (2008, 2010); value for percent Assessment points from . USFWS (1998) of > 0 habitat suitability.03 pools per meter from used to assess bull trout . This is similar to a threshold meter of stream MT DEQ (2010b), converted to pools per points from Assessmrent Assessment points from USFWS (1998) for “Functioning Appropriately” bull trout populations (scaled to the sampled reach length at GRKO of 500 m) populations (scaled to the sampled reach bull trout Appropriately” USFWS (1998) for “Functioning Assessment points from Higher S:N values suggest the metric had more signal or true information than variation due to crews, season and other sources (see text and Kaufmann et season and other sources true information than variation due to crews, signal or . Higher S:N values suggest the metric had more S:N is the ratio of information to noise in a metric 4 . tive macroinvertebrates 5 x (“Mountains ”). bioregion EMAP sites in a broader from data are et al. 2005). S:N source al. 1999, Stoddard reference/non-reference, respectively reference/non-reference, 1b at the <75 distribution are 2 3 14). states (see Figure and non-reference reference (A) channels for (F) and armored given for fining * Assessment points values for RBS are Notes 1a Table 10. 10. Table

50 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site largely following Kaufmann et al. (1999). determinant of the quality of a stream. The We interpret metrics following the general general size of a stream is a function of its ROMN assessment approach as outlined drainage area and discharge throughout above. Only a few have meaningful existing the year; however, anthropogenic activities assessment point(s) that we felt were frequently alter channel dimensions and relevant to GRKO. In particular, the State of flood or low flow discharge, all of which Montana, as part of the TMDL process for can alter the quantity and quality of aquatic the Upper Clark Fork basin, has developed habitat. “targets” for select habitat features and we use a few of these as reference assessment SEI data can be used to generate simple points (MT DEQ, 2010d). We also derive metrics of habitat volume such as mean assessment points for select metrics using wetted width and depth. While useful, Middle Rockies ecoregion reference sites. these metrics are sensitive to the flow stage For simplicity, we restrict the number of on the given day of sampling and do not assessment points derived from ecoregion necessarily reflect habitat volume during reference sites to a few key habitat metrics— limiting drought conditions. Therefore, we others may be derived upon request or estimated a residual mean depth metric in future reports. Note that the sample that accounts for stream flow (following size of ecoregion reference sites was a Kaufmann 1987, Stack 1989, and Kaufmann bit small (~35); however, a new data set et al. 1999). Mean residual depth is the from the EPA National Rivers and Stream mean across the sampled reach in the Assessment program will soon be available difference in depth or bed elevation between that will improve this. We interpret habitat a pool and its downstream riffle crest (i.e., metrics with no assessment points using hydraulic control point). This metric is qualitative comparisons to predicted independent of instantaneous discharge values from models or via comparisons to as it is based on the profile of the stream average conditions across reference sites bottom along its deepest point (regardless of in the ecoregion. All interpretations are how much water is in the channel). Reach- done in the context of the ecology of the wide residual mean depth varies naturally Clark Fork River and management needs across references streams as a function of for GRKO (i.e., maintaining the historic stream size and other geomorphological context of the Grant Kohrs ranch). Finally, attributes. However, within these natural as we accrue additional SEI data, we (will) gradients, Kaufmann et al. (2008, 2009) generate baselines to compare change in and demonstrate that anthropogenic stressors eventually trend in habitat condition. also play a strong role in the size and shape of streambeds and sediment dynamics such Quality Assurance and Quality Control that metrics like residual mean depth are We only have habitat results from one full somewhat independent of stream size (but sample event at GRKO and therefore have see below). Therefore, while caution must be no estimates of spatial or temporal variability used, ecoregion reference values generated (while SEI habitat data has many internal from ecoregion reference sites that span a subsamples or replicates, the proper scale for wide range in stream type and watershed SEI habitat data and metrics is the complete size are likely appropriate (Kaufmann et sample reach). Until we generate our own al. 1999, Stoddard et al. 2005). The mean QAQC data, as a rough proxy of precision residual depth at GRKO in 2009 of 20.4 cm in our habitat metrics we include estimates was higher than the ecoregion reference th of variability from the EMAP program (see assessment point (25 percentile) of 4.1 th above) and discuss them in relevant sections cm and well above the non-reference (5 below. percentile) of 1.9 cm (Table 8). This suggests the river at GRKO had more habitat volume Habitat Volume than reference streams across the ecoregion Habitat volume, or the amount of in-channel —or likely a meaningful amount of residual habitat available to stream biota, is a major habitat volume in the channel, even at low flow.

Results and Discussion 51 Mean residual depth is controlled to a large complexity is difficult to quantify, yet SEI degree by the size and power of the stream, protocols provide for several metrics that which in turn vary with geoclimatic setting may reasonably estimate this attribute of (i.e., drainage area, runoff, and slope; Stack streams. We selected three to interpret here: and Beschta 1989). Therefore, another useful (1) coefficient of variation in thalweg depth, metric for understanding habitat volume (2) the volume of woody debris per m2 of is the deviation of mean residual depth bankfull channel, and (3) the density of from a predicted value that accounts for large wood (Kaufmann 1999, Stoddard et al. geoclimatic setting. This metric may better 2005). deal with the confounding effects of stream type and watershed size on residual volume. The coefficient of variation in thalweg depth We estimated the predicted mean residual directly measures how variable (or complex) depth at GRKO as 1.06 cm using a model the bottom of a stream is. Higher values (on developed by the EPA (Stoddard et al. 2005). a scale of 0 to 1.0) suggest greater dispersion The model was derived from data collected among replicates of thalweg depth (or a by EMAP at over 1,000 streams across the more complex channel profile). The thalweg West and predicts mean residual depth profile coefficient of variation at GRKO was from basin area, long-term mean annual 0.31 (Table 10). There are no assessment precipitation, channel slope, and lithology. It points or predicted values available for this is important to note that this predicted value is metric. In qualitative comparisons (using not necessarily a reference condition (as used EMAP data) the Clark Fork at GRKO in other interpretations of SEI data). The EPA appears to have similar variability in its model is based on streams and rivers ranging thalweg profile to reference sites across the from pristine to disturbed, it merely corrects ecoregion (the mean coefficient of variation for geophysical drivers. The deviation of the in thalweg depths across the 39 ecoregion measured mean residual depth at GRKO in reference sites was 0.36 with a standard 2009 from its predicted value of was 19.1 deviation of 0.1). cm. A value of zero would suggest the Clark Woody debris is one of the most important Fork had a mean residual depth exactly as sources of structure (or complexity) in predicted or expected given the geoclimatic most stream types (Figure 12; Harmon et setting of GRKO. The large amount of al. 1986, Robison and Beschta 1990) and we residual habitat volume relative to what used two woody debris metrics as estimates is expected based on geoclimatic setting of habitat complexity. First, we used the further suggests that even at low flow there volume of woody debris per m2 of bankfull are likely to be remnant pools available for channel. This metric incorporates both the fish and other stream biota. frequency and amount of wood of any size. We use a raw value versus the log, as used Signal to Noise Ratio in Stoddard et al. (2005), to more intuitively The S:N of both the predicted (or scaled) include streams with no (or zero) wood and unscaled habitat volume metrics was volume (where a log value is undefined). The acceptable within the EMAP program, Clark Fork at GRKO had a volume of large with high values suggesting that they are woody debris (LWD) per area of bankfull estimating conditions in GRKO with channel of around 0.00005 m3/m2 (Table acceptable precision. It is more difficult to 10). There are no established assessment count wood within volume (size) classes as points available for this metric. In qualitative required for the volume metric. comparisons (using EMAP data) the Clark Fork at GRKO appears to have much Habitat Complexity lower total wood volume than reference All else being equal, more complex stream sites across the ecoregion (the mean wood habitat should generally support greater volume across ecoregion reference sites biodiversity and a more intact, higher was 0.01 m3/m2, with a standard deviation functioning system (Gorman and Karr of 0.01). Second, we used the density of 1978, Benson and Magnuson 1992). Habitat large wood (>0.3 m in diameter and >5 m in

52 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site NPS/BILLY SCHWEIGER

Figure 12. ROMN SEI field crew collect data on stream physical habitat in the Clark Fork in 2008. The area in the photo shows good examples of woody debris, complex off channel habitats, overhanging banks and riparian cover. length) given an assessment point developed Signal to Noise Ratio for bull trout habitat assessments by USFWS The S:N of the habitat complexity metrics (1998). The density of larger woody debris was generally acceptable within the EMAP was 1.82 per 100 m of stream at GRKO. program, with high values for thalweg depth Using the assessment points in USFWS variation and woody debris density, but only (1998) and similar guidance in Dunham et a moderate value for woody volume; it is al. (2003) and USFWS (2002) we estimated more difficult to count wood within volume that maintenance of a bull trout fishery in (size) classes as required for the volume the Clark Fork would require more than metric. This suggests that these metrics around 1.2 piece of large wood per 100 m of may have mixed capacity to estimate stream stream length (note that many other habitat habitat complexity at GRKO with acceptable conditions would also be needed). This precision. was exceeded at GRKO in 2009 suggesting that there was enough habitat complexity Cover for Stream Biota from large wood to support viable bull trout Another key function of stream habitat is the populations. Clearly many other attributes support of aquatic biota through physical are required for a bull trout fishery (and cover (i.e., for protection from predators or several may not be met in GRKO, i.e., as spawning habitat). While several of the stream temperature as described above). habitat metrics we interpret in this report are Our data on wood at GRKO are complex. indirectly related to this function, following They suggest that supply of large wood to Stoddard et al. (2005), Hankin and Reeves the Clark Fork, which depends on active (1988) and Kaufmann and Whittier (1997), floodplains that move naturally eroding we focus on four direct metrics of in-channel banks that then deliver more substantial cover. These include: (1) the cover of aquatic woody debris to a channel, may be largely macrophytes, (2) the cover of filamentous intact. However, wood volume (across all algae, (3) the combined cover of natural sizes) is lower than reference sites across features, and (4) pool density. the ecoregion, perhaps because of reduced amounts of smaller woody debris given the There are no established assessment points agricultural land uses upstream of GRKO. or regulatory criteria available for the first two cover metrics (macrophytes and algae).

Results and Discussion 53 In qualitative comparisons (using EMAP are to be quantified, but they are probably data), the Clark Fork at GRKO tended to most relevant at base flow when salmonids have low cover of macrophytes compared are most stressed. We used residual pool to that seen in reference sites across the depth measured at base flow to develop our ecoregion. While excessive macrophytes can metric. This includes all possible pools, not be problematic (indicating excess nutrients just those seen easily from the surface. The or reduced stream flow), the very small cover USFWS bull trout assessment point suggests at GRKO suggests that there may have been there may not be sufficient pool to support a lack of this cover type for fish and other a bull trout fishery at GRKO. However, biota in the channel. Algal cover was over 35 more generally the amount of residual pool times the mean ecoregion value (75 vs. 2%). at GRKO was higher than the mean value Filamentous algal beds can be important across ecoregion reference sites (0.004 pool habitat for small fish, but the high cover at meters per meter of stream) and we do not GRKO in 2008 was excessive. Suplee (2008) think the amount of pool in the Clark Fork suggests that in the Northwestern Glaciated is of real concern (unless a bull trout fishery Plains ecoregion (the high-line region of the becomes a management goal). state to the east of GRKO), streambed cover by filamentous algae should generally be less Signal to Noise Ratio than 30% for a single sampling event and The S:N of the cover for stream biota metrics less than 25% for the summertime average. was variable within the EMAP program, with Algae cover in more mountainous areas like low values for the cover of woody debris and GRKO should be even lower. As discussed macrophytes, high values for algae cover, above, the Clark Fork often has higher and moderate values for cover of all natural nutrient concentrations (Lohman and Priscu features. This suggests that these metrics may 1992, Dodds et al. 1997) that led to nuisance have mixed capacity to estimate conditions blooms of algae beginning in the 1970s in GRKO with acceptable precision. Wood and 80s. Suplee et al. (2012) suggests these and macrophytes cover can vary over time continued through at least 2011—matching naturally and can be hard to safely estimate our data. with precision in challenging stream habitats.

An assessment point from Stoddard et al. Substrate (2005) for the third metric (natural cover Stream substrates are major components types including woody debris, brush, of in-channel, shoreline and bank habitat undercut banks, overhanging vegetation, and can be important elements of overall boulders and rock ledges) suggests reference stream condition. In particular, substrate streams should have more than 34% cover size can impact the species composition from all natural features. At GRKO this of macroinvertebrate, periphyton, and metric was only 27% indicating that cover fish assemblages (Hynes 1972, Cummins of natural features in the Clark Fork was not 1974, Platts et al. 1983, Barbour et al. 1999). comparable to reference streams. However, Along with bedform, substrate particle natural cover at GRKO was intermediate size influences the hydraulic roughness, and above the assessment point for a non- water velocity, and the interstices among reference state (18%). substrate particles that provide habitat for Measured pool density on the Clark Fork macroinvertebrates and smaller vertebrates. (0.02 per meter) was lower than a State The sizes of substrates and there relative of Montana (MT DEQ 2010b) target or extents on streambeds are often sensitive reference pool frequency for the Clark Fork indicators of the effects of human activities of >0.04 per meter of stream. Likewise it was on streams (MacDonald et al. 1991, below an assessment point used to assess Barbour et al. 1999). Accumulations of fine bull trout habitat suitability (>0.03 pools substrate particles can fill interstitial spaces per meter; USFWS 1998). The MT DEQ and reduce habitat space for benthic fish, or USFWS bull trout assessment point do macroinvertebrates and diatoms (Platts et al. not specify the stream flow at which pools 1983, Rinne 1988). In addition, fines impede

54 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site the circulation of oxygenated water into on sediment assessment (Kusnierz et al. hyporheic habitats. Decreases in the mean 2013) and will be publishing formal western particle size and increases in streambed fine Montana sediment assessment methods and sediments can destabilize stream channels assessment points that may be pertinent to (Wilcock 1997, 1998) and may indicate GRKO and we will use these in the future as increases in the rates of upland erosion applicable. and sediment supply from anthropogenic disturbances (Lisle 1982, Dietrich et al. Percent fines, percent sand and percent 1989). Because of these and many other embeddedness were all above their important aspects of stream substrate, we assessment points at GRKO in 2009 (Table include six substrate metrics following 10). This suggests that the percentage of Stoddard et al. (2005), and Kaufmann et al. small, often embedded sediments in the (1999): (1) percent silt and finer (“fines”; channel were higher than the reference <0.6 mm), (2) percent sand and finer (<2 conditions for bull trout and select sediment mm), (3) percent embeddedness, (4 and sensitive macroinvertebrates. These 5) the deviation of fines and sand from results fit many of the biological responses predicted values, and (6) deviation of the seen in sediment metrics derived from log mean substrate diameter from expected macroinvertebrates and diatom assemblages (“Fining Index”). at GRKO as presented below. However, as shown above, sediment delivery to the bed Bryce et al. (2008, 2010) define select had a negative trend (using total suspended sediment assessment points based on a sediment; TSS), over 1986-2010 and suite of quantitative and semi-quantitative relatively low levels in 2008-2010. Therefore, approaches (including correlative studies sediment issues may be improving at GRKO. and examination of key species tolerances). We use a Bryce et al. (2008, 2010) reference Substrate sizes and density vary naturally in assessment point for fines (<6%) developed streams with different drainage areas, slopes, for bull trout (note it is more conservative and surficial geologies (Leopold et al. 1964, than a similar assessment point from USFWS Morisawa 1968). Therefore, other useful 1998). We feel that even though the Clark metrics for understanding sediment are the Fork at GRKO may never be managed as deviation of percent fines, sand and mean a bull trout fishery and may not be viable substrate size from predicted values that habitat for other reasons (i.e., temperature, account for geoclimatic setting (Kaufmann see above) this provides a useful comparison et al. 2008). To estimate predicted values we based on an important potential biological use models derived from data collected by function for the river. Bryce et al. (2008, EMAP at over 1,000 streams across the West. 2010) provide a value for percent sand The models predict values based on bankfull (<9.7%) based on sediment sensitive bed shear stress (see below), catchment macroinvertebrate assemblage response mean annual precipitation, ecoregion (using some of the same genera that are in type, and the method used to quantify the our GRKO samples). The State of Montana bed substrate. It is important to note that (MT DEQ 2010b) derives a very similar predicted values are not necessarily a reference target percent sand value for the Clark Fork condition (as used in other interpretations of <10% (we use the Bryce value as it is more of SEI data). The EPA models are based on specific and conservative). The best available streams and rivers ranging from pristine to assessment point for percent embeddedness disturbed, they merely correct for geophysical (<20%) is for bull trout from the USFWS drivers. The deviation of percent fines and (1998). Note that there are State of Montana percent sand from their predicted values narrative standards for sediment that state was -4.3, and -5.9, respectively (Table 10). that there are to be no increases above A value of zero would suggest the percent naturally occurring concentrations (MT of each substrate size was as predicted or DEQ 2002); however, as this is vague we use expected, given the geoclimatic setting. the numeric values described above. Finally, Interestingly, the negative values for percent MT DEQ has developed working guidance fines and sands suggests that GRKO had a

Results and Discussion 55 lower measured percentage of these small led to reduced stream condition. Finally, sediment size classes than expected based on because RBS depends on both the supply of the geoclimatic setting of the Clark Fork (or sediment and the capacity for sediment to be in other words, there should be even more transported, it may have a direct connection fines and sands at GRKO than the relatively to many anthropogenic disturbances. high values presented above). However, the deviation from the predicted mean substrate Following Kaufmann et al. (1999) and size was 0.10 (Table 10), suggesting that the Kaufmann et al. (2008), RBS is estimated average sediment size was a bit larger than as the ratio of observed geometric mean expected. Sediment size in the Clark Fork bed surface particle diameter (from the 155 was somewhat bimodal, with excessive substrate particles classified by field crews) smaller size classes but also higher density of and the critical maximum mobile particle cobble and above sized particles. diameter estimated for a streamflow likely to mobilize the streambed. The critical Signal to Noise Ratio bed particle diameter in a natural stream The S:N of the both the unscaled and is a function of the average streambed predicted bed particle size metrics was tractive force, or shear stress. Shear stress easily adequate within the EMAP program, is estimated using principles of sediment suggesting that they are estimating transport theory (e.g., Simons and Senturk conditions in GRKO with acceptable 1977) and quantitative SEI data on bankfull precision. For the predicted or scaled flow depth, slope, channel shape, and metrics this is despite removing the portion roughness, including the mitigating role of of the variance in bed particle size due to large woody debris and channel complexity natural geomorphic variability. (Kaufmann et al. 2008). Values of RBS for a given stream type and Relative Bed Stability geoclimatic setting can be either to low (finer Another metric used to characterize the substrates with a more unstable streambed) substrate of the Clark Fork is relative bed or to high (coarser substrates and to stable stability (RBS). RBS measures the tendency of a streambed) than expected (given the of streambed particles to resist transport complex placement of reference assessment under a given streamflow condition. points on both ends of the scale of RBS, Basically, the metric compares the measured Figure 13 shows these values graphically). median substrate size in the streambed to Lower than expected streambed stability the maximum substrate size carried during may result either from high inputs of fine bankfull events. In other words, if a stream sediments (i.e., when the anthropogenic is capable of moving large boulders during supply of sediments from the landscape bankfull flow events, yet the measured exceeds the ability of the stream to move median substrate is fines and silts, a RBS them downstream) or increases in flood metric would suggest that excessive fine magnitude or frequency. Low RBS resulting sediment loading is present. RBS is a from hydrologic alteration can be a critically important component of habitat precursor to channel incision. In contrast, structure for aquatic organisms. Species are high bed stability is typified by hard, adapted to natural streambed movement— armored streambeds, such as those often too much or too little can cause stress and

Fining, non-reference Fining and armoring, reference Armoring, non-reference

-1.8 -1.1 -0.4 0.1

Figure 13. Clarification on arrangement of finning and armoring reference and non-reference assessment points for relative bed stability (log RBS; Stoddard et al. 2005). Unlabeled yellow regions are intermediate between reference and non-reference states. Black arrow indicates approximate location of 2009 GRKO RBS value.

56 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site found below dams where fine sediment increases channel slope, and can lead flows are interrupted, or within channels to lower bed stability, greater sediment where banks are highly altered (e.g., paved transport, increased bank instability, channel or lined with rip-rap). Note that the data downcutting, and downstream flooding. presented here were collected before GRKO There are no assessment points or predicted began armoring banks within the SEI sample values available for this metric. In qualitative reach as part of the Superfund remediation comparisons to EMAP data, the Clark process. Fork at GRKO in 2009 had slightly lower sinuosity (1.1) than reference sites across In 2009 the observed mean particle size the ecoregion (mean log sinuosity=1.17, at GRKO was 1.77 mm and the estimated SD=1.23). However, at this point, it is critical maximum mobile diameter, 36.9 mm. unclear if reduced sinuosity at GRKO is Using reference assessment point values any cause for concern. The scale of the SEI developed by Stoddard et al. (2005; Table 10 sample reach may not be quite large enough and Figure 13), relative bed stability on the to appropriately measure this attribute of the Clark Fork (by convention this is expressed Clark Fork (a larger river that will express its as a log; RBS=-1.29) was intermediate, but sinuosity across a larger reach). We restricted more on the finning side of the gradient. This the sample reach to 500 meters as per the SEI suggests that streambed substrates at GRKO protocol where the ideal length of the reach were more mobile (but not markedly so) is likely closer to 750 meters based on the than in reference streams and is consistent mean wetted width at base flow. In addition, with the higher percentage of fine substrates the section of the river we sample appears seen in the river. less sinuous in aerial imagery than other reaches at GRKO. Signal to Noise Ratio The S:N of the RBS metrics was high Stream incision results when bankfull stage within the EMAP program, suggesting that is at a lower elevation than the top of either it is estimating conditions in GRKO with stream bank and is usually caused by an acceptable precision. increase of the erosive ability of a river, a decrease in the bedload sediment supply, Floodplain Interaction or when the stream bottom and banks are Following Stoddard et al. (2005), Madej destabilized by disturbances. There are no (2001) and Kaufmann et al. (1999), we use assessment points or predictive EPA models four metrics to estimate the interaction available for this metric. Therefore, given the between the Clark Fork channel and its importance of incision to many aspects of floodplain at GRKO. These include: (1) stream habitat condition (Kaufmann et al. channel sinuosity, (2) incision, (3) bankfull 1999) we compared the 2009 GRKO value width to depth ratio, and (4) bankfull width to assessment points from the distribution to wetted width ratio. of reference sites from across the ecoregion (75th percentile for reference=0; 95th Channel sinuosity of a stream or river reach percentile non-reference=0.43; note that is the ratio of channel length (as the water increasing incision is associated with a flows) divided by the straight-line distance declining condition, so a reference state between the two ends of a sample reach. is below the 75th percentile and a non- It ranges from 1.0 (straight) to around 3.0 reference is above a 95th percentile). Incision (tortuously twisty). Channel slope should at GRKO (1.2 meters in 2009) was in the be considered when interpreting sinuosity 97th percentile of the reference distribution, as very steep channels are more likely to be well above the 75th percentile reference straight naturally (the Clark Fork is fairly assessment point (the reference assessment flat so this is less relevant). The greater the point is no incision). This result is consistent sinuosity, the slower water flows through with patterns seen with stream sediments at a valley and the more chance it has to GRKO and with the agricultural Clark Fork interact with the terrestrial landscape. watershed. Incision can be caused by local Channel straightening decreases sinuosity, and/or upstream disturbance such as grazing

Results and Discussion 57 of riparian vegetation that protects stream Following Stoddard et al. (2005), we use two banks and channels. general metrics derived from habitat data as indirect indicators of the hydrologic regime. Our last two metrics of the interaction These are in addition to the direct measure between the Clark Fork and its floodplain, of stream flow using USGS data from the bankfull width to depth ratio and the nearby gauge as presented below. The habitat bankfull width to wetted width ratio, are based measures include: (1) the ratio of base relatively simple measures of the response flow to annual mean runoff, and (2) the ratio of the channel to changes in sediment of bankfull depth to wetted depth (both supply and transport. Channels that are at the time of sampling). The first index is over widened relative to their depth are an index of “droughtiness” that compares often associated with excess sediment instantaneous discharge (corrected for deposition and streambank erosion, contain watershed area) to the average annual shallower and warmer water (White et runoff (using a model developed by Wolock al. 1987), and provide fewer deep water and McCabe (1999) and precipitation, habitat refugia for fish (MT DEQ 2011c, evapotranspiration, topographic data from Kasahara and Wondzell 2003). There are no 1951-2000 within the Clark Fork watershed). established assessment points or regulatory If the instantaneous baseflow is high criteria available for these two metrics, and relative to the average, the site is likely not we only make qualitative comparisons to experiencing a drought. The second index values across the ecoregion (using EMAP is a morphometric index of “flashiness.” It data). The bankfull width-to-depth ratio estimates a range or the degree of variation at GRKO in 2009 (19.3) describes a fairly in a sites hydrologic regime by comparing wide, shallow channel, notably larger than high and low flows with bankfull depth the mean ecoregional reference site value as the high flow proxy and wetted depth of 8.5 (SD=1.5). Wider bankfull channels at base flow as the low flow proxy. If the relative to their wetted width often are ratio of high to low flow is high, peak flow related to higher fine and smaller sediments is much greater than baseflow suggesting and steep incised banks. The bankfull width the channel experiences more significant to wetted width ratio (1.02) was fairly low flooding and is more flashy (or variable) relative to the ecoregional reference mean in its hydrology. Both of these metrics are of 1.35 (SD=1.2). In general, both of these somewhat crude approximations, yet they metrics suggest that the Clark Fork had more may add a more integrated perspective to the constrained access to the flood plain in 2009. classic measures of hydrology (i.e., stream discharge) we also utilize. Their primary Signal to Noise Ratio value may be from simple comparisons to The S:N of the floodplain interaction ecoregional reference data and in future metrics was moderate within the EMAP trend analyses. program, suggesting that they are estimating conditions in GRKO with acceptable There are no established assessment points precision. or regulatory criteria available for these two metrics, and we only make qualitative Hydrologic Regime (Habitat-Based) comparisons to values across the ecoregion The hydrologic regime, or the amount and (using EMAP data). The Clark Fork at timing of streamflow, and its interaction GRKO had a ratio of base flow to annual with other elements of river habitat is a mean runoff in its watershed of 0.016. This fundamental component and determinant of is comparable to the mean across ecoregion the condition of rivers. For example, human reference sites (0.013) and suggests that use of water in rivers like the Clark Fork (i.e., potential water stress or localized drought withdrawal for irrigation) can cause stress was not much different from that seen in the to river ecosystems (Poff et al. 1997) when reference condition. The ratio of bankfull changes in hydrology disrupt reproduction, to wetted depth or flashiness index in 2009 survival, spawning, and migration of biota (1.3) was lower than at ecoregion reference (Poff and Ward 1989, Junk et al. 1989). sites (2.14), suggesting that the river was

58 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site a bit less flashy than reference sites in the two of these metrics, and we only make ecoregion. As discussed below, 2009 was a qualitative comparisons to values across the fairly high water year, which would increase ecoregion (using EMAP data). The Clark mean wetted depth (lowering the ratio with Fork at GRKO had lower canopy density on bankfull depth). This can also happen in its banks than the mean across ecoregion rivers like the Clark Fork with return flows reference sites (51.8 vs 83.1 (SD=2.5); from irrigation water. Table 10). Given the degree of historical disturbance in the riparian system of the Signal to Noise Ratio Clark Fork this is not a surprise. It will be The S:N of the habitat-based hydrologic important to track this over time as the regime metrics was very low for the restoration of the floodplain proceeds. The droughtiness index, but acceptable for lack of canopy may contribute to potentially the flashiness index within the EMAP elevated temperatures in the channel that program, suggesting that they are estimating can have important impacts on many conditions in GRKO with mixed precision. aquatic species (e.g., bull trout, see Table The droughtiness index uses instantaneous 7). However, total riparian vegetation cover discharge which can vary naturally over (versus just on the banks and in the canopy) short time periods. was more comparable to the ecoregion reference mean (124 vs. 136% (SD=11); Riparian Vegetation Cover Table 10). The banks of the Clark Fork do The importance of riparian vegetation tend to have high ground cover of forbs to channel structure, cover, shading, and grasses (perhaps due to the irrigation nutrient inputs, large woody debris, in nearby hayfields). This is confirmed with wildlife corridors, and as a buffer against the third metric in this group, total riparian anthropogenic disturbance is well woody cover. An assessment point does exist recognized (Naiman et al. 1988, Gregory for this metric from Stoddard et al. (2005). et al. 1991). Riparian vegetation not only Woody riparian cover on the Clark Fork moderates stream temperatures through (41%) was intermediate between reference shading, but also increases bank stability (67%) and non-reference (23%) assessment and the potential for inputs of coarse and point values (Table 10). There is little upper fine particulate organic material. Organic canopy (see first metric) and woody cover in inputs from riparian vegetation become food mid and ground strata is patchy, some areas for stream organisms and provide structure have dense thickets of willows and birch, but that creates and maintains complex channel other areas are devoid of nearly all woody habitat. vegetation. Woody cover is an important potential source of nutrients and structure Largely following Stoddard et al. (2005) and and this is another habitat attribute that will Ringold et al. (2008) we used three metrics need to be carefully tracked over time. describing riparian vegetation on the banks of the Clark Fork at GRKO. These include: Signal to Noise Ratio (1) canopy density at the bank, (2) total The S:N of the riparian vegetation metrics riparian vegetation areal cover in the canopy, was generally acceptable within the EMAP mid-level and ground layers, and (3) total program, suggesting that they are estimating riparian woody vegetation areal cover in conditions in GRKO with acceptable the canopy, mid-level and ground layers. precision. In particular, the instrument aided We use canopy density at the bank (versus canopy density metric was very precise midstream as in Stoddard et al. 2005) as this (S:N=34.8). is more appropriate for larger rivers like the Clark Fork where even reference systems Human Disturbance will often not have a closed canopy over the Following Stoddard et al. (2005), Ringold et middle of the stream (Vannote et al. 1980). al. (2008), and Kaufmann et al. (2008), we There are no established assessment points used eight metrics describing disturbances or regulatory criteria available for the first and alteration to riparian vegetation on the

Results and Discussion 59 banks of the Clark Fork at GRKO. These increases in riverside human activities, and include: (1) proximity weighted tallies of increases with increasing riparian woody riparian and near-stream agriculture (of all vegetation complexity and riparian cover types), (2) rowcrop agriculture, (3) roads, density (low values of the index correspond (4) walls/dikes/revetments, (5) an inverse to a higher risk of disturbance at sites given index of human disturbances of all types higher anthropogenic presence coupled combined, (6) a riparian vegetation alteration with low vegetation complexity that might metric, (7) frequency of select target invasive mitigate this disturbance). This index is species, and (8) all target invasive plants in relevant in riparian settings where a multi- riparian plots. storied riparian corridor (including woody vegetation) and a bankside canopy is the Anthropogenic disturbances in riparian expectation. While the ranching mission of systems have long been considered GRKO complicates this somewhat, we feel it important stressors on stream habitat is an appropriate metric for the Clark Fork condition (i.e., Patten 1998). Important at GRKO. In contrast, this metric would not potential anthropogenic stressors of the be appropriate for a prairie stream lacking riparian ecosystem at GRKO include changes naturally occurring woody vegetation on its in hydrologic regimes, livestock grazing floodplain. (upstream and potentially in the ranch), and the high concentrations of heavy metals Finally, we also include the presence of in channel and stream bank sediments. invasive plants as both indirect indicators of Naiman and Decamps (1997) provide an human disturbance and measures of direct overview of the potential mechanisms disturbance. Invasive plants are (usually non- behind these simple proxies and stream native) species whose introduction causes or riparian condition. The SEI protocol or is likely to cause economic or ecological does not directly measure the impact of harm or harm to human health (Clinton anthropogenic disturbances on streams. 1999). Invasive taxa in riparian systems are Rather, indirect proxies are developed especially important as the connectivity based on the spatial proximity of types of afforded by the stream network provides an anthropogenic landuse. Metrics are created additional pathway for invasion of riparian from these data as weighted averages across vegetation to a broader landscape (Ringold all riparian plots, with disturbances closer et al. 2008, Merritt and Wohl 2002). SEI to the Clark Fork assumed to be more data on invasive taxa are very simple and important (metrics range from 0 for no should not replace other efforts to catalog disturbance to a theoretical maximum of 22 (or manage) this issue. Future work will if a disturbance was in the channel at every compare SEI data with more rigorous data station). The index of human disturbances sets. of all types combines a proximity weighted average of the complete list of disturbance There are no established assessment points cataloged in the SEI protocol (channel or regulatory criteria available for the four revetment, pipes, channel straightening, single disturbance metrics or the all types bridges, culverts, buildings, lawns, roads, inverse index, and we only make qualitative pastures, orchards, row crops, and comparisons to mean values across the miscellaneous “trash”—e.g., car bodies, ecoregion (using EMAP data). The Clark grocery carts, pavement blocks, and other Fork at GRKO had much higher values trash) and transforms the value to vary (or these disturbance types were more inversely with disturbance (0 at a maximum prevalent)) for the agriculture and road disturbance, 1 with no disturbance). metrics than the means across ecoregion reference sites (Table 10). Similarly, the index We also develop a riparian vegetation of all disturbances was lower (indicating alteration metric that combines extant higher disturbance) at GRKO than the vegetation cover with the proximity- ecoregion reference mean (0.18 at GRKO weighted inverse tally of riverside human vs. 0.89 (SD=0.03) at ecoregion reference activities. The index value decreases with sites; Table 10). There were no revetments

60 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site (rip rap) in the SEI sample reach at GRKO 25th percentile reference assessment point. in 2009, and on average ecoregion reference The combination of potential disturbance sites do have some revetment. If revetment from close proximity anthropogenic land is added as part of the remediation process use and the lack of a well-developed woody this will of course change, but well placed riparian canopy places GRKO well outside and managed bank improvements can have the reference condition seen at ecoregion minimal negative impacts on ecological reference sites. condition. Disturbance metrics like those presented The riparian vegetation disturbance metric above are part of how EPA and MT DEQ result was also not a surprise; there is a lot place stream sites into reference and non- of disturbance and not a lot of woody or reference classes. Based on these scores canopy riparian vegetation at GRKO that alone, GRKO would clearly fall into a non- might mitigate some of these disturbances. reference class. However, the interpretation There are no established assessment points of this in the context of the condition of the or regulatory criteria available for this metric. Clark Fork at GRKO is complicated. It is possible that the best management practices Given the useful way this metric combines utilized by the park mitigate riparian disturbance and factors that might mitigate disturbance from being a source of local stress, we compared riparian vegetation disturbance and reduced condition in the alteration at GRKO to assessment points channel. from the distribution of reference sites th from across the ecoregion (25 percentile Invasive Riparian Plants th for reference=0.07; 5 percentile non- Four of fifteen targeted invasive taxa reference = 0.04; note that decreasing were present at sample transects at values are associated with a declining GRKO: Canada thistle (Cirsium arvense), condition; a reference state is above the Leafy spurge (Euphorbia esula), Spotted th 25 percentile and a non-reference is knapweed (Centaurea maculosa), and th below the 5 percentile; Table 10). The Dalmatian toadflax Linaria( dalmatica) vegetation disturbance metric at GRKO with frequencies ranging from 0.1 to 0.73 (0.002 in 2009) was in the 0 percentile of (Table 11). While SEI data does not quantify the reference distribution, well below the abundance, many patches appeared to

Table 11. Frequency of 15 target invasive taxa at GRKO during 2009 sample event. Values range from 0 to 1, with 1 describing a species that occurred at all 11 riparian plots. Taxa 2009 Spotted knapweed (Centaurea maculosa) 0.1 Oxeye daisy (Chrysanthemum leucanthemum) 0 Canada thistle (Cirsium arvense) 0.63 Houndstongue (Cynoglossum officinale) 0 Leafy spurge (Euphorbia esula) 0.73 St. Johnswort (Hypericum perforatum) 0 Dalmatian toadflax Linaria( dalmatica) 0.27 Yellow toadflax Linaria( vulgaris) 0 Common tansy (Tanacetum vulgare) 0 Orange hawkweed (Hieracium aurantiacum) 0 Meadow hawkweed Complex (Hieracium pratense, H. floribundum, H. piloselloides) 0 Purple loosestrife (Lythrum salicaria, L. virgatum, and any hybrid crosses) 0 Tall buttercup (Ranunculus acris) 0 Tansy ragwort (Senecio jacobea) 0 Japanese knotweed (Polygonium cuspidatum) 0 Index for all taxa (ip_score) 1.73

Results and Discussion 61 taxa included in this metric differ (as they do across SEI and EMAP). Moreover, one of the most significant limitations of SEI invasive plant data set is that it focuses on a

NPS/BILLY SCHWEIGER NPS/BILLY select set of species rather than on the full assemblage. The result is that the focus of the analysis is best placed on the status of a species rather than on the overall invasion status of the riparian community at a reach. We will therefore use this combined metric only in comparisons across time (and only over a period in which the target species list remains the same). Species-level comparisons will still be possible over time and with external data across the Middle Rockies ecoregion (below).

For comparisons across the ecoregion, we focus on Canada thistle and leafy spurge as Figure 14. Leafy be fairly large (Figure 14). Note that SEI these taxa were also included at EMAP sites. spurge (Euphorbia invasive plant data should not be considered Note that EMAP data was collected in 200- esula) on banks of the as rigorous as weed programs that GRKO or 2003 and is therefore almost a decade old Clark Fork at GRKO in compared to the SEI data used in this report. 2007. other collaborators conduct, especially any effort that generates a complete description Canada thistle occurred (i.e., was present of all species in a plot (such that a degree of in at least one plot at site) at about 8% of community invasion could be derived). the reference sites in the ecoregion, with an average frequency (i.e., the mean proportion There are no established assessment points of plots occupied across sites it was present or regulatory criteria available for these in of about 0.31, SD=0.29). Therefore metrics. However, Canada thistle, Leafy the degree of thistle invasion at GRKO spurge, Spotted knapweed and Dalmatian (frequency of 0.63) is higher than the typical toadflax are all Priority 2b on the Montana reference site. Interestingly, leafy spurge did Noxious weed list (MT DOA 2012). These not occur at any ecoregion reference site weeds are abundant in Montana and (it was at about 5% of all EMAP sites). At widespread in many counties. Management GRKO spurge was almost ubiquitous, with a criteria for these taxa requires eradication frequency of 0.73. Note that the EMAP data (or containment where less abundant) and is almost a decade old compared to the SEI shall be prioritized by local weed districts data and its possible that spurge had yet to (or parks). Montana Fish, Wildlife and Parks fully invade the Middle Rockies ecoregion (MT FWP 2008) states that all of these when the EMAP data was collected. In taxa may render land unfit or greatly limit general, our invasive plant results mirror beneficial uses. other metrics of disturbance in the GRKO riparian corridor and are to be expected We also use a metric of the summed given the agricultural context of the park. frequency of invasive taxa we searched for along the GRKO sample reach. Metric Signal to Noise Ratio values range from a maximum of 15 if all The S:N of the human disturbance metrics taxa occur in all pairs of riparian plots to was generally acceptable within the EMAP 0 if no taxa were found. At GRKO in 2009 program, suggesting that they are estimating this metrics was 1.72 (Table 10 and Table conditions in GRKO with acceptable 11). There are no established assessment precision. The agriculture proximity metric points or regulatory criteria available for was a bit low (but in GRKO this should be this metric. Comparisons to other datasets unequivocal, with fields often on the bank of are also complicated if lists of the target the river). The vegetation disturbance metric

62 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site was very precise measure, with a S:N ratio two year flood recurrence interval and of 16.6. S:N was not calculated for invasive total annual discharge values greater than plant metrics. Ringold et al. (2008) describe the average across the previous 30 years. the various issues with the quality of these Seasonal patterns were fairly typical for data, as we note above, they should not mountain valley streams with a discharge replace more focused efforts. peak in June to July driven by snowmelt and early summer rains. Streamflow Hydrology: Stream Flow decreased rapidly through July and was The hydrologic regime, or the amount and lowest in August and September due to timing of stream flow, is one of the most low precipitation rates and high rates of important aspects of stream habitat and a key evapotranspiration. During summer months, long-term monitoring response within the water also is diverted from the main channel SEI protocol. We used data available from for irrigation. There were higher spring and the nearby USGS gauge in Deer Lodge to early summer flows in all three years and develop detailed hydrographs for the 2007 to higher base flow in 2009 and 2010 versus the 2010 water years and present trend models 30 year average. This is a bit of a departure of discharge over a 30 year period of record from the long term trend (when discharge (in the trend analysis section above). We also was adjusted for seasonal patterns and conducted a flood frequency analysis of the tested for trend over 1978-2011 there was a period of record (note we use 1978 to 2013) marginal (p=0.13) negative trend of -1.72 cfs/ from the Deer Lodge gauge to determine the year, see Table 8). magnitude and recurrence interval of annual expected flood peaks and better place SEI Biology monitoring in context. Because the State of Summary Montana does not have numeric streamflow In general, the macroinvertebrate and criteria, our interpretation of the hydrologic diatom communities at GRKO were in an regime during SEI monitoring is largely intermediate to reference state in 2008 qualitative. and 2009, suggesting that the river is Hydrographs from the USGS gauge for the beginning to or has recovered in part from 2007-2010 water years are shown in Figure the degradation caused by the mine tailing 15. These plots show daily discharge and its floods. Table 14 presents a simple summary mean for a 30 year period of record (1978- with more detailed results below. 2011). They also include the median total Most synthetic metrics indicated a reference discharge for the water year and the historic conditions as they were above MT DEQ period of record. Table 12 presents the total, criteria and (for macroinvertebrates) minimum and maximum flow along with the ecoregion reference assessment points. date of maximum flow for these same time There was some indication of sediment periods. Table 13 shows the results from the issues and while the macroinvertebrate HEC-SSP model of flood frequency for the metal metric suggested an improved Deer Lodge gauge. Atkins (2010) shows that biological response to metals there is still the effect of Cottonwood Creek (in between some minor signal of the historic impact the gauge and GRKO) on flood frequency is of metals in the benthos community. There negligible, with values in GRKO adjusted for was a fairly strong nutrient signal in the peak flow in Cottonwood Creek within the benthos community and, depending on confidence interval of the predicted flows the year, in diatom taxa that have increased for all return intervals at the gauge in Deer abundances in response to elevated nutrient Lodge. concentrations. The biological response seen Qualitative comparisons of discharge in at GRKO generally matches our assessment the 2008 to 2010 water years to the period of SEI (and USGS) chemistry and habitat of record suggest SEI monitoring to date data and our results are similar to the general has occurred in somewhat wet years, with conclusions of McGuire (2010). While we all three annual peak discharges above the lack data to test for trends McGuire (2010)

Results and Discussion 63 Figure 15. Hydrographs for 2007-2010 water years at the Clark Fork River Deer Lodge USGS gauge (station 12324200). Figures include current water years (heavy blue or red lines), mean daily values for 1980-2010 (faint green lines) and the midpoint of current and historical discharge. Note different scale for 2007-2008 water year. Additional statistics are given in Table 12. Graphics were produced in the Climate Data Summarizer, version 3.2 courtesy of Mike Tercek.

showed that 2009 was the only year since Biological Metrics and monitoring began as part of the Superfund Assessment Points project (in 1986) that significant impairment The number and quality of biological was not indicated from macroinvertebrate metrics and their associated assessment communities in most monitored locations in points have been rapidly increasing over the Clark Fork River. the last decade nationwide and in particular in Montana through work by the State of Montana and its partners. We chose

64 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Table 12. Select stream flow statistics from the Deer Lodge USGS station on the Clark Fork River near GRKO. All discharge results are in cubic feet per second (cfs). Date of Standard Total Minimum Peak Median Water Year Maximum Deviation in Discharge Discharge Discharge Discharge Discharge Discharge 2008 93,329 64 957 June 5 206 164 2009 118,014 80 1140 June 1 264 208 2010 112,922 120 1470 June 17 245 202 Historic 88,538 61 2390 May 28 226 140 (1980-2010)

Table 13. Flood frequency analysis for the extended Deer Lodge USGS gauge period of record (1978 to 2013). Two year and ten year return intervals are highlighted in red. All discharge results are in cubic feet per second (cfs). Percent Chance Confidence Limit Confidence Limit Computed Flow Return Period (yrs.) Exceedance (0.05) (0.95) 4913.40 0.20 500.00 7946.10 3521.60 4135.30 0.50 200.00 6427.80 3038.70 3582.00 1.00 100.00 5390.20 2685.60 3057.20 2.00 50.00 4442.70 2341.70 2403.90 5.00 20.00 3320.40 1897.60 1935.90 10.00 10.00 2563.00 1565.20 1484.10 20.00 5.00 1877.20 1227.60 882.60 50.00 2.00 1057.60 737.20 517.20 80.00 1.25 625.00 409.30 388.80 90.00 1.11 481.60 293.00 306.30 95.00 1.05 389.60 220.10 194.20 99.00 1.01 262.40 126.40 biological metrics largely following MT stress biological condition in our informal DEQ guidance (Appendix B and Schweiger assessment of the status and trend in the et al. (In Review) provide an overview of condition of the Clark Fork at GRKO. the biological metrics presented here). Assessment points are primarily those Quality Assurance and Quality Control developed by MT DEQ but we do include Signal to noise values are available from some from other agencies or that are EMAP for a few macroinvertebrate metrics, found in the literature. Finally, we derive but not for diatom metrics. Therefore, ecoregion assessment points for select QAQC for benthos includes both field and macroinvertebrate metrics (we will do this lab procedures and our qualified application for diatom metrics in future reports). As of the S:N results from EMAP, while for noted above, when using MT DEQ metrics diatoms we must rely only on field and lab and associated assessment points NPS procedures alone. We report EMAP S:N cannot formally interpret data in the context for these benthos metrics and discuss them of meeting designated uses. in relevant sections below. Appendix E presents mode details on ROMN QAQC Using biological data is only one component procedures. of water quality assessment as practiced by MT DEQ (MT DEQ 2012a, b, d). Depending Macroinvertebrates on the availability and rigor of other data We present results for macroinvertebrates (i.e., chemistry), biological response cannot in Table 15. Our interpretation of biological be used alone by MT DEQ for beneficial condition at GRKO emphasizes the Montana use determination. However, because NPS Low Valley Multimetric Index (MMI; Jessup is not a regulatory agency, we are freer to et al. 2006) and two RIVPACS metrics

Results and Discussion 65 Table 14. Summary condition table excerpt for select biological metrics at GRKO in 2008 and 2009. We include example vital signs and assessment points, a brief description of results and patterns, and symbolize the status, trend, and our confidence in those summaries. See the Executive Summary for the complete Summary Condition Table. Vital Sign Summary Symbol (Example Indicators) Although there are several stressors impacting the Clark Fork at GRKO, they may not be having a strong effect on the overall integrity of the macroinvertebrate community as most synthetic metrics were in a reference state relative to MT DEQ Biological communities, and reference assessment points from across the ecoregion. Some nutrient metrics macroinvertebrates may suggest a small biological response to increased nutrients. We lack data to (MMI and RIVPACS metrics) assess trends, although longer-term monitoring by USGS and MT DEQ suggests there have been marginal improvements in macroinvertebrate communities in the Clark Fork. We have medium confidence in our assessment of benthos communities at GRKO. Like macroinvertebrates, most diatom metrics were above MT DEQ criteria (or in a reference state). Diatoms are the base of the food chain and the intact diatom community at GRKO may be one of the reasons why we also see fairly high- Biological communities, quality macroinvertebrate assemblages. There were subtle suggestions in some diatoms sediment metrics that the high level of fine sediments in the channel might be (increaser metrics, MMI) impacting diatom communities. Likewise, nutrient metrics were variable with some suggestion of issues, especially in 2008. We lack data to assess trends in diatom communities. We have medium confidence in our assessment of diatom communities at GRKO.

No aquatic invasive species were found, although Didymosphenia geminata (rock Aquatic invasives snot) and New Zealand mud snails (Potamopyrgus antipodarum) are in the Clark (presence) Fork watershed and spreading. SEI monitoring will watch closely for these and other invasive species over the coming years. We have medium confidence in our assessment of aquatic invasives at GRKO.

(O:E and Bray Curtis (BC); Hawkins 2005, conclusion that there were relatively Van Sickle 2008). These models were used high levels of ecological integrity in the by the State of Montana through 2011. macroinvertebrate community during these We also include a suite of classic and sample events, even in comparison to the disturbance-specific metrics that we feel generally higher quality Middle Rockies have interpretative value and/or connect ecoregion sites. SEI results to previous monitoring at or near GRKO and across the state. Finally, Specifically, the values for the MMI were we include interpretation of select taxa above the MT DEQ criterion of 48 in both using demonstrated associations between years. This indicates that through 2011 assemblage components, habitat and water MT DEQ would have likely considered quality variables from the literature and the life histories, tolerances to stressors professional judgment. We feel narratives and community composition attributes as of the presence/absence or abundance summarized in the MMI characteristic of of specific invertebrate taxa lend a more a reference condition macroinvertebrate intuitive nature to SEI results (even though assemblage. Similarly, the ratio of observed many of the species will not be familiar to to expected taxa (O:E) from the RIVPACS the casual reader). model was above the MT DEQ criterion (0.8) for both sample events. This indicates Current MT DEQ Metrics that a high enough proportion of the species In all cases, the MMI and RIVPACS observed at GRKO were characteristic metrics and criteria as used by MT DEQ of that expected from a reference stream (through 2011) suggest that during base in Montana. O:E values in the Clark flow in 2008 and 2009 the Clark Fork at Fork during these sample events actually GRKO was in reference condition. Similar exceeded 1.0, suggesting that the observed patterns were seen when using the (slightly taxonomic richness of reference taxa was more conservative) ecoregion assessment greater than expected for a reference stream. points. This supports a weight of evidence

66 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site x are not are ------2.6 3.05 2.22 2.46 1.58 2.11 1.67 EMAP Signal:Noise 2a 2a 2b ) ) ) 7 3 4 6 5 ------>4 <3 <25 <16 >8.9 are in reference. Metric values in Metric values are in reference. <47.3 <0.83 >0.36 ( ( ( Non-reference Non-reference Assessment Points Assessment bold green 4 3 7 ------5-8 75-25 Point(s) 7-6, 5-46 0 version of this metric (see text) and these thresholds the p > 0.0 version of this metric (see text) and these thresholds 44-38, 36-28, 26-18 Intermediate Assessment Intermediate values are associated with a declining condition at the >25th / <5th percentile values associated with a declining condition at the >25th / <5th percentile values are by MT DEQ; values less than 4 are in reference (or if assessed by MT DEQ, indicate in reference by MT DEQ; values less than 4 are does not use 2a ) 2a ) decreasing 2b value is associated with a declining condition and thresholds in ecoregion reference site reference in ecoregion thresholds value is associated with a declining condition and ) 3 4 6 5 7 >1 rence assessment point). rence assessment ( ------>60.1 1 <3 >8 <4 ( >75 >46 1 <0.30 ( Reference Reference >0.8 are in non-reference. Metric values in Metric are in non-reference. >48 increasing Assessment Points Assessment Grant-Kohrs Ranch National Historic Site, 2008-2009. Assessment points are derived from are derived points Assessment Site, 2008-2009. Historic National Ranch Grant-Kohrs . Note that MT DEQ (“Mountains”). bold red 1 3 56 36 1.3 0.7 58.7 0.20 4.47 0.60 5.03 5.07 16.5 27.1 15.4 2009 15.91 specific to the Clark Fork and may be used by MT DEQ. by MT DEQ; values above 8 are fine sediment intolerant (we use this as reference), 7-6 are moderately intolerant to fine sediment, 5-4 7-6 are reference), fine sediment intolerant (we use this as by MT DEQ; values above 8 are condition classes with the direction of the inequality indicating whether the state is above or below a value. Select or below a value. the state is above indicating whether of the inequality classes with the direction condition 1 3 65 61 30 1.1 0.1 9.8 by MT DEQ; classes are described by (Karr 1998) as Excellent (we use this as reference), with intermediate classes Good, Fair, Poor, and finally Very finally and Poor, with intermediate classes Good, Fair, (we use this as reference), described by (Karr 1998) as Excellent by MT DEQ; classes are 0.22 4.21 0.64 4.83 4.21 35.1 37.8 2008 15.85 are not used are 0 are in reference (or if assessed by MT DEQ, “full in reference and Sada de Suplee 2011); values less than 3.0 are in support of other nutrient data by MT DEQ (Suplee has never been used non-reference or can be used These thresholds . These thresholds by MT DEQ prior to the current MMI model; (Bukantis 1998, MT DEQ 2005); values above or equal to 75 were in reference (or if assessed by MT DEQ, were in “Full (or if assessed by MT DEQ, were in reference MT DEQ 2005); values above or equal to 75 were MMI model; (Bukantis 1998, by MT DEQ prior to the current has never been used reference was used 8 Summary of core macroinvertebrate metrics for the Clark Fork River in Clark Fork for the metrics macroinvertebrate of core Summary by MT DEQ. Fine Sediment Biotic Index Index Temperature Low Valley Multimetric Index (MMI) Multimetric Low Valley O:E (P>0.5) RIVPACS O:E (P>0.0) RIVPACS Bray Curtis dissimilarity (P>0.5) RIVPACS Percent Crustacea and Mollusca Percent Metal Tolerance Index Metal Tolerance EPT Percent Chironomidae Percent RIVPACS Bray Curtis dissimilarity (P>0.0) RIVPACS Valley/Foothill Prairies Multimetric Index Integrity Karr Benthic Index of Biotic Biotic Index (Nutrients) Hilsenhoff Shredder Taxa Richness Taxa Shredder Predator Percent are either intermediate or non-reference (depending on the existence of a non-refe on the existence (depending intermediate or non-reference are either 5) ecoregion thresholds generated from reference sites in the Middle Rockies ecoregion; sites in the Middle Rockies ecoregion; reference generated from thresholds MMI and O:E (p>0.5) ecoregion Bray Curtis ecoregion thresholds generated from reference sites in the Middle Rockies ecoregion; sites in the Middle Rockies ecoregion; reference generated from thresholds Bray Curtis ecoregion Metric Current MT DEQ Metrics Current MMI Component Metrics Classic or Disturbance Specific Metrics Classic or Disturbance Hilsenhoff Biotic Index (Hilsenhoff 1988) Biotic Index (Hilsenhoff Hilsenhoff Fine Sediment Biotic Index (Relyea et al. 2000) 1987, 1989; Ingman and Kerr 1989) is Index (McGuire Metal Tolerance Karr Benthic Index of Biotic Integrity 5) metrics and their criteria were used by MT DEQ through 2011 (MT DEQ 2012d); values above or equal to a criterion are in reference (or if assessed by MT DEQ, “Not Impaired” and (or if assessed by MT DEQ, “Not Impaired” in reference (MT DEQ 2012d); values above or equal to a criterion are 2011 used by MT DEQ through MMI and O:E (p>0.5) metrics and their criteria were The current Low Valley MMI % EPT excludes Hydropsychidae and Baetidae. MMI % EPT excludes Hydropsychidae Low Valley The current Valley/Foothill Prairies MMI Valley/Foothill Notes 5 life”). (or if assessed by MT DEQ, “impairment of aquatic in non-reference than 4.0 are support of aquatic life”), values greater 6 . moderately tolerant to fine sediment and <3 fine sediment tolerant (we use this as non-reference) are 7 . (or if assessed by MT DEQ indicate “metals tolerance”) in non-reference than 8.9 are intermediate, values greater “metals intolerance”); values between 5 and 8 are 8 (see text and Kaufmann et season and other sources signal or true information than variation due to crews, x S:N is the ratio of information to noise in a metric. Higher S:N values suggest the metric had more bioregion EMAP sites in a broader from data are et al. 2005). S:N source al. 1999, Stoddard for reference/non-reference, respectively for reference/non-reference, 2b respectively values for reference/non-reference, at the <75th / >95th percentile distribution are used 3 in non- less than 25 are violated”); values in “Partial support–moderate impairment–standards intermediate (or if assessed by MT DEQ, were 75 are not violated,”); values between 25 and support–standards violated”). impairment-standards (or if assessed by MT DEQ indicate “Non-support–severe reference 4 . Poor (we used this as non-reference) 1 designated use(s)). and do not support (or if assessed by MT DEQ, “Impaired” in non-reference are support designated use(s)), values less than a criterion 2a bold Table 15. Table and are for various sources in Metric values are given within parentheses. assessment points ecoregion

Results and Discussion 67 The Bray Curtis (BC) dissimilarity metric GRKO BC scores were below (or in a index has some advantages over O:E and MT reference condition) the ecoregion reference DEQ recommends using both indices (MT assessment point. This supports the general DEQ 2012d). The BC index can facilitate conclusion from the MMI and O:E metrics determining if expected reference taxa are that the community of macroinvertebrates at being replaced by more tolerant taxa (i.e., GRKO was higher quality. indicative of a non-reference state) in sites transitional between reference and non- Signal to Noise Ratio reference conditions. The BC index is most The S:N of both the MMI and O:E metrics useful when a high (i.e., good) O:E score is was acceptable within the EMAP program, generated from a known or suspected non- with relatively high values suggesting that reference stream (the Clark Fork may be an these metrics may be estimating conditions possible an example of this situation). O:E in GRKO with acceptable precision. can be relatively insensitive to stress-induced shifts in taxonomic composition that have Classic and Disturbance Specific Metrics little net effect on the number of reference- Although the classic and disturbance specific site taxa (Hawkins et al. 2000, Davies and macroinvertebrate metrics are not currently Jackson 2006). used by MT DEQ for regulatory decisions and the current MMI out-performed a Patterns in BC values at GRKO largely suite of classic indices (Jessup et al. 2006) agreed with O:E results when (following we feel they still have value in assisting in MT DEQ guidance) the taxa included were the general interpretation of the condition restricted to those that were more common of the macroinvertebrate assemblage at (with a probability of capture greater than GRKO. However, the following results and 0.5), with little indication of dissimilarity interpretation should be used with caution. between a reference assemblage and what was actually observed. However, if rarer The Valley/Foothills MMI (Bollman 1998, taxa (with a probability of capture less than 2006) in both 2008 and 2009 was near the 0.5) were included, the BC value increased, upper end (or closer to a reference state) suggesting that (rare) species indicative of of the intermediate class. This is similar a reference state were less likely to occur to the current MMI. If interpreted by MT at GRKO. The O:E metric, when also DEQ, this would suggest partial support of calculated using rare species, increased— designated uses. The Valley/Foothills MMI but this suggests the opposite pattern as was developed for lower-order streams the BC metric or that more reference taxa and may be a useful index for sites where than expected were actually observed. The sediment is an important stressor like the choice to include rare species or not in Clark Fork (see SEI substrate data; Table 10). bioassessment is a much debated subject, The Karr Benthic Index of Biotic Integrity with no clear direction as to what is the (B-IBI) was within the “fair” class in both right approach (Cao et al. 1998, Clarke and years. This suggests a slightly more impaired Murphy 2006, Van Sickle et al. 2007, but see condition in the Clark Fork than the MMIs Ostermiller and Hawkins 2004). Some taxa but as this metric was originally calibrated that are more sensitive to stress (or occur in for the Pacific Northwest it may not be reference sites) are naturally less widespread well suited for the specific disturbance and thus tend to have considerably lower regime at GRKO. Nevertheless, it has been average site-specific expected probabilities. successfully used across the West and we The use of higher probability of capture include it here for comparative purposes. thresholds (i.e., greater than 0.5) can exclude Again, we take its general agreement with these more sensitive taxa and may lead to the MMIs as confirmation that there are a lack of sensitivity in the estimation of not any overwhelming signs of stress on the condition. More data and analysis will be macroinvertebrate community in GRKO. needed to resolve this potentially complex result. However, while there are no State The Hilsenhoff Biotic Index (HBI; of Montana criteria for the BC index, Hilsenhoff 1987) is an index of stream

68 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site nutrient concentration based on similar to diatom metrics for sediment tolerance values of a large number of (see below). Fine sediments can limit macroinvertebrate taxa to organic pollution access for some taxa to key habitat and (Barbour et al. 1999). HBI is calculated as can have a dramatic impact on the overall a weighted average tolerance value of all macroinvertebrate community and stream individuals in a sample. Higher index values food web. This result will need close scrutiny indicate increasing tolerance to nutrient over the coming years. concentration. It has a sufficiently predictive response to nutrient gradients in Montana’s The Temperature Index results are close to mountainous streams that MT DEQ uses the mean August temperature at GRKO of it as a secondary response variable to help 16.3ºC from SEI and USGS temperature assess nutrient impacts (MT DEQ 2011a). At data (see Table 7 and Figure 10). There GRKO, the HBI was above its non-reference are no assessment points for this metric assessment point (4.0) suggesting a biological and MT DEQ does not use it in any response to elevated nutrients. SEI (and MT formal way. However, it supports direct DEQ data; MT DEQ 2008a) water chemistry measures of temperature and extrapolates monitoring do indicate that maximum N and this across a broader spatial and temporal P concentrations were often above MT DEQ scale via the preferred temperatures of (2012) nutrient criteria. Nutrient pollution macroinvertebrates. has been a widespread and significant The Metal Tolerance Index (MTI) stressor in the Clark Fork River Basin for metric quantifies changes in community many years. However, impacts attributable composition attributable to metals pollution. to nutrients have declined over the past 16 The format and calculation is similar to years at most monitoring sites on the river the HBI (see above), with tolerance values (McGuire 2010). Interestingly, there appears assigned to each taxon based on sensitivity to to be a complex interaction between metals metals. The scale of the index is 0 to 10 with and nutrients. Often, when the impacts higher values indicating communities more of metal concentrations on the benthic tolerant of metals pollution. McGuire (2010) community are reduced, elevated nutrient suggests that MTI values for communities concentrations are seen. dominated by species intolerant of metals The Fine Sediment Biotic Index (FSBI) are less than 4 while values for communities metric describes the response within composed of only the most metals-tolerant the benthic community to fine sediment species approach 10. Metals tolerance values (Relyea et al. 2000). Lower scores indicate for most taxa were developed from (Ingman a community with more taxa capable of and Kerr 1989) and (McGuire 1987 and persisting in systems with higher levels of 1989). The MTI index at GRKO in 2008 was fines. It was developed for the north western very close to a reference condition (4.21 vs. U.S. from 561 streams, many of which are 4). In 2009 the MTI value was a bit higher similar to the Clark Fork at GRKO and at 5.07. [Note that our 2009 value was very many taxa included in the model occur in similar (5.07 vs 5.05) to the same result at GRKO SEI samples. FSBI results for both the Clark Fork site monitored by McGuire years were in the second to lowest class (2010).] McGuire (2010), shows that these (out of four) indicating that the benthos values are the first time since 1984 (when assemblage consisted of taxa that were monitoring of macroinvertebrates began) moderately tolerant of fine sediment. This that the community response to metals is suggests that there was a biological response near a reference state and suggests that the to the moderate level of fine sediment measured high levels of metals in sediments loading occurring within the Clark Fork at may not still be having a strong impact on the GRKO. This matches SEI habitat data for macroinvertebrate community. This reduced percent fines, sands and embeddedness (see biological response generally matches the SEI substrate data; Table 10). Moreover, results from diatom metal metrics (see the response of the macroinvertebrate below). community to fine sediment is generally

Results and Discussion 69 Component MMI Metrics and MMI metric. Though the ecological Interpretation of Taxonomic Composition mechanisms of this response are not well The following section briefly summarizes understood (Jessup et al. 2006), the percent select patterns in the five component metrics of a sample composed of Chironomidae (Percent EPT, Percent Chironomidae, tends to increase with decreasing stress in Percent Crustacea and Mollusca, Shredder the lower elevation streams of the Middle Taxa Richness and Percent Predator) that Rockies ecoregion (in other ecoregions this comprise the current MMI. We also include relationship is usually opposite of this). The discussion of a few specific taxa of note (i.e., relatively high values of this component Bollman and Bowman 2007). A complete metric are an important part of the higher taxa list with counts and relative abundance scores of the overall MMI. values is given in Appendix D. The Percent EPT (Ephemeroptera, Using a consistent Operational Taxonomic Trichoptera, and Plecoptera) component Unit (OTU) level to define unique taxa, we metric increases with decreasing stress. collected a total of 76 taxa across the two Three taxa of caddis and mayflies sample events. The abundance of organisms included in the EPT component metric was quite surprising (anecdotally, as high as (Tricorythodes, Nectopsyche, and some of the streams sampled by the ROMN Ephemerella inermis) were relatively in Glacier National Park) and we easily common (5 to 16%) at GRKO. These taxa reached the maximum pick size of 600 in are sensitive to disturbance and likely the both samples. In general, the community primary taxa driving the higher scores for the composition was typical for the recent Clark EPT component metric at GRKO (especially Fork River (McGuire 2010), with caddisflies in 2008). Interestingly, net spinning (Trichoptera) and true flies (Diptera), major caddisflies in the family Hydropsychidae components of the assemblage. Mayflies (Figure 16) and the mayfly genusBaetis are (Ephemeroptera), stoneflies (Plecoptera), not included in the Low Valley MMI. These and beetles (Coleoptera) had more restricted two taxa were common in 2008 and 2009, distributions. with relative abundances ranging from 5 to 20%. While most caddisflies and mayflies are The taxon with highest total count and intolerant of disturbance, these two groups relative abundance across both GRKO are generally tolerant (Harris and Lawrence samples was the dipteran midge genus 1978) and many lower condition streams Orthocladius with a mean relative have high abundances of Hydropsyche. abundance around 25% in 2008. In 2009, Had these groups been in the Low Valley two other midge genera (Cricotopus spp. MMI it might not have scored GRKO as at 7% and Polypedilum at 5%) were also a higher quality stream. However, these relatively common. The higher abundances are the only abundant taxa in the GRKO of these taxa were likely a large part of macroinvertebrate assemblage that suggest the relatively high (38 and 27%) values of any sort of disturbance. the Percent Chironomidae component The Percentage of Crustacea and Mollusca at GRKO was very low and consisted of a single taxon in the family Sphaeriidae (a small to minute freshwater bivalve mollusk commonly known as fingernail clams). These snails are primarily collectors, scrapers, and filterers that are tolerant of disturbance.

NOAA PHOTO LIBRARY (NOAA 2012) NOAA PHOTO LIBRARY The low abundance of these taxa also likely drives the higher scores of the overall MMI.

Shredder taxa richness was very low (three Figure 16. Hydropsyche spp. caddisfly, common in both years). Shredder diversity typically in the Clark Fork at increases with increasing stress as they take GRKO.

70 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site advantage of coarse particulate matter that abundances suggest that the SEI site on the can be higher in degraded streams. Finally, Clark Fork in GRKO had about a 42 and 36 the percentage of the GRKO assemblage that percent probability of being impaired due consisted of predaceous taxa was relatively to sediment in 2008 and 2009, respectively. high at 10-15%. Predaceous taxa are more These probabilities did not exceed State common in sites with less stress and more of Montana criterion (51) suggesting intact aquatic insect food webs. that there is not a biological response in diatom communities at GRKO to excess Signal to Noise Ratio sediment. This matches the muted sediment The S:N of the component metrics within signal in macroinvertebrate metrics (Table the MMI were mixed within the EMAP 16). However, it is counter to the more program, with relatively high values for pronounced finer sediment signal seen in Percent EPT and Shredder Richness but physical habitat metrics (Table 10). However, lower values for Percent Chironomidae and as noted, the habitat data presented here are Predators. These likely reflect the difficulty in from a single sample and more data will be identifying these groups and suggest that the needed to resolve this discrepancy. application of the MT DEQ MMI to GRKO should be done with some caution. Mountain Bioregion Nutrient and Metal Periphyton (Diatoms) Increasers (Classic) We present results for diatoms in Table 16. The sediment model presented above is the Our interpretation of biological condition only current tool (as used by MT DEQ) for using diatoms at GRKO emphasizes the diatoms increasers in the Middle Rockies. Middle Rockies ecoregion sediment Teply and Bahls (2005) present similar increaser metric developed by Teply (2010a models for nutrients and metals for the and b, MT DEQ 2011b) and currently used broader Mountain Bioregion that, while by the State. We do present results for recent MT DEQ no longer uses, we feel may be (but not current and less precise) increaser appropriate for our more general needs models (nutrients and metals; Teply and at GRKO to interpret the composition Bahls 2005) followed by a suite of historic and condition of the diatom community. diatom metrics (Bahls 1993) including a Autecological data for the taxa included in MMI and its component metrics. Finally, we these metrics supports their inclusion as include select interpretation of specific taxa increasers in response to these disturbance using demonstrated associations between (van Dam et al. 1994) but the models may assemblage components, habitat and water not describe well the specific patterns quality variables from the literature and and tolerances for the taxa in the Middle professional judgment. These narratives of Rockies. However, the nutrient and the presence/absence and abundance of metal Mountain Bioregion models were specific diatom taxa may lend an intuitive significant (p<0.0001) with jackknifed total nature to SEI results (even though many of classification accuracies of 84% and 80% the species will not be familiar to the casual respectively. reader). The process for interpreting the nutrient and metal increaser metrics is essentially Middle Rockies Ecoregion Sediment the same as with the current sediment Increasers (Current) model. The classic nutrient increaser metric The Middle Rockies sediment increaser indicated that the assemblage in 2008 had model summarize the relative abundance a high abundance of diatoms that tolerate of diatom taxa that, as a group, exist in nutrients, suggesting that there was a high detectable amounts and demonstrate a probability of nutrient issues. In 2009, there meaningful, measurable, and significant were fewer taxa on the nutrient increaser response to sediment. We found seven and list and the probability of impairment ten (2008 and 2009, respectively) sediment decreased, suggesting a high degree of increaser taxa, representing a relative temporal variability in this response. Given abundance of around 13 and 11%. These the difference between 2008 and 2009, the

Results and Discussion 71 ------<1 <1 >60 >75 <1.5 Point Non-reference Assessment Assessment Non-reference ------2-3,1-2 Point(s) 2.5-2, 2-1.5 20-39, 40-60 25-50, 50-75 2.5-1.75, 1.7-1 Intermediate Assessment Intermediate

are either intermediate or non-reference (depending on the existence on the existence (depending intermediate or non-reference are either <20 <25 >=4 >2.5 >2.5 Point <~28 (51) <~29 (51) <15.3 (51) bold Reference Assessment Assessment Reference 4 6 37 4.4 2.61 2009 17, (26) 12, (14) 10.8, (35.8) 4 2 10 3.1 2.34 2008 8, (9) are in reference. Metric values in Metric values are in reference. 65, (99) 13.1, (42.4) bold green condition classes with the direction of the inequality indicating whether the state is above or below a value. Metric values Metric values a value. or below is above the state whether indicating inequality of the direction with the classes condition 2

1 2 by MT DEQ (Bahls 1993); metric values=4 have excellent biological integrity and no overall impairment, metric values between 2 and 3 have good biologi by MT DEQ (Bahls 1993); metric values=4 have “None” impairment, disturbance and “Excellent” biological integrity, number have no stress/siltation/pollution/ by MT DEQ (Bahls 1993); values above the first non-reference 3 or The RA at the 51% level DEQ defined as impairment are both given. are and Bahls, 2005 and 2006). The RA at the 51% level DEQ defined as impairment by MT DEQ (Teply The relative abundance (RA) at the 51% level is defined by DEQ as impairment abundance (RA) at the 51% level is defined. 2010a, b). The relative by MT DEQ (Teply reference were used were historically used were historically used were 4 4 4 4 currently used currently are in non-reference. Metric values in Metric are in non-reference. nts are derived from various from various derived points are Assessment 2008-2009. Site, Historic Ranch National in Grant-Kohrs Fork River for the Clark metrics diatom of core Summary Shannon Diversity Siltation Index Pollution Index Disturbance Index .) (RA, prob Middle Rockies Sediment Increasers, .) (RA, prob Mountains (classic) Nutrient Increasers, .) (RA, prob Mountains (classic) Metal Increasers, Index Montana Diatom Multimetric Assessment points Criteria are Criteria are Assessment points Assessment points bold red impairment. 4 for designated uses; “Minor” impairment and “Partial support” and “Good” biological integrity, range have minor stress/siltation/pollution/disturbance or “Full support” for designated uses; in the second have “Moderate” impairment and partial support of designated uses; and below the fourth number and “Fair” biological integrity, range have moderate stress/siltation/pollution/disturbance in the third for designated uses. and “Non-support” impairment or stress “Severe” and “Poor” biological integrity, stress/siltation/pollution/disturbance severe Notes 1 2 3 overall integrity and severe values between 1 and 2 have fair biological integrity and moderate overall impairment, metric values=1 have poor biological cal integrity and minor overall impairment, metric Classic MMI Component Metrics Classic MMI Component Current Disturbance Specific Metrics Current Disturbance Specific Metrics Classic or Disturbance Metric Table 16. Table for and are sources in points). assessment of a non-reference

72 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site lack of a nutrient signal in SEI chemistry (i.e., more nutrient tolerant species in 2008) and macroinvertebrate data and the broader may not have adversely affected diversity. spatial scale of the modeling behind The siltation and pollution index were also this metric we are somewhat reluctant above their impairment criteria but both to interpret this result as meaningful. did indicate partial support in the 2009 and Nevertheless, given the many sources of 2008 sample events, respectively. Finally, nutrients in and above GRKO it does suggest the disturbance index was well below the future work should continue to carefully impairment criteria, although this metric look at multiple nutrient indicators. is likely better suited to higher elevation steeper gradient streams. MT DEQ (2005) The classic metals increaser metric indicated combined these four metrics into a MMI that the diatom assemblage had relatively synthetic index of stream condition. low abundances of taxa tolerant of metals, Applying this to the GRKO data suggested suggesting that there was not an impact of that diatom communities were intact with metals on the diatom community and that excellent biological integrity and no overall there was a low probability of metal issues. impairment—matching the general pattern Interestingly, this contradicts the metals in nearly all more current and stressor in sediment and water column results. We specific metrics. interpret this as either a spatial disconnect between where diatoms, sediments Interpretation of Taxonomic Composition and water are sampled or the biological We next discuss select periphyton taxa irrelevance of the high levels of metals in within the GRKO samples (e.g., Bollman and the water column and fine sediments as we Bowman 2007, Teply and Bahls 2006). Using also saw no indication of metal issues in the autecology of the dominant species in macroinvertebrate metrics. the samples allows additional interpretation and may help resolve the reasons or causes Comparison to Middle Rockies Ecoregion behind the increaser metrics or the more Assessment Points general indicators of ecological integrity. These comparisons will be done in a future Diatoms ranked first in biovolume for both iteration of this report. sample events at around 51% and coupled with the less well resolved tolerances for Classic and Disturbance Specific Metrics soft-bodied algae in Montana streams we Bahls (1993) developed diatom metrics focus on these taxa. that were used by the State of Montana as a diagnostic tool in support of monitoring and A complete taxa list with counts and relative assessment for nearly a decade. However, abundance values is given in Appendix Teply and Bahls (2005) show that these D. We collected 84 unique taxa (largely metrics have relatively low capacity to identified to species level) over the two discriminate impairment. Therefore, we use sample events at GRKO. Community these results only in a general way to help composition in the two periphyton samples interpret the composition and condition of from GRKO was fairly typical for the recent the diatom community at GRKO. Clark Fork River (L. Bahls, pers. comm., 2011). There were no exceedances of the assessment points for any of these metrics Non-Diatom Algae during either sample event at GRKO. While The periphyton community at GRKO some metrics scores fell within intermediate consisted of diatoms, cyanobacteria, and classes of impairment, most were actually green and red algae. The cyanophytes well within the highest quality range. (blue-green algae) Homoeothrix janthina Shannon’s diversity was well above the and Leptolyngbya spp. were the most reference assessment point of 2.5, suggesting abundant soft-bodied algae ranging in that the diatom community included a large relative abundance from 9 to 20%. Potapova number of species with generally equitable et al. (2005) presents water chemistry distributions. This implies that the species optima for several species of soft-bodied replacement seen in the increaser metrics algae including Homoeothrix janthina. This

Results and Discussion 73 species tends to occur at sites with relatively western Montana and some of the highest good water quality with larger substrates densities of attached algae are found in but moderately high total nitrogen perhaps the reach just downstream from GRKO due to the absence of heterocytes in this (Johnson 2002). Deer Lodge municipal organism, and therefore to its inability to fix wastewater is handled by the Deer Lodge free nitrogen. Other notable taxa include the wastewater treatment facility, located just pollution-sensitive red alga Audouinella spp. inside the northern park boundary (see during the 2008 sample event. However, the Figure 2 and Figure 3). Since 2000, the ranch red alga genus Phormidium, which contains has used secondary treated sewage effluent several pollution-tolerant species, was more from the wastewater treatment facility to abundant and occurred in both 2008 and irrigate some of its hayfields during summer 2009. months. In 2000, the NPS funded a study to determine the fate of sewage effluent Diatoms nutrients (primarily nitrogen) applied to soils The diatom taxon with highest total in GRKO by a sprinkler irrigation system count and relative abundance across both (Johnson 2002). The study determined that GRKO samples was Diatoma moniliformis most of the applied nitrogen was retained (Figure 17). It was common in 2008, with by plants and in the soil but that as much as a relative abundance greater than 50%. 40% of the nitrogen might be leached to the D. moniliformis is a fairly common species shallow ground-water system. These diatom across the West, and it appears to be species and the high levels of the nutrient generally intolerant of general anthropogenic metric in 2008 may suggest that there disturbance when considered across a are nutrient issues in the reach, perhaps broader region (Spaulding et al. 2010). connected to the sewage effluent applied to Gomphonema parvulum was also somewhat GRKO fields, moving from groundwater to common in the GRKO samples. It can the Clark Fork. However, the signal may also indicate organic enrichment, especially be temporally variable and more research is when associated with sedimentation (van needed. Dam et al. 1994). Finally, Cocconeis pediculus also occurred in the samples in higher Reimeria sinuate was also relatively abundant abundances. It is primarily an epiphyte on in both years. This species is on the current Cladophora spp. (a filamentous green algae), Middle Rockies sediment increaser list and which prospers in nutrient-rich waters with is the primary contributor to this metric at slow to moderate current velocities where GRKO. It was the only species on the list sedimentation is an issue. that occurred with more than 5% relative abundance, with almost all other increaser The high abundances of these species taxa at less than 1% relative abundance. (especially D. moniliformis in 2008) are the Similarly, Amphora pediculus was the primary reason why the classic nutrient second most common species across both increaser metric was so high in 2008. The samples. It was on the Teply and Bahls (2005) Clark Fork historically has some of the most sediment increaser list for the mountainous productive (nutrient rich) stream water in part of the state. Nitzschia inconspicua and

N. dissipata were also relatively abundant.

(2010) N. dissipata was on the Teply and Bahls (2005) sediment increaser list but not on POTAPOVA M . POTAPOVA

OBTAINED FROM OBTAINED the current list of increaser taxa used in SPAULDING ET AL . SPAULDING (POTAPOVA 2009); (POTAPOVA the Middle Rockies. In general, these taxa are motile and can deal with mobile sand Figure 17. grains and increased fines. They favor slower Scanning electron current velocities where sediments are prone micrograph image to accumulate. The marginal indications of of the most sediment issues in the diatom community at common algae at GRKO, Diatoma GRKO are likely due to these species. moniliformis.

74 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Aquatic Invasive Species There were no known zebra or quagga mussel (Dreissena polymorpha, and D. rostriformis bugensis) or rusty crayfish (Orconectes rusticus) occurrences in or near the Clark Fork watershed (the mussels have been found outside Rocky Mountain National Park and the crayfish on the western slope of Colorado). Confirmed New Zealand mudsnail populations (Potamopyrgus antipodarum) in south west Montana and neighboring states from the USGS Nonindigenous Aquatic Species program (Benson 2011) are shown in Figure 18 (as of 2011). There are no known current populations in the Clark Fork watershed, but the species does occur in nearby locations in rivers similar to the Clark Fork (and outside Rocky Mountain National Park in Figure 18. Point locations of the New Zealand mudsnail (Potamopyrgus Boulder, CO). Mudsnails tolerate siltation, antipodarum) in the Montana/Idaho/Wyoming area. Red dots indicate thrive in disturbed watersheds, and benefit established populations. Yellow dots indicate confirmed collections. from high nutrient flows. They have the Approximate location of GRKO is shown with the red star. Data are ability to reproduce quickly and mass in high from several (uncoordinated) sample designs and therefore are not a densities and can impact trophic dynamics valid sample of this region as a whole. Data span 2004-2009. Data and map courtesy of the USGS Nonindigenous Aquatic Species program. of native trout fisheries and alter the physical characteristics of the streams themselves. Given these possible impacts and the suitability of the Clark Fork to the snail, we will closely watch for this species.

Didymo (Didymosphenia geminata) have also been confirmed in the Montana region from work conducted by the U.S. EPA (EPA 2012) the ROMN (in GLAC) and other partners (Figure 19). As of 2011 there were no known current populations on the Clark Fork in the GRKO watershed, but the species does occur in immediately adjacent watersheds. Didymo is a diatom native to mountain habitats of North America and Europe (Blanco and Ector 2009). In recent years didymo has expanded into lower elevations, latitudes, and new regions of the globe (Kumar et al. 2009). In Montana, didymo was first reported in 1929 at Flathead Lake (Prescott and Dillard 1979) and has likely been present Figure 19. Point locations of confirmed presence of didymo in the northern Rockies since at least the end (Didymosphenia geminata) in the Montana region. Data are from of the last ice age, about 10,000 years ago several (uncoordinated) sample designs and therefore are not a valid sample of this region as a whole. Data span 2004-2009. Records outside (Bahls 2007). Didymo can form extensive of Glacier National Park are based on data from USGS National Water mats (or blooms), which can be several Quality Assessment (NAWQA), EPA Environmental Monitoring and centimeters thick and up to 20 km (12 mi) Assessment (EMAP), and samples from other studies and are courtesy in length (Blanco and Ector 2009). Larger Karl Hermann, Sarah Spaulding, and Tera Keller. Data in Glacier blooms can inhibit growth of other algal National Park are from ROMN SEI monitoring. species, change the composition of aquatic

Results and Discussion 75 communities, decrease the amount of of the drivers of didymo abundance in suitable spawning habitat for fish, and cause another ROMN park, Glacier NP (see Figure changes in stream chemistry (Spaulding and 19 for these data as presence/absence). As Elwell 2007). Blooms of didymo also greatly with mudsnails, given the possible impacts decrease the aesthetic appeal of streams, and the suitability of the Clark Fork to this an important consideration for the NPS. diatom, the ROMN will closely watch for Schweiger et al. (2011) recently described the this species in GRKO. distribution and developed simple models

76 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Summary and Conclusions

This report presents data, results, and concentrations in water and sediment select interpretations from ROMN SEI (although these have been dropping over monitoring on the Clark Fork River during time), variable and sometimes elevated 2008-2010 at GRKO. We employed several nutrient concentrations with evidence vital signs to monitor the status and trend of an episodic nutrient signal in benthos (as possible) of stream condition. These and diatoms, a suggestion of rising water included multiple measures of water and temperatures, a somewhat elevated level sediment physiochemistry, physical habitat, of fine embedded substrates, and reduced and biology (macroinvertebrate and diatom riparian vegetation cover with some invasive communities). The approach of the SEI plants of concern. However, the majority of protocol is to integrate or interpret these the other measures and metrics, including measures in concert as indicators of the the synthetic biological indices of ecological ecological integrity of the system and to condition, suggest that most aspects of the emphasize individual metrics that are likely Clark Fork in GRKO were in a reference direct estimators of integrity. We interpret or near reference state in 2008-2010. With results using a variety of assessment this combination of results we feel the mechanisms including regulatory criteria overall score (within the constraints of the (but in a non-regulatory way), other relevant Summary Condition process) is a “Caution/ assessment points from the literature and Intermediate.” We do not have enough where possible, reference values derived data to speak on the trend in this overall from state and federal monitoring reference condition, but suspect it is improving. sites in the Middle Rockies ecoregion. We have “Medium” confidence in these Because SEI monitoring is not regulatory in conclusions as discussed further below. nature, we can consider assessment points that may have more of an ecological basis Confidence or are relevant to a specific GRKO resource The SEI protocol is based on established management need. and accepted methods, used by many ROMN partners. Field methods and the Table 17 presents an overall assessment of analyses employed on these data were the condition of the Clark Fork in 2008- chosen because of their general ease of 2010. In general, the river and its riparian implementation and quality of the results. corridor was largely intact and medium to However, budget and other constraints high in quality. The few areas of concern do limit some uses of the data, especially might include residual elevated metal so early in ROMN monitoring when we

Table 17. Summary condition table excerpt for the overall assessment of the condition of the Clark Fork at GRKO in 2008-2010. We include example vital signs and indicators, a brief description of results and patterns, and symbolize the status, trend, and our confidence in those summaries. See the Executive Summary for the complete Summary Condition Table.

Vital Sign Summary Symbol (Example Indicators)

The ecological integrity of the Clark Fork from 2008 to 2010 was largely intact and of medium to higher quality—a somewhat surprising result given the history of intensive land uses in the watershed. However, there are lingering issues with Overall ecological metals in water and sediment from historical mining in the integrity watershed, and select aspects of in channel and riparian habitat are in a non-reference condition. In general the condition of the system may be slowly improving. We have medium confidence in our interpretation.

Summary and Conclusions 77 have a small temporal sample size, and the calculation for energy loss from channel following briefly summarizes important form roughness (Kaufmann et al 2008). caveats. Biology Water Physiochemistry The bioassessment models we use are of Most water physiochemistry constituents course imperfect. As with any modeling are naturally variable seasonally or even effort, there is error and uncertainty daily and SEI survey methods may not associated with data sampling and capture some of this temporal variance. processing, model calibration, validation, Water physiochemistry also has a high and model use. The current MT DEQ Low degree of spatial structure, with meaningful Valley MMI had a discrimination efficiency differences across macrohabitats within of 94%, indicating that the MMI was unable streams. Many of the chemistry parameters to distinguish between reference and we include are likely conservative and may degraded sites in only approximately 6% of integrate across habitats. SEI base flow the samples. We feel this is an acceptable sample events are conducted when most error rate for our application of the tool to biota are stressed (due to lower stream flow) GRKO. The Montana RIVPACS model is or at key points in their life cycles. Water or comparable to or better than most RIVPACS sediment physiochemistry at this point in models in use in the U.S. and elsewhere in the hydrologic cycle may be a useful index of terms of model precision (Hawkins 2006). stress driving biological condition. Good models typically have a standard deviation in O:E less than 0.18. The Montana Finally, while the status of water model SD in O:E was 0.17 with the model physiochemistry is important, given its diel accounted for ~88% of the explainable and seasonal fluctuations, it may not be as variability in taxonomic composition among relevant as long-term, seasonally adjusted samples. Moreover, because the OTUs trends. After we collect sufficient data at the used in RIVPACS modeling often represent SEI sentinel site we will conduct detailed relatively coarsely resolved taxa (e.g., many trend analyses and we will continue to use genera, some families, a few species), our USGS data, where they are available, to assessments will be conservative with analyze for trend. respect to what we would see with models based on species-level data (Hawkins et Habitat al. 2000). An important consideration in The SEI protocol adopts a suite of field bioassessment using RIVPACS is the match methods for habitat developed for regional or fit between an assessed site and the surveys (i.e., EMAP). These methods are appropriate reference condition. The Clark relatively simple and inexpensive. They Fork at GRKO was within the experience have been shown to provide a diverse suite of the RIVPACS model, suggesting that of useful data, with most metrics generated the environmental conditions at GRKO from these data having acceptable precision that drive macroinvertebrate community that allow meaningful comparisons across assemblage were comparable to those seen large numbers of sites within a region. in reference sites. However, these methods are not equivalent to an engineering grade survey for a specific Finally, it is sometimes difficult using stream site. Current SEI habitat results might biological metrics to be diagnostic or to prudently be confined to identifying severe ascribe a causal relationship between a cases of sedimentation, channel alteration, biological response and a stressor. While etc. in the Clark Fork at GRKO. Greater not necessarily needed from a strict long- confidence to discern subtle differences term monitoring perspective focused on could be gained by using more precise ecological integrity, knowing, at least in a field measurements of channel slope, bed correlative sense, what might be causing particle size and bankfull dimensions, and by a lower quality biological status is useful refining some of the adjustments in metric for park management and interpretation

78 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site for visitors. As we accrue more data and et al. 2004), a lower snowpack volume at our understanding and modeling of the lower to mid-elevations (Knowles et al. biological response in the Clark Fork 2006), and increased flooding due to rain- improves, we may be able to make more on-snow events in winter (Heard 2005). definitive statements about why a particular Hydrographs (i.e., the magnitude and timing response is seen at GRKO. Moreover, of spring run-off) will likely shift to earlier methods being developed for the ROMN floods. These changes will, in turn, likely wetland protocol on causal modeling (Grace affect the seasonal dynamics of stream et al. 2013) may be applied to future SEI and riparian biota (Palmer and Bernhardt data to help refine our understanding of 2006). Climate may drive changes in water relationships in these data. quality due to several mechanisms such as stream temperature, increased erosion, and Management Implications decreased dilution of pollutants. Decreases Pending review and input from the park, the in snow cover and more winter rain on bare SEI results at GRKO will be applied toward: soil are likely to lengthen the erosion season, which can increase nutrient concentrations 1. Continued long-term monitoring of the in streams. Predicted increases in the ecological integrity of the Clark Fork severity and frequency of floods may also in GRKO, especially as a response to contribute to increases in erosion, as well as climate change. affect ecological processes that are sensitive to changes in the probability distributions 2. Contributing to interpretative material of high flow events such as habitat stability, for park visitors. biodiversity, and trophic structure (Hamlet 3. Assisting park staff and management in and Lettenmaier 2007). understanding the ongoing restoration Whether climate change is having a of the river and its floodplain as part of measurable impact on the Clark Fork in the Superfund project. GRKO is unclear. In the northern and 4. Regulatory applications as needed (in central Rockies, streamflow has generally collaboration with MT DEQ). shifted toward earlier peak runoff, which has been attributed to more precipitation Climate Change falling as rain rather than snow and earlier One of the most relevant applications of SEI snowmelt (Knowles et al. 2006, Mote et al. monitoring in GRKO will be to helping the 2005). The midpoint of annual discharge park understand any changes in the integrity and the date of highest flows (not driven by of the Clark Fork that might be due to ice jams) at GRKO for 2007-2010 were all climate change. This will of course take more later in the year than during historic period and longer-term data for SEI responses, of record. This suggests a hydrograph that is although the USGS stream flow and water shifting forward in time, with delayed runoff. physiochemistry data from the Deer Lodge There is also a suggestion that the GRKO gauge are already offering some glimpses at hydrograph is becoming more variable. It possible climate mediated shifts in the Clark is unclear if the three recent water years Fork. at GRKO are aberrations from patterns in climate driven streamflow phenology seen Changes in stream hydrology and across the west over longer time periods. temperatures in the West over the past More data will be required to test this. Isaak fifty years due to climate changes are well et al. (2011) show an interesting pattern documented (Barnett et al. 2008) and these across select Pacific Northwest streams changes are expected to continue. Even from 1980 to 2009 (two of which were in modest ambient temperature increases in the Montana) with a cooling trend apparent western U.S. may cause significant changes during the spring and warming trends during to the hydrologic cycle, often manifested in the summer, fall, and winter. The amount earlier snowmelt, earlier ice-out on lakes, of warming more than compensated for reduced summer base flows (Dettinger spring cooling to cause a net temperature

Summary and Conclusions 79 increase, and rates of warming were highest The future of the SEI protocol in GRKO during the summer. Warmer stream water, includes several important actions. Primarily as we see a suggestion of in GRKO, can we must continue to accumulate and analyze have a cascade of impacts on the condition SEI data relative to criteria and assessment of a river like the Clark Fork. Increased points from state and developed from the temperatures can lead to oxygen depletion, ecoregion. These may help us focus on a change in fish distribution, and a loss of monitoring the most parsimonious and some species of benthos and periphyton, management relevant suite of indicators to especially those that are isolated in habitats evaluate the ecological integrity of the Clark near thermal tolerance limits or that occupy Fork in GRKO. We may want to consider rare and vulnerable habitats (Williams et al. improving a subset of the habitat methods. 2007). Fish such as bull trout, dependent on Our current protocol includes relatively cool waters, will likely decline (Pederson et simple and inexpensive approaches shown al. 2010). In contrast, many fish species that to provide a diverse suite of useful data. prefer warmer water, such as largemouth However, these methods are not equivalent bass and carp, may expand their ranges if to an engineering grade survey of a specific surface waters warm (Battin et al. 2007). stream site. More precise field measurements Warmer waters may also cause aquatic of channel slope, cross-section geometry, diseases and parasites to become more and bed surface particle size would be widespread (Hari et al. 2006). required to use some of our metrics in site- specific assessments at individual streams. Future Work Finally, we may want to consider including This effort has been, and will continue to be, episodic monitoring over extended time a cooperative undertaking between ROMN periods of in situ parameters (pH, DO, staff; aquatic scientists from USGS, the EPA, conductivity, temperature) and adding and MT DEQ; and most importantly, GRKO turbidity to quantify shorter term changes management and staff. Our results represent that may occur with the ongoing restoration the efforts of many dedicated scientists and of the floodplain at GRKO. resource managers over several years and we feel are valid, representative and address our objectives. They should be useful for understanding and, within the constraints of the protocol, managing the Clark Fork.

80 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Literature Cited

Allan, J. D. 2004. Landscapes and USA: A critical element of ecological riverscapes: The influence of land use integrity. Hydrobiologia 422:453-464. on stream ecosystems. Annual Review of Ecology Evolution and Systematics Barbour, M. T., J. Gerritsen, B. D. 35:257-284. Snyder, and J. B. Stribling. 1999. Rapid bioassessment protocols American Public Health Association for use in streams and wadeable (APHA). 1995. Standard methods rivers: Periphyton, benthic for the examination of water and macroinvertebrates, and fish. U.S. wastewater. American Public Health Environmental Protection Agency, Association, American Water Works Office of Water, Washington, DC. Association, and Water Pollution Control Federation. 19th edition, Barnett, T. P., D. W. Pierce, H. G. Hidalgo, Washington, D.C. C. Bonfils, B. D. Santer, T. Das, G. Bala, A. W. Wood, T. Nozawa, A. A. Mirin, Arismendi, I., M. Safeeq, S. L. Johnson, D. R. Cayan, and M. D. Dettinger. J. B. Dunham and R. Haggerty. 2008. Human-induced changes in the 2012. Increasing synchrony of hydrology of the western United States. high temperature and low flow in Science 319:1080-1083. western North American streams: Double trouble for coldwater Battin, J., M. W. Wiley, M. H. Ruckelshaus, biota? Hydrobiologia DOI 10.1007/ R. N. Palmer, E. Korb, K. K. Bartz, s10750-012-1327-2. and H. Imaki. 2007. Projected impacts of climate change on salmon habitat Bahls, L. L. 2007. Diatom indicators of restoration. Proceedings of the National climate change in Glacier National Academy of Sciences of the United States Park. Intermountain Journal of Sciences of America 104:6720-6725. 13:99–109. Bennetts, R. E., J. E. Gross, K. Cahill, C. Bahls, L. L. 2004. Northwest diatoms: A McIntyre, B. B. Bingham, A. Hubbard, photographic catalogue of species L. Cameron, and S. L. Carter. 2007. in the Montana diatom collection, Linking monitoring to management with ecological optima, associates, and planning: Assessment points as and distribution records for the nine a generalized approach. The George northwestern United States. Volume 1. Wright Forum 24:59-77. Hannaea, Helena, Montana. Benson, A. J. 2011. New Zealand mudsnail Bahls, L. L. 1993. Periphyton bioassessment sightings distribution. Available at: New protocols for Montana streams. Zealand mud snail distribution.aspx Montana Department of Health and (accessed 18 December 2011). Environmental Sciences. Water quality Bureau. Helena, Montana. Benson, B. J. and J. J. Magnuson. 1992. Spatial heterogeneity of littoral fish Bailey, R. C., R. H. Norris, and T. B. assemblages in lakes: Relation to species Reynoldson. 2004. Bioassessment diversity and habitat structure. Can. J. of freshwater ecosystems using the Fish. Aquat. Sci. 49:1493-1500. reference condition approach. New York: Kluwer Academic Publishers. Bestelmeyer, B. T. 2006. Threshold concepts and their use in rangeland management Barbour, M. T., W. F. Swietlik, S. K. Jackson, and restoration: The good, the bad, D. L. Courtemanch, S. P. Davies, and and the insidious. Restoration Ecology C. O. Yoder. 2000. Measuring the 14:325-329. attainment of biological integrity in the

Literature Cited 81 Biggs, T. W., T. Dunne, and L. A. Martinelli. Bray, J. R. and J. T. Curtis. 1957. An 2004. Natural controls and human ordination of upland forest impacts on stream nutrient communities of southern Wisconsin. concentrations in a deforested region Ecological Monographs 27: 325-349. of the Brazilian Amazon basin. Biogeochemistry 68: 227–257. Brick, C. and J. N. Moore. 1996. Diel variation of aqueous trace metals in Biggs, B. J. F. 1996. Patterns of benthic the upper Clark Fork River, Montana algae in streams. Pages 31-55 in R. Environmental Science and Technology J. Stevenson, M. Bothwell, and R. 30:1953-1960. L. Lowe, editors. Algal ecology: Freshwater benthic ecosystems. Bright, D. A. and J. Addison. 2002. Derivation Academic Press, San Diego, California. of matrix soil standards for salt under the British Columbia Contaminated Blanco, S. and L. Ector. 2009. Distribution, Sites Regulation. Royal Roads ecology, and nuisance effects of University. Prepared for BC Ministry the freshwater invasive diatom of Water, Land and Air Protection, Didymosphenia geminata (Lyngbye) BC Ministry of Transportation and M. Schmidt: A literature review. Nova Highways, BC Buildings Corp., and the Hedwigia 88:347–422. Canadian Association of Petroleum Producers. Victoria, BC. Bollman, W. 2006. Summary of aquatic invertebrate data from the Montana Briske, D. D., S. D. Fuhlendorf, and F. E. statewide monitoring network, Smeins. 2006. A unified framework 2001-2005. Rhithron Associates, Inc. for assessment and application of Missoula, Montana. ecological thresholds. Rangeland Ecology and Management 59: 225-236. Bollman, W. 1998. Improving stream bioassessment methods for the Briske, D. D., B. T. Bestelmeyer, T. K. Montana Valley and Foothill Prairies Stringham, and P. L. Shaver. 2008. Ecoregion. M.S. Thesis. University of Recommendations for development of Montana, Missoula, Montana. resilience-based state-and-transition models. Rangeland Ecology and Bollman, W. and J. Bowman. 2007. Management 61:59-367. An assessment of the ecological conditions of the streams and rivers Britten, M., E. W. Schweiger, B. Frakes, of Montana using the Environmental D. Manier, and D. Pillmore. 2007. Monitoring and Assessment Program Rocky Mountain Network vital signs (EMAP) method. A report prepared monitoring plan. Natural Resource for The Montana Department of Report NPS/ROMN/ NRR-2007/010. Environmental Quality. Rhithron Fort Collins, Colorado. Associates, Inc. Helena, Montana. Bryce, S. A., G. A. Lomnicky, and P. R. Bowman, M. F. and K. M. Somers. 2006. Kaufmann. 2010. Protecting sediment- Evaluating a Novel Test Site Analysis sensitive aquatic species in mountain (TSA) Bioassessment Approach. streams through the application of 2006. Journal of the North American biologically-based criteria streambed Benthological Society 25: 712-727. sediment criteria. Journal of the North American Benthological Society Brandt, D. 2001. Temperature preferences 29(2):657-672. and tolerances for 137 common Idaho macroinvertebrate taxa. Report to the Bryce, S. A., G. A. Lomnicky, P. R. Kaufmann, Idaho Department of Environmental L. S. Mcallister, and T. L. Ernst. Quality, Coeur d’ Alene, Idaho. 2008. Development of biologically- based sediment criteria in mountain

82 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site streams of the western United States. Dodge, K. A., M. I. Hornberger, and J. North American Journal of Fisheries L. Dyke. 2005. Water quality, bed- Management 28:1714-1724. sediment, and biological data (October 2003 through September 2004) and Bukantis, R. 1998. Rapid bioassessment statistical summaries of data for macroinvertebrate protocols: Sampling streams in the upper Clark Fork basin, and sample analysis SOPs. Working Montana. U.S. Geological Survey Draft. Montana Department of Open-File Report 2005–1356, 124 p. Environmental Quality. Planning and Prevention and Assistance Division. Dodge, K. A., M. I. Hornberger, and J. Helena, Montana. L. Dyke. 2009. Water-quality, bed- sediment, and biological data (October Cao, Y., D. D. Williams, and N. E. Williams. 2007 through September 2008) and 1998. How important are rare species in statistical summaries of long-term data community ecology and bioassessment. for streams in the Clark Fork basin. U.S. Limnology & Oceanography Geological Survey Open-File Report 43:1043-1049. 2009–1178, 140 p. Clarke, R. T. and J. F. Murphy. 2006. Effects of locally rare taxa on the precision and Dodds,W. K., V. H. Smith, and B. Zander. sensitivity of RIVPACS bioassessment 1997. Developing nutrient targets to of freshwaters. Freshwater Biology control benthic chlorophyll levels in 51:1924–1940. streams: A case study of the Clark Fork River. Water Resources 31:1738-1750. Clinton, W. J. 1999. Executive Order 13112 establishing the National Invasive Dunham, J., B. Rieman, and G. Chandler. Species Council. Federal Register 64: 2003. Influences of Temperature 6183–6186. and Environmental Variables on the Distribution of Bull Trout within Cummins, K. W. 1974. Structure and Streams at the Southern Margin of function of stream ecosystems. Its Range. North American Journal of Bioscience 24:631-641. Fisheries Management 23:894–904.

Davies, S. P. and S. K. Jackson. 2006. The Fancy, S., J. Gross, and S. Carter. 2009. biological condition gradient: A Monitoring the condition of natural descriptive model for interpreting resources in U.S. national parks. change in aquatic ecosystems. Environmental Monitoring and Ecological Applications 16:1251-1266. Assessment 151:161–174.

Dettinger, M., K. Redmond, and D. Cayan. Feldman, D. 2006. A report to the DEQ 2004. Winter orographic precipitation Water Quality Planning Bureau on ratios in the Sierra Nevada large- the proper interpretation of two scale atmospheric circulations and recently developed macroinvertebrate hydrologic consequences. Journal of bioassessment models. Montana Hydrometeorology 5:1102-1116. Department of Environmental Quality. Planning Prevention and Assistance Dietrich, W. E., J. W. Kirchner, H. Ikeda, Division. Helena, Montana. and F. Iseya. 1989. Sediment supply and the development of the coarse Fitzpatrick, F. A., I. R. Waite, P. J. D’Arconte, surface layer in gravel bed rivers. Nature M. R. Meador, M. A. Maupin, and M. 340(20):215-217. E. Gurtz. 1998. Revised methods for characterizing stream habitat in the DiTomaso, J. M. and E. A. Healy. 2003. National Water-Quality Assessment Aquatic and riparian weeds of the West. Program. U.S. Geological Survey Water Report 3421. University of California Resources Investigations Report 98- Agriculture and Natural Resources, 4052, 67 p. Oakland, California.

Literature Cited 83 Frissell, C. A., W. J. Liss, C. E. Warren, and pollution tolerance of Trichoptera. M. D. Hurley. 1986. A hierarchical EPA-600/4-78-061. United States framework for stream habitat Environmental Protection Agency, classification: Viewing streams in a Washington, D.C. watershed context. Environmental Management 10:199-214. Harmon, M. E., J. F. Franklin, F. J. Swanson, P. Sollins, S. V. Gregory, J. D. Lattin, N. Gorman, O. T. and J. R. Karr. 1978. Habitat H. Anderson, S. P. Cline, N. G. Aumen, structure and stream fish communities. J. R. Sedell, G. W. Lienkaemper, K. Ecology 59(3):507-515. Cromack Jr., and K. W. Cummins. 1986. Ecology of Coarse Woody Debris in Grace, J.B., Schoolmaster, D.R. Jr., Pearl, Temperate Ecosystems. Adv. Ecol. Res. J., Guntenspergen, G.R., Little, A.M., 15:133–302. Mitchell, B.R., Miller, K.M., and Schweiger, E.W. 2012. Guidelines for Hauer, F. R., J. A. Stanford, and M. S. a graph-theoretic implementation of Lorang. 2007. Pattern and process in structural equation modeling. Ecosphere northern Rocky Mountain headwaters: 3:73. Ecological linkages in the headwaters of the Crown of the Continent. Journal Gregory, S. V., F. J. Swanson, W. A. McKee, of the American Water Resources and K.W. Cummins. 1991. An Association 43:104-117. ecosystem perspective of riparian zones. BioScience 41:540-551. Hauer, F. R. and G. A. Lamberti, editors. 2006. Methods in stream ecology, 2nd Groffman, P., J. Baron, T. Blett, A. Gold, I. edition. Academic Press. New York, Goodman, L. Gunderson, B. Levinson, New York. M. Palmer, H. Paerl, G. Peterson, N. Poff, D. Rejeski, J. Reynolds, M. Hauer, F. R., D. B. Fagre, and J. A. Stanford. Turner, K. Weathers, and J. Wiens. 2002. Hydrologic processes and 2006. Ecological thresholds: The key to nutrient dynamics in a pristine successful environmental management mountain catchment. Verhandlungen or an important concept with no des Internationalen Verein Limnologie practical application? Ecosystems 9:1-13. 28:1490-1493.

Hamlet, A. F. and D. P. Lettenmaier. 2007. Hauer, F. R., J. A. Stanford, J. J. Giersch, Effects of 20th century warming and and W. H. Lowe. 2000. Distribution climate variability on flood risk in the and abundance patterns of western U.S. Water Resources Research macroinvertebrates in a mountain 43:W06427. stream: an analysis along multiple environmental gradients. Verh. Internat. Hankin, D. G., and G. H. Reeves. 1988. Verein. Limnol. 27(3):1485-1488. Estimating total fish abundance and total habitat area in small streams based Hawkins, C. P., J. R. Olson, and R. A. on visual estimation methods. Can. J. Hill. 2010. The reference condition: Fish. Aquat. Sci. 45:834-844. predicting benchmarks for ecological and water-quality assessments. Journal Hari, R. E., D. M. Livingstone, R. Siber, P. of the North American Benthological Burkhardt-Holm, and H. Guttinger. Society 29:312-343. 2006. Consequences of climatic change for water temperature and brown Hawkins, C. P. 2006. Quantifying biological trout populations in Alpine rivers integrity by taxonomic completeness: and streams. Global Change Biology evaluation of a potential indicator 12:10-26. for use in regional- and global-scale assessments. Ecological Applications Harris, T. L. and T. M. Lawrence. 1978. 16:1277-1294. Environmental requirements and

84 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Hawkins, C. P. 2005. Development of a family-level biotic index. Journal of the RIVPACS (O:E) Model for assessing the North American Benthological Society biological integrity of Montana streams. 7:65. The Western Center for Monitoring and Assessment of Freshwater Ecosystems, Hilsenhoff, W. L. 1987, An improved biotic Utah State University, 10 November index of organic stream pollution. 2005. Great Lakes Entomologist 20:31-39.

Hawkins, C. P., J. Ostermiller, M. Vinson, R. Hirsch, R. M., J. R. Slack, and R. A. Smith. J. Stevenson, and J. Olsen. 2003. Stream 1982. Techniques of trend analysis algae, invertebrate, and environmental for monthly water-quality data. Water sampling associated with biological Resources Research 18:107-121. water quality assessments: Fieled Hooper, A. 1993. Effects of season, habitat, protocols. Utah State University, Logan, and an impoundment on twenty UT. five benthic community measures Hawkins, C. P., R. H. Norris, J. N. Hogue, used to assess water quality. M.S. and J. W. Feminella. 2000. Development Thesis. College of Natural Resources, and evaluation of predictive models University of Wisconsin-Stevens Point. for measuring the biological integrity 443 pp. of streams. Ecological Applications Hughes, R. M. 1995. Defining acceptable 10:1456-1477. biological status by comparing with Heard, A. M. 2005. Effects of landscape scale reference conditions. Pages 31–47 in prescribed fire on hydrology and stream W. Davis and T. Simon, eds., Biological chemistry. Colorado State University. assessment and criteria: Tools for water Fort Collins, Colorado. resource planning and decision making for rivers and streams. Boca Raton, Fla.: Heitke, J. D., E. K. Archer, R. J. Leary, Lewis Publishers. and B. B. Roper. 2011. Effectiveness monitoring for streams and riparian Hughes, R.M., D.P. Larsen, and J.M. areas: Sampling protocol for stream Omernik. 1986. Regional reference channel attributes. Unpublished paper sites: A method for assessing stream on file at:http://www.fs.fed.us/biology/ potentials. Environmental Management fishecology/emp. 10:629-635.

Helsel, D. R. 2005. Nondetects And Data Hunsaker, C. T. and D. A. Levine. 1995. Analysis: Statistics for censored Hierarchical approaches to the study environmental data. John Wily and of water quality in rivers. Bioscience Sons, New York. 250 p 3:193-203.

Hill, B. H., A. T. Herlihy, P. R. Kaufmann, Hynes, H. B. N. 1972. Ecology of running S. J. DeCelles, and M. A. Vander waters. University of Toronto Press, Borgh. 2003. Assessment of streams Toronto, Canada. of the eastern United States using a Hynes, H. B. N. 1966. The biology of periphyton index of biotic integrity. polluted waters, 3rd ed., Liverpool Ecological Indicators 2:325-338. University Press, Liverpool, UK. Hilsenhoff, W. L. 1998, A modification of the Ingman, G. L. and M. A. Kerr. 1989. Water biotic index of organic stream pollution quality in the Clark Fork River Basin, to remedy problems and to permit its Montana: State fiscal years 1988- use throughout the year. Great Lakes 1989. Montana Dept. of Health and Entomologist 31:12pp. Environmental Sciences, Water Quality Bureau. Helena, Montana. Hilsenhoff, W. L. 1988. Rapid field assessment of organic pollution with a

Literature Cited 85 Irwin, R. J. 2006. Draft Part B lite (Just the Karr, J. R. 1998. Rivers as sentinels: Using Basics) QA/QC Review Checklist for the biology of rivers to guide landscape Aquatic Vital Sign Monitoring Protocols management. Pages 502-528 in R. J. and SOPs, National Park Service, Naiman and R. E. Bilby, editors. River Water Resources Division. Fort Collins, ecology and management: Lessons Colorado, distributed on Internet only from Pacific coastal ecosystems. at http://www.nature.nps.gov/water/ Springer, New York, New York. Vital_Signs_Guidance/Guidance_ Documents/PartBLite.pdf Karr, J. R. and E. W. Chu. 1999. Restoring Life in Running Waters: Better Isaak, D. J., S. Wollrab, D. Horan, and G. Biological Monitoring. Washington, Chandler. 2011. Climate change effects DC: Island Press. on stream and river temperatures across the northwest U.S. from 1980-2009 Karr, J. R. and E. W. Chu. 1997. Biological and implications for Salmonid fishes. monitoring: Essential foundation for Climatic Change 0165-0009: 1-26. DOI ecological risk assessment. Human and 10.1007/s10584-011-0326-z. Ecological Risk Assessment 3:993-1004. Jackson, S. and W. Davis. 1995. Meeting Karr, J. A. and D. R. Dudley. 1981. Ecological the goal of biological integrity in perspectives on water quality goals. water-resource programs of the U.S. Environmental Management 5:55-68. Environmental Protection Agency. Kasahara, T. and S.M. Wondzell. 2003. Journal of the North American Geomorphic controls on hyporheic Benthological Society 13:592–597. exchange flow in mountain streams. Jessup, B., C. Hawkins, and J. Stribling. Water. Resour. Res. 39(1):1005, 2006. Biological indicators of stream doi:10.1029/2002WR001386. condition in Montana using benthic Kaufmann, P. R. (ed.) 1993. Physical Habitat. macroinvertebrates. Prepared pp. 59-69 In: R.M. Hughes (ed.).Stream for the Montana Department of Indicator and Design Workshop. Environmental Quality by TetraTech, EPA/600/R_93/138. U.S. Environmental Inc., Owings Mills, MD, and the Protection Agency, Office of Research Western Center for Monitoring and and Development, Corvallis, Oregon. Assessment of Freshwater Ecosystems, Utah State University, Logan, UT. Kaufmann, P. R. 1987. Slackwater Habitat in Torrent-impacted Streams. In: Erosion Johnson, M. M. 2002. The fate of sewage and Sedimentation in the Pacific Rim, effluent nutrients applied to coarse- R. L. Beschta, T. Blinn, G. E. Grant, grained soils by sprinkler irrigation. F. J. Swanson, and G. E. Ice (Editors). University of Montana, Bozeman, International Association of Hydrologic Montana. Science, Pub. No. 165, Proceedings of Junk, W., P. B. Bayley, and R. E. Sparks, an International Symposium, August 1989. The flood pulse concept in river- 3-7, 1986, Ore. State Univ., Corvallis, floodplain systems. Proceedings of the OR. pp. 407-408. International Large River Symposium Kaufmann, P. R. and E. G. Robison. 1998. (LARS). Canadian Special Publications Physical Habitat Characterizaton. in Fisheries and Aquatic Sciences 106: Pages 77-118 in J. M. Lazorchak, D. 110-127. J. Klemm, and D.V. Peck, editors. Karr, J. R. 1991. Biological integrity: A long- Environmental Monitoring and neglected aspect of water resource Assessment Program—surface waters: management. Ecological Applications Field operations and methods for 1:66-84. measuring the ecological condition of wadeable streams. EPA/620/R-94/004F. U.S. Environmental Protection Agency,

86 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Office of Research and Development, Lange-Bertalot, H. 1979. Pollution tolerance Washington, D.C. as a criterion for water quality estimation. Nova Hed. 64:283-304. Kaufmann, P. R. and T. R. Whittier. 1997. Habitat Assessment. Pages 5-1 to Lambing, J. H. 1991. Water-quality and 5-26 In: J. R. Baker, D. V. Peck, and transport characteristics of suspended D. W. Sutton (eds.) Environmental sediment and trace elements in Monitoring and Assessment Program streamflow of the upper Clark -Surface Waters: Field Operations Fork basin from Galen to Missoula, Manual for Lakes. EPA/620/R-97/001. Montana, 1985-90. U.S. Geological U.S. Environmental Protection Agency, Survey Water-Resources Investigations Washington, D.C. Report 91-4139:73.

Kaufmann, P. R., J. M. Faustini, D. P. Larsen, Leopold, L. B., M. G. Wolman, and J. P. and M. A. Shirazi. 2008. A roughness- Miller. 1964. Fluvial processes in corrected index of relative bed geomorphology. W.H. Freeman and stability for regional stream surveys. Co., San. Fran. CA. 522 p. Geomorphology 99: 150-70. Lindenmayer, D. B., and G. Luck. 2005. Kaufmann, P. R., P. Levine, E. G. Robison, C. Synthesis: Thresholds in conservation Seeliger, and D. Peck. 1999. Quantifying and management. Biological physical habitat in wadeable streams. Conservation 124:351–354. U.S. Environmental Protection Agency, Washington, D.C. Lisle, T. E. 1982. Effects of aggradation and degradation on riffle-pool Knowles, N., M. D. Dettinger, and D. R. morphology in natural gravel channels, Cayan. 2006. Trends in snowfall versus northwestern California. Water. Resour. rainfall in the western United States. Res. 18(6):1643-1651. Journal of Climate 19:4545-4559. Lohman, K. and J. C. Priscu. 1992. Kolkwitz, R. and M. Marsson. 1908. Physiological Indicators of Ecology of plant saprobia. [Translated Nutrient Deficiency in Cladophora 1967]. Pages 47-52 in L. E. Keup, W. (chlorophyta) in the Clark Fork of the M. Ingram, and K. M. MacKenthum, Columbia River, Montana. Journal of editors. Biology of water pollution. Phycology 28: 443-448. Federal Water Pollution Control Administration, Washington, DC. Lowe, R. L. and Y. Pan. 1996. Benthic algal communities and biological monitors. Kumar, S., S. A. Spaulding, T. J. Stohlgren, Pages 705-39 in R. J. Stevenson, M. K. A. Hermann, T. S. Schmidt, and Bothwell, and R. L. Lowe, editors. L. L. Bahls. 2009. Potential habitat Algal ecology: Freshwater benthic distribution for the freshwater diatom ecosystems. Academic Press, San Diego, Didymosphenia geminata in the California. continental U.S. Frontiers in Ecology and the Environment 7:415–420. MacDonald, D. D., C. G. Ingersoll, and T. A. Berger. 2000. Development Kusnierz, P., A. Welch and D. Kron. and evaluation of consensus-based 2013. The Montana Department of sediment quality guidelines for Environmental Quality Sediment freshwater ecosystems. Archives of Assessment Method: Considerations, Environmental Contamination and Physical and Biological Parameters, Toxicology 39(1):20-31. and Decision Making. Helena, MT: Montana Dept. of Environmental MacDonald, L. H., A. W. Smart, and R. C. Quality. Wissmar. 1991. Monitoring guidelines to evaluate effects of forestry activities on streams in the Pacific Northwest

Literature Cited 87 and Alaska. Prepared for Region 10, Meador, M. R., C. R. Hupp, T. F. Cuffney, U.S. Environmental Protection Agency, and M. E. Gurtz. 1993. Methods for Seattle, Washington. EPA 910/9-91-001. characterizing stream habitat as part of the National Water-Quality Assessment Madej, M. A. 2001. Development of channel Program. U.S. Geological Survey Open- organization and roughness following File Report 93-408. U.S. Geological sediment pulses in single-thread, gravel Survey, Raleigh, N.C. bed rivers. Water Resource Research 37(8):2259-2272. Merritt, R.W., Cummins, K.W., and Berg, Mast, M. A. 2007. Assessment of historical M.B.. eds., 2008. An introduction to water-quality data for National Park the aquatic insects of North America th units in the Rocky Mountain Network, (4 ed.): Dubuque, Iowa, Kendall Hunt Colorado and Montana, through 2004. Publishing Company, 1,158 p. 2007–5147, U.S. Geological Survey. Merritt, D. M. and E. E. Wohl. 2002. Denver, CO. Processes governing hydrochory Mast, M. A., D. H. Campbell, D. P. along rivers: Hydraulics, hydrology, Krabbenhoft, and H. E. Taylor. 2005. and dispersal phenology. Ecological Mercury transport in a high-elevation Applications 12:1071-1087. watershed in Rocky Mountain National Montana Department of Agriculture (MT Park, Colorado. Water Air and Soil DOA). 2012. Montana Noxious Weed Pollution 164:21-42. Program. Noxious Weed List. Available McGuire, D. L. 2010. Clark Fork River at http://agr.mt.gov/agr/Programs/ Biomonitoring: Macroinvertebrate Weeds/PDF/weedList2010.pdf Community Assessments in 2009. (accessed 14 December 2012). United States Environmental Protection Agency, Washington, D.C. Montana Department of Environmental Quality (MT DEQ). 2012a. Water McGuire, D. L. 1993. Clark Fork River quality planning bureau field biomonitoring: Macroinvertebrate procedures manual for water community assessments in 2009. United quality assessment monitoring States Environmental Protection Version 3.0. Montana Department Agency, Washington, D.C. of Environmental Quality. Helena, Montana. McGuire, D. L. 1989. Clark Fork River Aquatic Macroinvertebrate Survey, Montana Department of Environmental August, 1987. Technical report prepared Quality (MT DEQ). 2012b. for the Montana Department of Health Montana 2012 Final Water Quality and Environmental Sciences/Water Integrated Report. Montana Dept. of Quality Bureau, Helena, Montana. Environmental Quality, Helena, MT.

McGuire, D. L. 1987. Clark Fork River Montana Department of Environmental Macroinvertebrate Study, 1986. Quality (MT DEQ). 2012c. Analysis Technical report prepared for the of Base Parameter Data and Erosion Montana Governor’s Office and Inventory Data for Sediment TMDL Montana Water Quality Bureau. Development within the Upper Clark Helena, Montana. Fork TPA. Prepared by Watershed McKnight, D. M., R. Harnish, R. L. Consulting, LLC Missoula, MT. Wershaw, J. S. Baron, and S. Schiff. Montana Department of Environmental 1997. Chemical characteristics of Quality (MT DEQ). 2012d. Sample particulate, colloidal, and dissolved Collection, Sorting, Taxonomic organic material in Loch Vale Identification, and Analysis of Benthic Watershed, Rocky Mountain National Macroinvertebrate Communities Park. Biogeochemistry 36:99-124.

88 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Standard Operating Procedure, of Environmental Quality. Helena, WQPBWQM-009. Montana Dept. of Montana. Environmental Quality, Helena, MT. Montana Department of Environmental Montana Department of Environmental Quality (MT DEQ). 2008a. Monitoring Quality (MT DEQ). 2012e. Current and maintenance plan for the MT Circular DEQ-7, Montana numeric DEQ 2011. Department Circular water quality standards. Available at Deq-12, Parts A and B Montana http://deq.mt.gov/wqinfo/Standards/ base numeric nutrient standards and default.mcpx (accessed May 2013). nutrient standards variances. Montana Department of Environmental Quality. Montana Department of Environmental Helena, Montana. Quality (MT DEQ). 2011a. Department Circular Deq-12, Parts A and B Montana Department of Environmental Montana base numeric nutrient Quality (MT DEQ). 2008b. Monitoring standards and nutrient standards and maintenance plan for the Clark variances. Montana Department Fork Operable Unit. Montana of Environmental Quality. Helena, Department of Environmental Quality. Montana. Helena, Montana.

Montana Department of Environmental Montana Department of Environmental Quality (MT DEQ). 2011b. Periphyton Quality (MT DEQ). 2006. Standard Operating Procedure, Sample collection, sorting, and WQPBWQM-010. Montana Dept. of taxonomic identification of benthic Environmental Quality, Helena, MT. macroinvertebrates. Water Quality Planning Bureau Standard Operation Montana Department of Environmental Procedure WQPBWQM-009. Montana Quality (MT DEQ). 2011c. Tobacco Department of Environmental Quality. Planning Area Sediment TMDLs Helena, Montana. and Framework Water Quality Improvement Plan. Helena, Montana Department of Environmental MT: Montana Department of Quality (MT DEQ). 2005. Quality Environmental Quality. http://deq. assurance project plan, sampling mt.gov/wqinfo/TMDL/finalReports. and water quality assessment of mcpx. Accessed 1/16/2013. stream and rivers in Montana, 2005: WQPBQAP–02, 80 p. Montana Department of Environmental Quality (MT DEQ). 2010a. Interim Montana Department of Environmental comprehensive long-term monitoring Quality (MT DEQ). 2002. Water-use plan for the Clark Fork River classifications: Administrative Rules of operable unit. Prepared by PBS&J for Montana 17.30.607 to 17.30.611, p. 17- Prepared for: Montana Department 2697 to 17-2707. of Environmental Quality Mine Waste Cleanup Bureau. Project No. Montana Department of Fish, Wildlife 100007965. Montana Department and Parks (MT FWP). 2012. Aquatic of Environmental Quality. Helena, Invasive Species - Identification and Montana. Distribution website. http://fwp.mt.gov/ fishing/guide/ANS/identification.html Montana Department of Environmental (accessed 14 December 2012) Quality (MT DEQ). 2010b. Upper Clark Fork River Tributaries Sediment, Montana Department of Fish, Wildlife and Metals, and Temperature TMDLs Parks (MT FWP). 2008. Statewide and Framework for Water Quality integrated noxious weed management Restoration. C01-TMDL-02a-F plan. Helena, Montana. Version 1.0. Montana Department

Literature Cited 89 Moore, J. N. and H. W. Langner. 2012. Can a National Oceanic and Atmospheric River Heal Itself? Natural Attenuation Administration (NOAA). 2012. of Metal Contamination in River NOAA Photo Library. http://www. Sediment. Environmental Science & photolib.noaa.gov/htmls/nerr0965.htm Technology 46 (5):2616-2623. (accessed 14 December 2012).

Morgan, P., G. H. Aplet, H. B. Haufler, National Park Service (NPS). 2012. NPS H. C. Humphries, M. M. Moore, Water Resource Division Hydrographic and W. D. Wilson. 1994. Historical & Impairment Statistics web site. http:// range of variability: A useful tool for www.nature.nps.gov/water/HIS/index. evaluating ecosystem change. Journal of cfm (accessed 14 December 2012). Sustainable Forestry 2:87-111. National Park Service (NPS). 2006. Morisawa, M. 1968. Streams, their dynamics Management Policies 2006. U.S. and morphology. McGraw-Hill Book Department of the Interior. National Company, New York. 175 p. Park Service. Washington, D.C.

Mote, P. W., A. F. Hamlet, M. P. Clark, and National Park Service (NPS). 2005. Park D. P. Lettenmaier. 2005. Declining Planning Program Sourcebook: General mountain snowpack in western north Management Planning. Final draft. America. Bulletin of the American Washington, D.C.: NPS. Meteorological Society 86:39-49 National Park Service (NPS). 2002. Muhlfeld, C., and B. Marotz. 2005. Seasonal Recommendations for core water movement and habitat use by sub-adult quality monitoring parameters bull trout in the upper Flathead River and other key elements of the NPS system, Montana. North American vital signs program water quality Journal of Fisheries Management monitoring component. Unpublished 25:797-810. Report-573285

Mulholland, P. J., J. L. Tank, D. M. Sanzone, National Park Service (NPS). 1970. Act W. M. Wollheim, B. J. Peterson, J. to improve the administration of R. Webster, and J. L. Meyer, 2000. the National Park System (General Nitrogen Cycling in a Forest Stream Authorities Act). Available at http:// Determined by a 15n tracer addition. www.cr.nps.gov/history/online_books/ Ecological Monographs 70:471-493. anps_7a.htm (Accessed 18 December 2012) Naiman, R. J. 1992. Watershed management Nimick, D.A., 2003, Diurnal variation in balancing sustainability and trace-metal concentrations in streams: environmental change. Springer-Verlag, U.S. Geological Survey Fact Sheet FS- New York, New York. 086-03, 4 p. Naiman, R. J. and H. Decamps. 1997. Ode, P. R., Hawkins, C.P., and Mazor, R.D. The ecology of interfaces: Riparian 2008. Comparability of biological zones. Annual Review of Ecology and assessments derived from predictive Systematics 28:621-658 models and multimetric indices of Naiman, R. J., H. DeCamps, and M. Pollack. increasing geographic scope. Journal 1993. The role of riparian corridors of the North American Benthological in maintaining regional biodiversity. Society 27:967-985. Ecological Applications 3:209-212. Ostermiller, J. D. and C.P . Hawkins. Naiman, R. J., H. Decamps, J. Pastor, and 2004. Effects of sampling error on C.A. Johnston. 1988. The potential bioassessments of stream ecosystems: importance of boundaries to fluvial application to RIVPACS-type models. ecosystems. J. North Amer. Benthol. Soc. Journal of the North American 7(4):289-306. Benthological Society 23:363-382.

90 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Palmer, M. A. and E. S. Bernhardt. 2006. Potapova, M. 2009. Diatoma moniliformis. In Hydroecology and river restoration: Diatoms of the United States. Available Ripe for research and synthesis. at: http://westerndiatoms.colorado.edu/ Water Resources Research 42:W03S07. taxa/species/Diatoma_moniliformis doi:10.1029/2005WR004354. (accessed 10 April 2012).

Patten, D. T. 1998. Riparian ecosystems of Potapova, M., J. F.Coles, E. M. P. Giddings, semi-arid North America: Diversity and and H. Zappia. 2005. A comparison human impacts. Wetlands 18:498-512. of the influences of urbanization on stream benthic algal assemblages in Peck, D. V., J. M. Lazorchak, and D. J. contrasting environmental settings. Klemm, editors. 2006. Environmental Pages 333-359 in L. R. Brown, R. H. monitoring and assessment program, Gray, R. M. Hughes, and M. R. Meador, surface waters: Western pilot editors. Effects of urbanization on study-field operations manual for stream ecosystems. American Fisheries wadeable streams. U.S. Environmental Society, Symposium 47, Bethesda, Protection Agency, EPA/620/R–06/003, Maryland. Washington, D.C. Prescott, G. W. and G. E. Dillard. 1979. Pederson, G. T., L. J. Graumlich, D. B. Fagre, A checklist of algal species reported T. Kipfer, and C. C. Muhlfeld. 2010. from Montana, 1891-1977: Montana A century of climate and ecosystem Academy of Sciences, Supplement to change in western Montana: What do the Proceedings, Monograph no. 1, v. temperature trends portend? Climatic 38, 102 p. Change 98:133-154. Radford, J. Q. and A. F. Bennett. 2004. Plafkin, J. L., M. T. Barbour, K. D. Porter, Thresholds in landscape parameters: and R. M. Hughes. 1989. Rapid Occurrence of the white-browed bioassessment protocols for use treecreeper Climacteris affinis in streams and rivers—benthic in Victoria, Australia. Biological macroinvertebrates and fish. U.S. Conservation 117:375-391. Environmental Protection Agency, EPA 440.4–89–001, 194 p. Rahel, F. J. and J. D. Olden. 2008. Assessing the effects of climate change on aquatic Platts, W. S., W. F. Megahan, and G. W. invasive species. Conservation Biology Minshall. 1983. Methods for evaluating 22:521-533. stream, riparian and biotic conditions. Gen. Tech. Rep. INT-138. U.S. Forest Relyea, C., G. Minshall, and R. Danehy. 2000. Service, Intermountain Forest and Stream insects as bioindicators of fine Range Experiment Station, Ogden, UT. sediment. Watershed Management 70 p. 2000 Conference. Stream Ecology Center, Department of Biological Poff, N. L., J. D. Allan, M. B. Bain, J. R. Karr, Sciences, Idaho State University. Boise, K. L. Prestegaard, B. D. Richter, R. Idaho. E. Sparks, and J. C. Stromberg, 1997. The Natural Flow Regime. BioScience Resh, V. H. and J. K. Jackson. 1993. 47(11):769-784. Rapid assessment approaches to biomonitoring using benthic Poff, N. L. and J. D. Ward. 1989. Implications macroinvertebrates. Pages 195-233 of streamflow variability and in D. M. Rosenberg and V. H. Resh, predictability for lotic community editors. Freshwater biomonitoring and structure: A regional analysis of benthic macroinvertebrates. Chapman streamflow patterns.Can. J. Fish. Aq. and Hall, New York. Sci. 46:1805-1818. Resh, V. H., R. H. Norris, and M. T. Barbour. 1995. Design and implementation of

Literature Cited 91 rapid assessment approaches for water theoretic approach to multimetric index resource monitoring using benthic construction. Ecological Indicators macroinvertebrates. Australian Journal 26:14-23. of Ecology 20:108-121. Schweiger, E. W., R. Hauer, I. Ashton, D. Ringold, P. L., T. K. Magee, and D. V. Peck. Shorrock, L. O’Gan, and M. Britten. 2008. Twelve invasive plant taxa in US In Review. Stream ecological integrity western riparian ecosystems. Journal protocol. Rocky Mountain Network of the North American Benthological Version 2.0 DRAFT. NPS/XXXX/ Society 27(4):949-966. NRTR—2014/XXX.

Rinne, J. 1988. Effects of livestock Schweiger, E. W., I. A. Ashton, C. C. grazing exclosure on aquatic Muhlfeld, L. A. Jones, and L. L. macroinvertebrates in a montane Bahls. 2011. The distribution and stream, New Mexico. Great Basin abundance of a nuisance native alga, Naturalist 48(2):146-153. Didymosphenia geminate, in streams of Glacier National Park. Park Science Robison, E. G., and R. L. Beschta, 1989. 9:78-81. Estimating Stream Cross Sectional Area from Wetted Width and Thalweg Scudder, B. C., L. C. Chasar, D. A. Wentz, N. Depth. Phys. Geogr. 10(2):190-198. J. Bauch, M. E. Brigham, P. W. Moran, and D.P. Krabbenhoft. 2009. Mercury Rosenberg, D. M. and Resh, V. H. 1996. Use in fish, bed sediment, and water from of aquatic insects in biomonitoring. streams across the United States, 1998- Pages 87-97 in R. W. Merritt and K. 2005. U.S. Geological Survey Scientific W. Cummins. An introduction to the Investigations Report 2009–5109. aquatic insects of North America, 3rd Ed., Kendall/Hunt Publishing Seastedt, T. R., W. D. Bowman, T. N. Caine, Company, Dubuque, Iowa. D. McKnight, A. Townsend, and M. W. Williams. 2004. The landscape Schertz, T. L., R. B. Alexander, and D. J. Ohe. continuum: A model for high-elevation 1991. The computer program EStimate ecosystems. Bioscience 54:111-121. TREND (ESTREND): A system for the detection of trends in water-quality Selong, J. H., T. E. McMahon, A. V. Zale, data. U.S. Geological Survey Water and F. T. Barrows. 2001. Effect of Resources Investigations Report. temperature on growth and survival of bull trout, with application of an Schoolmaster, D. R. Jr., J. B. Grace, and improved method for determining E. W. Schweiger. 2012. A general thermal tolerance for fishes. theory of multimetric indices Transactions of the American Fisheries and their properties. Methods Society 130:1026-1037. in Ecology and Evolution. doi: 10.1111/j.2041-210X.2012.00200.x. Simons, D.B. and F. Senturk. 1977. Sediment Transport Technology. Water Resources Schoolmaster, D. R. Jr., J. B. Grace, E. Publications. Fort Collins, CO. 80522, W. Schweiger, B. R. Mitchell, and USA. 807 p. G. R. Guntenspergen. 2013a. A causal examination of the effects of Simonson, T. D., J. Lyons, and P. D. Kanehl. confounding factors on multimetric 1994. Quantifying fish habitat in indices. Ecological Indicators streams: transect spacing, sample size, 29:411–419. and a proposed framework. North Amer. J. Fish. Mgmt.14:607-615. Schoolmaster, D. R. Jr., J. B. Grace, E. W. Schweiger, G. R. Guntenspergen, B. R. Slack, J. R. and D. L. Lorenz. 2003. USGS Mitchell, K. M. Miller, and A. M Little. library for S-PLUS for Windows, 2013b. An algorithmic and information- Release 2: U.S. Geological Survey

92 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Open-File Report 2003–357 Available Stevenson, R. J. and J. P. Smol. 2003. Use of at: http://pubs.er.usgs.gov/usgspubs/ofr/ algae in environmental assessments. ofr03357 (accessed December 2007). Pages 775-804 in J. D. Wehr and R. G. Sheath, editors. Freshwater algae Smith, J. G., J. J. Beauchamp and A. J. in North America: Classification and Stewart. 2005. Alternative Approach ecology. Academic Press, San Diego, for Establishing Acceptable Thresholds California. on Macroinvertebrate Community Metrics. Journal of the North American Stoddard, J. L., A. T. Herlihy, D. V. Peck, Benthological Society 24:428-440 R. M. Hughes, T. R. Whittier, and E. Tarquinio. 2008. A process for Smith, J. D., J. H. Lambing, D. A. Nimick, creating multimetric indices for large- C. Parrett, M. Ramey, and W. Schafer. scale aquatic surveys. Journal of the 1998. Geomorphology, flood-plain North American Benthological Society tailings, and metal transport in the 27:878-891. upper Clark Fork valley, Montana. U.S. Geological Survey. Stoddard, J. L., D. P. Larsen, C. P. Hawkins, R. K. Johnson, and R. H. Norris. 2006. Spaulding, S. A., D. J. Lubinski, and M. Setting expectations for the ecological Potapova. 2010. Diatoms of the United condition of streams: The concept States. http://westerndiatoms.colorado. of reference condition. Ecological edu (Accessed on 13 December, 2012.) Applications 16:1267-1276.

Spaulding, S. and W. Elwell. 2007. Increase Stoddard, J. L., D. V. Peck, S. G. Paulsen, J. in nuisance blooms and geographic Van Sickle, C. P. Hawkins, A. T. Herlihy, expansion of the freshwater R. M. Hughes, P. R. Kaufmann, D. diatom Didymosphenia geminata: P. Larsen, G. Lomnicky, A. R. Olsen, Recommendations for response. EPA S. A. Peterson, P. L. Ringold, and Region 8 and Federation of Fly Fishers, T. R. Whittier. 2005. An ecological Denver, Colorado. assessment of western streams and rivers. U.S. Environmental Protection Stack, W. R. 1989. Factors Influencing Agency, Washington, D.C. Pool Morphology in Oregon Coastal Streams. M.S. Thesis, Oregon State Stottlemyer, R., C. A. Troendle, and D. University. 109 p. Markowitz. 1997. Change in snowpack, soil, water, and streamwater chemistry Stack, W. R. and R. L. Beschta. 1989. Factors with elevation during 1990. Fraser influencing pool morphology in Oregon Experimental Forest, Colorado. Journal coastal streams. pp. 401-411 In: W. of Hydrology 195:114-136. W. Woessner and D. F. Potts (eds.). Headwaters Hydrology Symposium. Suplee, M.W., V. Watson, W. K. Dodds, American Water Resources Association. and C. Shirley, 2012. Response of 708 p. Algal Biomass to Large-scale Nutrient Controls in the Clark Fork River, Steedman, R. J. 1994. Ecosystem health as a Montana, United States. Journal management goal. Journal of the North of the American Water Resources American Benthological Society 13: Association 48(5): 1008-1021. DOI: 605–610. 10.1111/j.1752-1688.2012.00666. Stevenson, R. J. 1996. An introduction to Suplee, M.W. and R. Sada de Suplee. algal ecology in freshwater benthic 2011. Assessment Methodology habitats. Pages 3-30 in R. J. Stevenson, for Determining Wadeable Stream M. Bothwell, R. L. Lowe, editors. Impairment Due to Excess Nitrogen Algal ecology: Freshwater benthic and Phosphorus Levels. Helena, MT: ecosystems. Academic Press, San Diego, Montana Dept. of Environmental California. Quality.

Literature Cited 93 Suplee, M. W., V. Watson, A. Varghese, and J. U.S. Environmental Protection Agency Cleland, 2008. Scientific and technical (EPA). 2012. Didymosphenia geminata: basis of the numeric nutrient criteria for A nuisance freshwater http://www.epa. Montana’s wadeable streams and rivers, gov/region8/water/didymosphenia/ and addendums. Montana Department (accessed 14 December 2012). of Environmental Quality, Helena, Montana. U.S. Environmental Protection Agency (EPA). 2009a. National recommended Suplee, M., R. S. de Suplee, D. Feldman, water quality criteria. United States and T. Laidlaw. 2005. Identification Office of Water, Office of Science and assessment of Montana reference and Technology. U.S. Environmental streams: A follow-up and expansion Protection Agency, Washington, D.C. of the 1992 Benchmark Biology Study. Montana Dept. of Environmental U.S. Environmental Protection Agency Quality, Helena, Montana. (EPA). 2009b. Clark Fork River Operable Unit Site Description and Teply, M. 2010a. Diatom biocriteria Cleanup Progress. U.S. EPA 2009. for Montana streams. Montana Available at: http://www.epa.gov/ Department of Environmental Quality. region8/superfund/mt/milltowncfr/cfr/ Helena, Montana. index.html#1

Teply, M. 2010b. Interpretation of U.S. Environmental Protection Agency periphyton samples from Montana (EPA). 2007. National rivers and streams. Montana Department of streams assessment: Field operations Environmental Quality. Helena, manual. EPA-841-B-07-009. U.S. Montana. Environmental Protection Agency, Washington, D.C. Teply, M. and L. Bahls. 2006. Interpretation of periphyton samples for U.S. Environmental Protection Agency Montana streams: Middle Rockies (EPA). 2004. Clark Fork River Operable Ecoregion. Montana Department Unit of the Milltown Reservoir/Clark of Environmental Quality. Helena, Fork River Superfund Site, Record of Montana. Decision Part 1: Declaration. US EPA BOI032130003.DOC/KM. Washington, Teply, M. and L. Bahls. 2005. Diatom D.C. biocriteria for Montana streams. Montana Department of U.S. Environmental Protection Agency Environmental Quality. Helena, (EPA). 1996. Sampling ambient water Montana. for trace metals at EPA water quality criteria levels. U.S Environmental Tri-State Water Quality Council. 2009. Final Protection Agency, Office of Water, Report: The Clark Fork River Voluntary Washington, D.C. Nutrient Reduction Program, 1998- 2008. The Progress of a Voluntary U.S. Environmental Protection Agency Program in Reducing Nutrients and (EPA). 1988. Ambient water quality Noxious Algae in the Clark Fork River criteria for chloride, 1988. U.S. in Montana. The Tri-State Water Environmental Protection Agency Quality Council, VNRP Committee. Office of Water. EPA 440/5-88-001. Washington, D.C. U.S. Army Corps of Engineers (USACE), 2010. HEC-SSP Statistical Software U.S. Fish and Wildlife Service (USFWS). Package, User’s Manual, Version 2.0, 1998. A framework to assist in making Hydrologic Engineering Center, Davis, Endangered Species Act determinations California. of effect for individual or grouped

94 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site actions at the bull trout subpopulation White, D. S., C. H. Elzinga, and S. P. watershed scale. Washington, D.C. Hendricks. 1987. Temperature patterns within the hyporheic zone of a northern U.S. Fish and Wildlife Service (USFWS). Michigan river. Journal of the North 2002. Chapter 3, Clark Fork River Amer. Benthological Society 6:85-91. Recovery Unit, Montana, Idaho, and Washington. 285 p. U.S. Fish and Whittier, T. R., R. M. Hughes, J. L. Stoddard, Wildlife Service. Bull Trout (Salvelinus G. A. Lomnicky, D. V. Peck, and confluentus) Draft Recovery Plan. A. T. Herlihy. 2007. A structured Portland, Oregon. approach for developing indices of biotic integrity: Three examples from U.S. Geological Survey (USGS). 2012a. streams and rivers in the western USA. National Water Information System: Transactions of the American Fisheries Web Interface http://waterdata.usgs. Society 136:718-735. gov/nwis/uv?12324200 (accessed 14 December 2012). Wilcock, P. R. 1998. Two-fraction model of initial sediment motion in gravel-bed U.S. Geological Survey (USGS). 2012b. rivers. Science 280:410-412. Water Quality Technical Memorandums http://water.usgs.gov/admin/memo/ Wilcock, P. R., 1997. The Components QW/qw91.10.html (accessed 14 of Fractional Transport Rate. Water December 2012). Resour. Res. 33(1): 247-258. van Dam, H., A. Mertens, and J. Sinkeldam. Williams, J. E., N. G. Gillespie, H. M. 1994. A coded checklist and ecological Nelville, and W. T. Colyer. 2007. indicator values of freshwater Healing troubled waters: Preparing diatoms from The Netherlands. trout and salmon habitat for a changing Netherlands Journal of Aquatic Ecology climate. Trout Unlimited, Arlington, VA. 28(1):117-133. Woessner W. W. and M. Johnson. 2001. Vannote, R. L., G. W. Minshall, K. W. Water resource characterization report: Cummins, J. R. Sedell, and C.E. 2000 and 2001 field seasons, Grant Cushing. 1980. The river continuum Kohrs Ranch National Historic Site. concept. Canadian Journal of Fisheries National Park Service, Department and Aquatic Sciences 37:130-137. of Interior. Department of Geology, The University of Montana. Missoula, Van Sickle, J. 2008. An index of Montana. compositional dissimilarity between observed and expected assemblages. Wolock, D. M. D. M. and McCabe. 1999. Journal of the North American Explaining spatial variability in mean Benthological Society 27:227-235. annual runoff in the conterminous United States. Climate Research Van Sickle, J., D. P. Larsen, and C. P. 11:149-159. Hawkins. 2007. Exclusion of rare taxa affects performance of the O:E index in bioassessments. Journal of the North American Benthological Society 26(2):319-331.

Walking Shadow Ecology. 2012. Climate data Summarizer Version 4.4. http:// www.yellowstoneecology.com/ research/COOP_data.html (accessed 14 December 2012).

Literature Cited 95

Appendix A: Field Method Overview and Justification

We present summaries of field data quality conditions among multiple NPS collection methods and a brief justification monitoring programs (NPS 2002). The for the major responses we collect in the use of the word “core,” however, does not following sections. Schweiger et al. (In imply that these parameters are more or less Review) presents full details for all methods. important than other parameters. In situ data were collected with a handheld multi- Water and Sediment Physiochemistry parameter probe (also known as a sonde). Water and sediment physiochemistry In 2007 and 2008 data were collected with are indicators that have been historically an In Situ 9500. Beginning in 2009, all data used to monitor the condition of a stream; were collected with a YSI ProPlus. These however, they are just two indicators we two instruments were compared in the use to estimate “water quality” within lab before deployment, following WRD the SEI protocol. We also rely on habitat guidance, but were not used congruently and biological measures. Moreover, the in the field due to budget restrictions. We sampling methodology we use for water also collected continuous data on stream physiochemistry has implications for data temperature using small sensors submerged interpretation. Episodic grab samples in the channel near the bridge. All field represent conditions at the point and time methods follow NPS WRD protocols for of sampling; they do not represent the quality assurance (i.e., calibration routines condition of the entire water body, spatially and criteria). or temporally. We collected multiple water Water temperature is a critical variable physiochemistry samples on the limbs of controlling many ecosystem processes, the hydrograph (rising, peak, falling, base, both physical and biological, and it can and winter) in order to better understand impact almost all functions within an the range of conditions that can occur ecosystem (Allan 2004). Rates of most across a water year. A water year is defined physical, chemical, and biological processes as the 12-month period beginning October are strongly influenced by temperature. 1 through September 30. The water year is It is also a critical parameter for tracking designated by the calendar year in which climate change response in park ecosystems. it ends and which includes 9 of the 12 Water pH (the measure of water hydrogen months. For example, the year October ion concentration) has many physical and 1, 1998-September 30, 1999 is called the biological effects. Most aquatic species “1999” water year. Thus, important diel occur within specific habitat envelopes of patterns in some water quality parameters pH conditions, and changes in pH will likely are not captured. Over the duration of the result in changes in species assemblages. In GRKO SEI pilot we gradually increased addition, the pH determines the solubility of sampling frequency from one to four times many heavy metals, which is very important per year in an attempt to capture all of these in GRKO given contaminated sediments. phases of the water cycle at GRKO. Specific conductance is the ability of a water body to conduct an electric current and In situ is directly correlated with dissolved ion Four in situ (measured within the concentrations in water bodies. In essence, stream channel) core field parameters the more dilute the water, the lower the (temperature, pH, specific conductance, concentrations of dissolved salts and thus and dissolved oxygen) are measured as the lower the conductance. Changes in part of all funded NPS WRD water quality conductance suggest changes in major ions monitoring protocols. As such, these or nutrients, such as potassium, calcium, and attributes contribute some measure of other anions and cations. Dissolved oxygen consistency and comparability of water is closely linked to physical and biological

Appendix A: Field Method Overview and Justification 97 processes. For instance, respiration, few trace elements in 2010 were analyzed photosynthesis, and atmospheric exchange by the Environmental Testing Corporation (through turbulence in rapids and riffles) lab. Each lab was selected given expertise are the principle processes that affect or are and operational constraints within the affected by dissolved oxygen concentrations. SEI protocol development at GRKO. Each In addition to high water temperatures, high lab followed rigorous internal QAQC that microbial activity, which is driven by organic ensured comparability across lab results. pollution, drives demand for dissolved Future monitoring at GRKO will use the oxygen resulting in anoxic conditions. High Kiowa and EPA labs only. oxygen levels are especially critical for the metabolism of aquatic insects and salmonid Sampled nutrients include the dominant eggs. forms of nitrogen and phosphorous (both total and dissolved). We also sampled Water and Sediment Grab Samples organic and total dissolved carbon. Nutrients Given similar parameters collected, may be limiting in aquatic ecosystems, analytical methods, application and controlling ecosystem productivity, as well interpretation of criteria, we group water as being indicators of eutrophication caused and sediment physiochemistry together by external stressors (e.g., atmospheric in the following sections. We use the deposition or visitor use activities). Total less-common term “physiochemistry” organic carbon and dissolved organic carbon because we include both physical (i.e., are essential components of the carbon cycle temperature) and chemical (i.e., nutrients) in streams and their watershed. Dissolved parameters in the group. We measured organic matter may impact contaminant almost 60 parameters based on a variety transport as well as drinking water quality. of considerations. Parameters included 11 Major ions include two predominant nutrients (and carbon), 11 major ions, 12 anions (sulfate and chloride) and four trace elements (both dissolved and total cations (calcium, sodium, potassium, recoverable metals) in water and 13 metals and magnesium). These six ions, along in sediment. Water samples were taken using with carbonates, make up most of the single location, depth-integrated thalweg ions in stream water. They are important methods (at low flow) or equal width/ indicators of the watershed context of the depth-integrated samples taken from the stream, with different ion concentrations bridge at higher flows. Bulk bed sediment reflecting variation in geology, vegetation, samples were composited from seven to and weathering processes. However, ten individual samples of fine-grained bed sulfate is also common as an indicator sediment collected by scooping material of pollution (e.g., from mining waste or from the surfaces of representative deposits agricultural runoff). We also include total along pool or low-velocity areas at each suspended solids (TSS); high concentrations transect location or at ad hoc locations of particulate matter can cause increased near where water samples were collected. sedimentation and siltation in a stream, Sediments were not sieved. All ROMN SEI which in turn can impact important habitat Quality Assurance/Quality Control (QAQC) areas for aquatic life. Suspended particles field procedures were followed. also provide attachment places for other Samples were analyzed at four laboratories pollutants, such as metals and bacteria. Trace depending on the sample date and elements in water include those that typically parameter. In 2008, all samples went to the occur only in minute concentrations, such Flathead Lake Biological Station (FLBS) lab. as metals. Contamination of a stream with In 2009 and 2010, all nutrient samples, major trace metals is not always detectable in the elements and some trace elements went to water column, however, because they may the University of Colorado Kiowa lab. Most have precipitated or adsorbed to organic trace elements in water during 2009 and particulates or fine sediments. Therefore, all sediment parameters in 2009 and 2010 given the importance of this issue at GRKO, were analyzed by the EPA Region 8 lab. A we also sample sediments deposited from

98 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site the water column to detect trace metal AFDM also includes bacteria, fungi, small contaminants. Many metal ions are lethal fauna, and organic detritus. Together these to fish and other aquatic life forms. Metals measures complement the species list-based often are bio-concentrated, leading to diatom metrics described above and may be increasing concentrations in species higher especially important in studies that address in the food chain. We include total mercury potential nutrient enrichment or toxicity. in the suite of parameters analyzed in the sediment samples. Mercury has no known metabolic purpose and is toxic to living Physical Habitat organisms. In humans, mercury adversely Note: much of the approach in the SEI affects the central nervous system. Mercury protocol to habitat characterization and the can be converted from an inorganic narrative used in this report for describing compound (which is what we measure), to physical habitat methods, metrics, and organic forms such as methylmercury, which interpretation is based on similar treatments is easily absorbed by organisms, but harder graciously provided by researchers from the and more expensive to monitor. EPA (Kaufmann et al. 1999, Stoddard et al. 2005).

Stream Productivity Quantitative characterization of stream Two indicators of stream productivity were physical habitat is a core element of long- also created from the composite periphyton term monitoring of the Clark Fork at GRKO. samples (see below): chlorophyll-a and ash Stream habitat is an important component free dry mass (AFDM). Known volumes of aquatic resource monitoring because it were filtered from the composite samples in can help describe the context or template for the field and frozen for later analysis. ecosystem function and condition (Frissell AFDM and chlorophyll-a samples were et al. 1986, Kaufmann et al. 1999). Habitat processed using standard methods (APHA is often a key covariate of stream biology 1995). Filters were pre- and post-weighed and can help us understand patterns in after combustion in a muffle furnace for aquatic communities by helping predict the AFDM. Chlorophyll-a was extracted and presence or absence of organisms. It may analyzed via fluorescence. AFDM and also help explain important spatial variation chlorophyll-a concentration per unit area in other aspects of stream community were generated using the area and volume structure. Additionally, habitat is a useful of sample collected and volume of sample monitoring endpoint itself (e.g., increases in filtered. Note that this method is not the the percent of fine sediments) and can help same as the current protocol used by MT us understand changes in stressors such as DEQ (Suplee and Suplee 2011). particular land uses outside or inside a park, including possible visitor impacts. We present the stream productivity measures within the water chemistry The physical habitat components of the section as they are used as indicators of ROMN SEI protocol are based on research general nutrient loads within a stream. conducted by the EPA over the last two Nutrient concentration is correlated with decades (Kaufmann et al. 1999, Stoddard et ecosystem disturbance (e.g., deforestation al. 2005). Many of these methods have also and agriculture) as periphyton production been adopted by MT DEQ. Importantly, our declines with increasing river size and approach confines observations to habitat turbidity (Naiman et al. 1993). Chlorophyll-a characteristics themselves versus subjective typically ranges from 0.5 to 2% of total algal field evaluation of the quality of habitat (e.g., biomass at a typical stream (APHA 1995), SEI methods do not include crew members but this ratio varies with taxonomy, light, determining in the field if some aspect of and nutrients. Ash free dry mass is also a stream habitat is in “good condition” or not). measure of the organic matter in samples, Patterns and trends in habitat characteristics but in contrast to chlorophyll-a, where developed from SEI data may themselves only photosynthetic algae are the source, be of interest, as well as any association

Appendix A: Field Method Overview and Justification 99 with stream biota, watershed land use, park substrate, fish cover, bank characteristics, visitation, and other factors. Furthermore, riparian vegetation structure, nonnative an objective habitat quality index may be invasive plants, and evidence of human derived by ascribing quality scoring to the disturbances. The human disturbance habitat measurements as part of the data measures include in channel, bank side, and analysis process (Kaufmann et al. 1999, indicators of human-caused stress in the Simonson et al. 1994, Meador et al. 1993). nearby floodplain. These data were obtained We feel the extra fieldwork and analyses at each of the 11 equally spaced transects required to deal with these data are worth established within the sampling reach. The the effort. As with several other aspects of second component was a thalweg profile or the SEI protocol, using these methods also a longitudinal survey of depth, habitat class, allows us to connect our work to other presence of soft/small sediment deposits, similar monitoring programs in stream and and presence of off-channel habitats at 100 rivers across Montana and the West. equally spaced intervals along the thalweg (deepest part of the channel) between the We included five major classes of physical two ends of the sampling reach. Wetted habitat characterization at GRKO (Table width was measured and substrate size A1). The first component, a channel and evaluated at 21 equally spaced cross-sections riparian characterization, included measures (at 11 regular transects, and 10 supplemental and visual estimates of channel dimensions, cross-sections spaced midway between

Table A1. Categories of ROMN SEI physical habitat measures.

Component Description Up to 11 cross-section transects (with an additional 20 for substrate size) placed at equal intervals along reach length are used to measure:

• Channel cross-section dimensions, bank height, bank undercut distance, bank angle, slope and compass bearing (backsight), and riparian canopy density (densiometer).

Substrate, Channel, and Riparian • Substrate size class and embeddedness; areal cover class and type (e.g., woody trees) of riparian vegetation in canopy, mid layer, and ground cover; areal cover class of fish concealment features, aquatic macrophytes, and filamentous algae.

• Presence and proximity of human disturbances, presence of large trees, and presence of target invasive plants.

• Measure maximum depth, classify habitat and pool-forming features, presence of backwaters, side channels, and deposits of soft, small sediment at 10-15 equally spaced intervals between each of 11 channel cross-section transects (100 or 150 individual measurements along entire reach). Thalweg Profile • Measure wetted width and evaluate substrate size classes at 11 regular channel cross-section transects and midway between them (for 21 width measurements).

• Between each of the channel cross-sections, tally large woody debris numbers Woody Debris Tally within and above the bankfull channel according to length and diameter classes (10 separate tallies).

• After completing thalweg and transect measurements and observations, identify features causing channel constraint, estimate the percentage of Assessment of Channel Constraint, constrained channel margin for the whole reach, and estimate the ratio of Debris Tor-ents, and Major Floods bankfull/valley width. Check evidence of recent major floods and debris torrent scour or deposition.

• Instantaneous measure of discharge with each full sample event. Hydrology • Continuous USGS data on discharge at nearby gauge(s), if present.

100 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site NPS/BILLY SCHWEIGER each of these). In addition, measurements of the stream slope and compass bearing between stations were obtained, providing information necessary for calculating reach gradient, residual pool volume, and channel sinuosity. The third component was data on woody debris at each of 10 segments of stream plots located between the 11 regular transects. The fourth component, assessment of channel constraint, debris torrents, and major floods, was an overall assessment of these characteristics for the whole reach. Finally, the fifth component, stream discharge was measured at the time of each full sample event within the sample reach using a Flowtracker and taken from the USGS gauge in Deer Lodge as a time head capsule mounts for midges, are housed Figure A1. Jason Smith series of continuous data. with NPS. Nomenclature follows the (NPS, GRKO) collects Integrated Taxonomic Information System macroinvertebrates on the Clark Fork in 2009 Macroinvertebrates (ITIS). All identifications were cross-walked Stream macroinvertebrates, also known using the standardized to Operational Taxonomic Units (OTUs) methods of the SEI as benthos, include crustaceans, mollusks, as developed by the state of Montana protocol. aquatic worms, and most importantly Department of Environmental Quality (because of their dominance and ecological (Jessup et al. 2006) used to standardize function), the immature forms of aquatic identifications to a consistent level for insects such as stonefly and mayfly nymphs. some analyses and to NPS NPSpecies The term “benthic” means “bottom-living,” nomenclature. as these organisms usually inhabit stream bottoms for at least part of their life cycle. Macroinvertebrates are among the most widely used organisms for bioassessment ROMN SEI benthos sample collection and because they can be sampled relatively processing follow well-established and efficiently and effectively (Resh and Jackson standardized EPA, USGS, and MT DEQ 1993); they are widespread in aquatic methods and are described in detail in environments (Merritt et al. 2008), there Schweiger et al. (In Review). Quantitative are a large number of species that have a benthos samples were composited from wide range of responses to environmental 2 eleven 1 ft subsamples taken at each transect impacts (Resh et al. 1995), and since they along the reach using a D-net with 500 um are relatively sedentary, they can be used mesh net (Figure A1). Each subsample was to determine the spatial extent of impacts. collected over a constant time (30 seconds). In addition, since macroinvertebrates are Samples were preserved in 95% alcohol relatively long-lived, community response (ETOH) for later identification. Samples can integrate the high temporal variability were spread on a gridded tray or Caton-type associated with traditional physical and splitter and picked from a randomly selected chemical analyses (Rosenberg and Resh subset of grid cells until 600 organisms were 1996). removed (a search for large, rare specimens was also conducted in the whole sample). Periphyton At least 10x magnification was used to sort Periphyton includes algae, fungi, bacteria, invertebrates from debris. All specimens and protozoa associated with channel were identified to the lowest practical substrates. Periphyton can be further level or as specified in Schweiger et al. (In grouped into growth forms, either as Review). Voucher specimens, including microalgae (microscopic, appearing as

Appendix A: Field Method Overview and Justification 101 pigmented accumulations or films attached valves (300 cells) along a scribed line with a to submerged surfaces, typically single- random start was then conducted. A 100x celled algae), or macroalgae (visible without scan and count of all valves present on the magnification, typically filamentous). Note entire slide for Didymosphenia geminata and that, while all periphyton algae taxa are any novel taxa not in the focal search was included in SEI monitoring, diatoms (algae conducted. All specimens were identified with hard, silica “shells”) may be more to species. Voucher slides are housed with useful as ecological indicators because they NPS. Nomenclature follows the Montana are found in abundance in most stream Diatom Database (Bahls 2004) as there is not ecosystems, have a well understood range in well developed taxonomic data in ITIS for tolerance to stressors, and are the focus of algae. Any taxon that is identified in ITIS is other similar monitoring efforts in the west crosswalked to the nomenclature used in our (Spaulding et al. 2010). analyses. Moreover, all taxa are crosswalked to NPS’s NPSpecies nomenclature although ROMN SEI periphyton sample collection many diatoms and algae are not yet included and processing follow well-established in NPSpecies. and standardized EPA, USGS, and MT DEQ methods and are described in detail Periphyton has important functions in in Schweiger et al. (In Review). Periphyton aquatic habitats as producers of organic samples were composited from eleven matter and plays a vital role in inorganic subsamples taken at each transect along the nutrient retention, transfer, and cycling reach. The specific method used depended (Stevenson and Smol 2003). Periphyton on the dominant substrate type present are useful indicators of environmental at the chosen microhabitat. For erosional condition because they respond rapidly and habitat, a piece of cobble within 50 cm of the are sensitive to a number of anthropogenic surface was randomly chosen and a small disturbances, including habitat destruction, known area of the “sunny side” was scraped contamination by nutrients, metals, of all benthic algae. For depositional habitat herbicides, hydrocarbons, and acidification a small area of organic and mineral fines (e.g., Hill et al. 2003). In streams where was vacuumed up with a syringe. A known flow and substratum characteristics create volume from this composite was preserved efficient interactions between water with M-fixative (Lugols and dilute formalin) and the benthic periphyton assemblage, for later identification. The composite field benthic algae typically reflect recent water samples were well mixed in the lab, sub- chemistry (Lowe and Pan 1996). Periphyton sampled, cleaned with nitric acid digestion, assemblage composition is strongly and mounted on four slides using Naphrax. influenced by land-water interactions, and A Palmer-Maloney counting chamber also by stream size and the level of human count of 300 soft-bodied algae cells at 400x disturbance. and a proportional count of 600 diatom

102 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Appendix B: Macroinvertebrate and Diatom Bioassessment Metrics

The following present summaries of the EPT (excluding Hydropsychidae and bioassessment metrics used to summarize Baetidae), percent Chironomidae, percent macroinvertebrate and diatom assemblage Crustacea and Mollusca, Shredder Taxa samples collected at GRKO. Richness, and percent Predator. Note that all macroinvertebrate tolerances used in Macroinvertebrate Multimetric the model were derived from a suite of Index of Biotic Integrity references and are specific for Montana Multimetric or Indices of Biotic Integrity (Plafkin et al. 1989, Bukantis 1998, Relyea et (MMI; Karr and Chu 1997) models are al. 2000, Brandt 2001, Merritt et al. 2008, W. based on the presence/absence or sometimes Bollman pers. comm., 2011). We do present relative abundance of taxa in a sample. Often and interpret the five component metrics taxa presence/absence or abundances are in the MMI as they often help interpret the weighted based on autecological attributes more general index. (i.e., the tolerance of a taxon to a specific environmental condition such as sediment River Invertebrate Prediction and or nutrient concentration). Individual Classification System metrics (i.e., the proportion of a community The River Invertebrate Prediction and that consists of taxa that are tolerant of Classification system (RIVPACS) (Hawkins high metal concentrations) that individually et al. 2000) was used to predict the Expected respond to measures of stress in a stream (E) macroinvertebrate taxa at the GRKO are combined into a synthetic composite SEI site. RIVPACS models predict the index. Metrics that describe characteristics specific taxa that should occur at a site given of biota that change in predictable or its natural (i.e., reference) environmental interpretable ways with anthropogenic characteristics. Specifically, these models stress (stream physiochemistry, habitat describe how probabilities of capture of conditions, landscape composition, etc.) are all taxa of interest vary across naturally ideal and selected for inclusion in the final occurring environmental gradients. model (Barbour et al. 1999). MMI models Expected taxa are those specific taxa are empirical in that they are built from and that should occur in a sample collected calibrated by streams that are known to be in from a site with specific environmental reference or degraded condition. features assuming that site was in reference condition. By sampling the actual or Jessup et al. (2006) develop MMI models observed (O) assemblage at the site, the for macroinvertebrates within the state simple ratio of O and E estimates the of Montana. As of 2012 MT DEQ was taxonomic completeness of the assemblage. updating and evaluating these models and This is a fundamental aspect of biological the ROMN will use new versions of these integrity. The O:E ratio is both a site-specific tools as available in future assessments. Note and standardized index with values that that even if MT DEQ elects to discontinue theoretically range from 0 to 1. Values of 1 use of a MMI, NPS may still use the model. imply reference conditions and values <1 The MT DEQ MMI models discriminate imply biological impairment. RIVPACS well between reference and degraded sites, models are calibrated only with reference are ecologically meaningful (mechanisms site data, and the accuracy and precision of of responses can be explained), and RIVPACS assessments depend solely on how represent diverse types of information well models predict the taxa expected under (multiple metric categories). GRKO reference conditions. If reference sites are falls in the Middle Rockies ecoregion not in natural condition, then models predict where the MT DEQ applies the “Low the biota that should occur given the quality Valley” version of the MMI. This model of the sites used in modeling. includes five component metrics: percent

Appendix B: Macroinvertebrate and Diatom Bioassessment Metrics 103 Jessup et al. (2006) develop a RIVPACS variety of stream types (Karr 1998). Karr is models for macroinvertebrates within the considered the seminal author on the MMI state of Montana. MT DEQ (2012) presents approach and the B-IBI model has received an updated version of the model models much review and use. The theoretical scale and the ROMN will use these tools in future of the index is 0 to 100 with higher values assessments. The Montana RIVPACS model indicating more impaired or disturbed is a state scale model; however, a single communities. Its assessment points are model (Hawkins et al. 2000) can be used simple percentiles (that decrease with for an entire state if the effect of natural increasing disturbance) and have no formal gradients can be adequately modeled, as meaning within MT DEQ. it was in Montana (Jessup et al. 2006). We also include a Bray-Curtis measure of the The Hilsenhoff Biotic Index (HBI) is distance between O and E for each sample. another regionally scaled metric that has This is an alternative way to measure the demonstrated a broad degree of applicability compositional dissimilarity between an in bioassessments. It was originally intended observed and expected assemblage and as an assay of low dissolved oxygen caused may perform better as it can include low- by organic loading (Hilsenhoff 1987) probability taxa without reducing the power but it may be sensitive to the effects of to detect non-reference conditions (Van impoundment, thermal pollution, and some Sickle 2008, MT DEQ 2012d). types of chemical pollution (Hilsenhoff 1998, Hooper 1993). It is similar to many In general, the performance of the Montana other tolerance indices discussed below in RIVPACS model is comparable to or better that a tolerance values are developed for taxa than most RIVPACS models in use in the and then applied as weighted sums based U.S. and elsewhere in terms of model on occurrence within a sample. MT DEQ precision (Hawkins 2006). Good models adjusted HBI tolerance scores for taxa in typically have O:E standard deviations Montana and the HBI actually appears as a less than 0.18 and the Montana model constituent metric in both the current and was 0.17. The model accounted for 76% historic mountain MMI (Bukantis 1998, of the variation in O and the slope of the Jessup et al. 2006) used in Glacier National relationship was not different from 1. In Park. However, in Montana’s montane general, O:E values effectively discriminated foothills (where GRKO occurs), the HBI was the stressed sites from the reference sites, demonstrated to be significantly associated especially for the upland streams in western with conductivity, pH, water temperature, Montana and the valley streams of the sediment deposition, and the presence Middle Rockies (where GRKO falls). of filamentous algae (Bollman 1998). The theoretical scale of the index is 0 to 10 with higher values indicating communities more Other Macroinvertebrate Metrics tolerant of organic pollution. Assessment There are many “classic” metrics used in points in the HBI metric (which increase monitoring by several agencies over the with organic pollution) are largely derived last decade or two. These include other from professional judgment and have no variants on the current MMI constructed formal meaning with MT DEQ. following similar methods as the current MMI, but with different scoring criteria The Fine Sediment Biotic Index (FSBI; and component metrics. Given GRKOs Relyea et al. 2000) is based on tolerance intermediate elevation and setting in a scores for macroinvertebrates in Montana. mountainous valley floor we use the Valley/ FSBI increases with sediment intolerance Foothill model from Bollman (1998). (higher scores will contain more taxa that are intolerant to increased sediment or The Karr Benthic Index of Biotic Integrity fines), with assessment point levels set by (B-IBI) is a MMI developed at a regional professional judgment (Relyea et al. 2000). scale that has shown a surprising degree of The theoretical scale of the index is 0 to 10 applicability to assessments across a broad with higher values indicating communities

104 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site more tolerant of fine sediment. MT DEQ demonstrate a meaningful, measurable does not use the FSBI in any formal way. and significant response to sediment Moreover, assessing differences between (in the Middle Rockies or nutrients and natural levels of bed load sediment and sediment (in the Northern Great Plains) anthropogenically increased levels may be were identified on a specific list based on difficult given fine scale spatial structure in stream groups (derived using the MT DEQ stream habitat (Relyea et al. 2000). However, fisheries classification and level III and level Bollman and Bowman (2007) suggest the IV ecoregions following the same structure index has promise as a screening filter for than the MT DEQ nutrient assessment characterizing site sediment impairment. methodology), Discriminant analysis was then used to evaluate the ability of the total The Metal Tolerance Index (MTI) relative abundance (RA) of taxa on the metric quantifies changes in community Increaser Taxa list to discriminate between composition attributable to metals pollution. impaired and non-impaired streams, and The format and calculation is similar to to provide a probability of impairment for the HBI (see above), with tolerance values a given RA. A single criterion set by MT assigned to each taxon based on sensitivity to DEQ at ≥51% suggests the probability of metals rather than organics. The theoretical impairment for sediment in the river (MT scale of the index is 0 to 10 with higher DEQ 2011b). values indicating communities more tolerant of metals pollution. McGuire (2010) suggests Other Diatom Increaser Models that MTI values for communities dominated Increaser models in Teply and Bahls (2005, by species intolerant of metals are less than 2006) were created using a similar process 4 while values for communities composed as summarized above. They were used by of only the most metals-tolerant species MT DEQ until 2010. Teply and Bahls (2005) approach 10. Metals tolerance values for develop nutrient increaser models at a fairly most taxa were developed from (Ingman and coarse scale (mountains and plains). The Kerr 1989) and (McGuire 1987 and 1989). model has a lower accuracy (~50%) than Finally, the Temperature Index (Brandt 2001) the current sediment model. We use Teply developed for Idaho streams (but applicable (2010a) for sediment and Teply and Bahls to northwestern streams in general; (2005, 2006) mountain models for metals Bollman and Bowman 2007) was around and nutrient increasers. The state does 16 (equivalent to a stream temperature not use the older models but there use is of 16ºC). Scores (really a mean stream encouraged it if done by qualified ecologists temperature optima for taxa in a sample) and in concert with other monitoring data increase with warmer water, with lower as they offer an additional way to interpret scores characterizing sites with a higher the diatom assemblages. They also connect relative abundance of stenothermic (or cold SEI results to previous monitoring around tolerant taxa). GRKO.

Diatom Increaser Models Other Diatom Metrics The State of Montana’s current approach Bahls (1993) developed a suite of diatom to bioassessment using diatoms is based metrics for MT DEQ based on common on stressor-specific increaser diatom taxa, community level measures, species as described in Teply (2010a, 2010b). autecology or tolerance scores of groups of Metrics derived from these taxa are used diatom species. Diatom species tolerances as a diagnostic tool (along with others) in used in these metrics follow Bahls (2004) stressor specific assessment. Briefly, from and are specific to Montana. Metrics Teply (2010b) increaser taxa were identified included Shannon diversity, a Pollution from a large sample of sites across the state Index (a composite numeric expression with known impairments. Taxa that, as a of the pollution tolerances assigned by group, exist in detectable amounts and Lang-Bertalot (1979) to common diatom

Appendix B: Macroinvertebrate and Diatom Bioassessment Metrics 105 species), a Siltation Index (sum of the integrity of streams within plains and percent abundances of species in the mountain bioregions. Assessment points genera Navicula, Nitxschia, and Surirella based on the distribution of metric values predominantly motile taxa that are able to measured in least-impaired reference maintain their positions on the substrate streams (Bahls et al. 1992) and metric values surface in depositional environments), and measured in streams that are known to be a Disturbance Index (percent abundance impaired by various sources and causes of of Achnanthes minutissima, which pollution (Bahls 1993). As with the non- resists chemical, physical and biological current increaser metrics, we feel these disturbances in the form of metals toxicity, metrics are useful in that they characterize substrate scour by high flows and fast basic aspects of the diatom community and currents). The metrics were intended build upon the long-standing history of to allow an assessment of the biological algae-based bioassessment.

106 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Appendix C: Taxa Lists

The following tables list taxa collected at GRKO SEI sample events in 2008 and 2009.

Macroinvertebrates Table C1 presents the list of macroinvertebrate taxa from 2008-2009 sampling.

Table C1. Macroinvertebrate taxa list and relative abundances from GRKO SEI sample events in 2008 and 2009. Counts are followed by the relative abundances (in parentheses) within each sample. Bold values indicate dominant (>10% relative abundance) within a sample. Benthos taxa were identified by multiple taxonomist so we use Operational Taxonomic Units (OTUs) to collapse identifications to comparable values. A (blank) family indicates the level of identification was above family.

Level of Sample Date Family Taxa (OTU) Total Count ID 10/6/2008 8/13/2009 Ameletidae Ameletus Genus 1 (0.17) 1 Athericidae Atherix Genus 2 (0.33) 2 Baetidae Acentrella Genus 1 (0.15) 14 (2.32) 15 Baetis Genus 27 (4.47) 27 Baetis tricaudatus Species 5 (0.74) 14 (2.32) 19 Plauditus Genus 3 (0.44) 10 (1.66) 13 Ceratopogonidae Ceratopogoninae Family 9 (1.33) 9 Mallochohelea Family 3 (0.5) 3 Chironomidae Cardiocladius Genus 2 (0.3) 3 (0.25) 5 Cricotopus Genus 5 (0.74) 51 (4.23) 56 Cryptochironomus Genus 3 (0.44) 3 Diamesa Genus 7 (1.04) 7 Eukiefferiella Genus 24 (1.78) 19 (3.15) 43 Heterotrissocladius Genus 1 (0.17) 1 Micropsectra Genus 18 (2.67) 5 (0.83) 23 Microtendipes Genus 7 (1.04) 3 (0.5) 10 Odontomesa Genus 1 (0.17) 1 Orthocladius Genus 175 (12.96) 1 (0.17) 176 Pagastia Genus 1 (0.17) 1 Parametriocnemus Genus 3 (0.22) 13 (2.15) 16 Phaenopsectra Genus 1 (0.15) 24 (3.98) 25 Polypedilum Genus 32 (5.3) 32 Tanytarsus Genus 1 (0.15) 1 (0.17) 2 Thienemanniella Genus 1 (0.15) 2 (0.33) 3 Thienemannimyia Gr. Group 4 (0.59) 4 Tvetenia Genus 4 (0.3) 7 (0.58) 11 Elmidae Elmidae Family 9 (1.33) 6 (0.99) 15 Heterlimnius Genus 20 (1.48) 20

Elmidae Elmidae Family Optioservus Genus 14 (2.32) 14 Optioservus quadrimaculatus Genus 2 (0.33) 2 Zaitzevia Genus 21 (3.11) 9 (1.49) 30 Zaitzevia parvula Genus 1 (0.17) 1 Empididae Hemerodromia Genus 2 (0.3) 2

Appendix C: Taxa Lists 107 Table C1. Macroinvertebrate taxa list and relative abundances from GRKO SEI sample events in 2008 and 2009 (continued).

Level of Sample Date Family Taxa (OTU) Total Count ID 10/6/2008 8/13/2009 Enchytraeidae Enchytraeidae Family 3 (0.44) 3 Ephemerellidae Attenella margarita Species 2 (0.33) 2 Ephemerella Genus 13 (2.15) 13 Ephemerella inermis Species 109 (16.15) 109 Ephemerellidae Family 2 (0.33) 2 Glossosomatidae Glossosoma Genus 1 (0.17) 1 Gomphidae Ophiogomphus Genus 1 (0.15) 1 (0.17) 2 Haliplidae Haliplus Genus 1 (0.15) 1 Haplotaxidae Haplotaxis Genus 1 (0.15) 1 Helicopsychidae Helicopsyche Genus 49 (7.26) 4 (0.66) 53 Heptageniidae Heptageniidae Family 1 (0.15) 2 (0.33) 3 Hyalellidae Hyalella Genus 1 (0.15) 1 Hydropsychidae Arctopsyche grandis Species 1 (0.17) 1 Cheumatopsyche Genus 7 (1.04) 4 (0.66) 11 Hydropsyche Genus 41 (6.07) 123 (20.38) 164 Hydropsychidae Family 12 (1.78) 65 (5.39) 77 Hydroptilidae Hydroptila Genus 1 (0.15) 5 (0.83) 6 Hydroptilidae Family 1 (0.17) 1 Lepidostomatidae Lepidostoma Genus 4 (0.59) 2 (0.33) 6 Leptoceridae Nectopsyche Genus 45 (6.67) 3 (0.5) 48 Oecetis Genus 9 (1.33) 7 (1.16) 16 Oecetis avara Species 3 (0.5) 3 Leptohyphidae Tricorythodes Genus 17 (2.52) 47 (7.79) 64 Naididae Nais Genus 6 (0.89) 6 Perlodidae Isogenoides Genus 1 (0.15) 1 (0.17) 2 Perlodidae Family 2 (0.33) 2 Skwala Genus 3 (0.5) 3 Planariidae Polycelis coronata Species 5 (0.74) 5 Pteronarcyidae Pteronarcella Genus 1 (0.15) 1 (0.17) 2 Simuliidae Simuliidae Family 2 (0.33) 2 Simulium Genus 6 (0.89) 22 (3.65) 28 Sphaeriidae Sphaeriidae Family 4 (0.66) 4 Tipulidae Hesperoconopa Genus 2 (0.3) 2 Hexatoma Genus 1 (0.15) 1 (0.17) 2 Tipula Genus 11 (1.63) 2 (0.33) 13 Tipulidae Family 3 (0.5) 3 Tubificidae Limnodrilus Genus 1 (0.15) 1 Naididae Family 1 (0.17) 1 Tubificidae Family 10 (1.48) 10 Misc. Acari Subclass 4 (0.59) 4 Atractides Subclass 1 (0.17) 1 Nematoda Phylum 5 (0.83) 5 Turbellaria Class 4 (0.66) 4 675 605 1280

108 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Diatoms Table C2 presents the diatom species detected in the 2008-2009 sampling.

Table C2. Diatom species list and abundance from long-term stations from GRKO SEI sample events in 2008 and 2009. Counts are followed by the relative abundances (in parentheses) within each sample. Bold values indicate dominant (>10% RA) within a sample. Nearly all diatoms were identified to the species level by a single taxonomist so we do not require OTUs. P indicates the taxon was present in a whole slide search (P is arbitrarily given a count of 0.5). Sample Date Species Total Count 10/6/2008 8/13/2009 Achnanthes lauenbergiana 4 (0.52) 4 Achnanthidium gracillimum 2 (0.26) 2 Achnanthidium minutissimum 15 (2.22) 37 (4.85) 52 Achnanthidium neomicrocephalum P (0.07) P Achnanthidium pyrenaicum 5 (0.66) 5 Amphipleura pellucida 2 (0.26) 2 Amphora copulata P (0.07) P Amphora dusenii 2 (0.3) 2 Amphora montana 2 (0.26) 2 Amphora pediculus 16 (2.37) 158 (20.72) 174 Amphora veneta 1 (0.15) P (0.07) 1.5 Aneumastus tusculus 1 (0.15) 1 Cocconeis pediculus 20 (2.97) 10 (1.31) 30 Cocconeis placentula 24 (3.56) 33 (4.33) 57 Craticula ambigua P (0.07) P Cyclotella meneghiniana 2 (0.3) 2 Cyclotella ocellata 4 (0.59) 4 Cymatopleura solea 2 (0.26) 2 Cymbella excisa 8 (1.19) 8 Denticula kuetzingii P (0.07) P Diadesmis gallica P (0.07) P Diatoma moniliformis 360 (53.37) 24 (3.15) 384 Diatoma vulgaris 1 (0.15) 2 (0.26) 3 Encyonema reichardtii 3 (0.39) 3 Encyonema silesiacum 5 (0.74) 14 (1.84) 19 Eolimna minima 2 (0.3) 15 (1.97) 17 Epithemia sorex 1 (0.15) 1 Fragilaria capucina 2 (0.3) 2 Fragilaria capucina v. mesolepta 11 (1.44) 11 Fragilaria vaucheriae P (0.07) P Geissleria acceptata 4 (0.52) 4 Geissleria decussis 2 (0.3) 2 Gomphoneis eriense 1 (0.15) 1 (0.13) 2 Gomphoneis olivaceum 31 (4.6) 31 Gomphonema brebissonii 2 (0.26) 2 Gomphonema kobayasii 8 (1.19) 8 Gomphonema olivaceum 21 (2.75) 21 Gomphonema parvulum 4 (0.59) 4 (0.52) 8 Gomphonema spp. 2 (0.26) 2 Gomphonema truncatum 1 (0.15) 4 (0.52) 5

Appendix C: Taxa Lists 109 Table C2. Diatom species list and abundance from long-term stations from GRKO SEI sample events in 2008 and 2009 (continued).

Sample Date Species Total Count 10/6/2008 8/13/2009 Mayamaea atomus 2 (0.26) 2 Meridion circulare 5 (0.74) 5 Navicula antonii 7 (0.92) 7 Navicula canalis 1 (0.15) 1 Navicula cryptotenella 5 (0.74) 14 (1.84) 19 Navicula cryptotenelloides 3 (0.39) 3 Navicula gallica P (0.07) P Navicula lanceolata P (0.07) P Navicula reichardtiana 6 (0.89) 3 (0.39) 9 Navicula tripunctata 2 (0.3) 1 (0.13) 3 Navicula veneta 1 (0.15) 1 Neidium dubium 2 (0.3) 2 Nitzschia alpinobacillum 3 (0.39) 3 Nitzschia amphibia 1 (0.13) 1 Nitzschia dissipata 10 (1.48) 82 (10.75) 92 Nitzschia fonticola 2 (0.3) 2 Nitzschia frustulum 17 (2.23) 17 Nitzschia heufleriana P (0.07) 4 (0.52) 4.5 Nitzschia inconspicua 19 (2.82) 87 (11.41) 106 Nitzschia linearis 6 (0.89) 16 (2.1) 22 Nitzschia palea 2 (0.3) 6 (0.79) 8 Nitzschia paleacea 4 (0.59) 4 Nitzschia sigmoidea P (0.07) P Nitzschia umbonata 1 (0.15) 1 Pinnularia obscura P (0.07) P Pinnularia subcapitata 2 (0.26) 2 Planothidium frequentissimum P (0.07) P Planothidium lanceolatum 2 (0.3) 5 (0.66) 7 Pseudostaurosira brevistriata 12 (1.78) 41 (5.38) 53 Reimeria sinuata 57 (8.45) 44 (5.77) 101 Reimeria uniseriata 8 (1.05) 8 Rhoicosphenia abbreviata 5 (0.74) 13 (1.7) 18 Sellaphora pupula 2 (0.26) 2 Sellaphora seminulum 4 (0.52) 4 Staurosira venter P (0.07) P Staurosirella leptostauron 1 (0.15) P (0.07) 1.5 Staurosirella pinnata 2 (0.3) 10 (1.31) 12 Staurosirella spp. 8 (1.19) 8 Surirella angusta 2 (0.3) 2 (0.26) 4 Surirella minuta 1 (0.15) 9 (1.18) 10 Synedra acus 1 (0.15) 1 Synedra ulna 3 (0.44) 8 (1.05) 11 Synedra ulna v. contracta P (0.07) P Tabularia fasciculata 2 (0.26) 2 674.5 762.5 1437

110 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Appendix D: Quality Assurance and Quality Control Overview

The ROMN SEI protocol and associated standardized field training session. Field SOPs (Schweiger et al. In Review) including trainers were experienced NPS ROMN the ROMN Quality Assurance Performance staff. Each training session was 1-2 days, and Plan (QAPP) specify various quality included lectures, field demonstrations, and assurance and quality control (QAQC) at least one practice field exercise. The field procedures and should be consulted for a operations manual served as the basis for the complete treatment of this important topic. field training program. Field crews often had Table D1 lists the various steps associated experienced ROMN staff or collaborators with the protocol to ensure data quality and as core members. Field crews were offered assurance throughout the data lifecycle. All an opportunity at the end of each year to collection and measurement procedures suggest improvements to the field operations were standardized among all field crews. manual and other aspects of field operations. Field protocols and other activities were Systematic errors were minimized by using documented in peer reviewed SOPs. All validated methodologies and standardized field crew personnel participated in a procedures.

Table D1. ROMN SEI quality control and quality assurance measures.

Core QA and QC Measure Description of the Action

Documentation Documentation via protocol, SOPs, methodology, etc. readily accessible

Training Pre-season training of field crew Datasheets with standardized codes, precision guidelines for measured values, and Field datasheets clearly documented methodology to complete the form

Crew leader/protocol lead oversight during the field season and or periodically during Field oversight data collection

End of the day or end of week quality assurance review of field datasheets for complete- Field datasheet review ness, recording errors, and legible handwriting.

Data entry quality control measures, including electronic data collection devices, to Data entry quality control eliminate misspellings, to standardize entry, and reduce errors (drop down menus, value range, limited lists, check boxes, and species lists)

Quality assurance review following SOP DM Quality Assurance and Quality Control Quality assurance post-data entry Queries are applied to the data post-data entry to isolate errors. All tests and procedures should be documented in SOP SEI: Quality Control and Quality Assurance.

Quality assurance Quality assurance tests for derived data prior to uploading to the database. All tests and post-analyses procedures should be documented in SOP SEI: Quality Control and Quality Assurance. Data collected with electronic equipment must undergo tests using QA software, queries Quality assurance queries for in MS Access, or graphical review in MS Excel for outliers and missing data. Criteria data recorded by a digital must be developed for accepting or omitting data. All tests and procedures should be device documented in SOP SEI: Quality Control and Quality Assurance

Protocol methodology establishes quality control pertaining to the sample collection. Lab data quality control ROMN assumes contract labs maintain a level of quality control that is tractable internally.

Chain of custody forms are crosswalked with lab sample results for completeness. Data Lab data quality assurance results are reviewed by the Protocol lead for outliers and missing data prior to uploading to the database.

At the end of the field season, the species list, with additions from the current year, Species lists are crosswalked and updated with code and names as listed in NPSpecies, ITIS, CNHP, Weber, and USDA Plants. Database is updated prior to data entry.

Appendix D: Quality Assurance and Quality Control Overview 111 The ROMN QAPP presents details on forms were randomly selected and all data Measurement Quality Objectives intended was confirmed. Any systematic errors were to demonstrate how ROMN monitoring evaluated to assess if there was a persistent generates data of known and documented error with a need for a global correction. quality, resulting in complete, accurate and Data were reviewed and validated to be transferable information. Data credibility sufficiently representative, accurate, precise, necessary for the intended uses will be and complete. achieved when it is: Laboratory Quality Assurance and • consistent over time and consistent Quality Control between staff members Quality Assurance and Control (QAQC) • collected and analyzed using performed by ROMN contract labs involves standardized and acceptable techniques specific tasks undertaken to determine the reliability of field and laboratory results. • comparable to data collected in other It is accomplished internally by routinely assessments using the same methods analyzing blanks, duplicates, and spikes in the day-to-day operation of a laboratory, or • used appropriately to make decisions externally by incorporating field-originated based on sound statistics blanks, duplicates, and spikes into the set of the samples collected within the SEI The ROMN QAPP also provides a summary protocol. The general objectives of lab of Data Quality Objectives or quantitative QAQC are to ensure that: (1) the integrity and qualitative terms that describe how of data generated in a monitoring effort is good data need to be in order to meet not compromised by extraneous sources project objectives. Data Quality Objectives of contamination (blanks), (2) the reported for measurement data (or data quality data are favorably comparable to the true indicators) are discussed in detail below but values (accuracy), and (3) the results of a are listed by NPS WRD as: sample collection and measurement process • Target population (data generated by analytical procedures) are reproducible (precision). • Sensitivity Rigorous QAQC procedures were also • Representativeness applied to periphyton and macroinvertebrate samples. Diatom and macroinvertebrate • Completeness data accuracy and precision were assured • Data comparability by a blind, quantitative quality-assurance approach, with a minimum of 10% of • Measurement sensitivity and detection samples randomly selected for analysis. limits Bray-Curtis similarity between counts and identification performed independently by • Measurement precision as repeatability two analysts on QAQC samples was required (accuracy) to exceed 95% (Bray and Curtis 1957). If the • Measurement systematic error/bias required accuracy was not met, the samples were reanalyzed with lessons learned from All of the various data components of the this applied to all samples. SEI protocol were managed at the ROMN offices. Completed field forms were reviewed Table D2 presents laboratory detection limits and entered into a database with several for GRKO chemistry data collected between internal QAQC checks. At least 10% of the 2007 and 2009.

112 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Table D2. Water chemistry detection limits for GRKO 2008-2010 data.

Standard Constituent Units Valid N Mean Median Minimum Maximum Std.Dev. Error

Alkalinity Carbonate as CaCO3 Total mg CaCO3/L 1 0.8 0.8 0.8 0.8 Aluminum Dissolved µg/L 8 532.5 100.0 100.0 3560.0 1223.29 432.50 Aluminum Total µg/L 7 100.0 100.0 100.0 100.0 0.00 0.00 Arsenic Dissolved µg/L 7 7.4 10.0 4.0 10.0 3.21 1.21 Arsenic Total µg/L 7 7.4 10.0 4.0 10.0 3.21 1.21 Barium Dissolved µg/L 3 4.0 4.0 4.0 4.0 0.00 0.00 Barium Total µg/L 3 4.0 4.0 4.0 4.0 0.00 0.00 Beryllium Dissolved µg/L 3 1.0 1.0 1.0 1.0 0.00 0.00 Beryllium Total µg/L 3 1.0 1.0 1.0 1.0 0.00 0.00 Cadmium Dissolved µg/L 3 0.2 0.2 0.2 0.2 0.00 0.00 Cadmium Total µg/L 3 0.2 0.2 0.2 0.2 0.00 0.00 Calcium Dissolved mg/L 7 0.0 0.0 0.0 0.0 0.00 0.00 Calcium Total mg/l 0 Carbon organic Dissolved mg C/L 8 0.1 0.1 0.1 0.2 0.06 0.02 Carbon organic Suspended mg/L 0 Carbon organic Total mg C/L 4 0.5 0.5 0.5 0.5 0.01 0.00 Chloride Dissolved mg/L 8 0.0 0.0 0.0 0.3 0.08 0.03 Chromium Dissolved µg/L 3 2.0 2.0 2.0 2.0 0.00 0.00 Chromium Total µg/L 3 2.0 2.0 2.0 2.0 0.00 0.00 Color True Total sµ 0 Conductivity Total µmhos/cm 1 0.5 0.5 0.5 0.5 Copper Dissolved µg/L 7 7.1 5.0 5.0 10.0 2.67 1.01 Copper Total µg/L 7 7.1 5.0 5.0 10.0 2.67 1.01 Fluoride Dissolved mg/L 3 0.2 0.2 0.2 0.2 0.00 0.00

Hardness carbonate Dissolved mg CaCO3/L 7 0.6 0.1 0.1 1.3 0.64 0.24 Iron Dissolved µg/L 7 100.0 100.0 100.0 100.0 0.00 0.00 Iron Total µg/L 7 100.0 100.0 100.0 100.0 0.00 0.00 Lead Dissolved µg/L 7 3.9 6.0 1.0 6.0 2.67 1.01 Lead Total µg/L 7 3.9 6.0 1.0 6.0 2.67 1.01 Magnesium Dissolved mg/L 7 0.0 0.0 0.0 0.0 0.00 0.00 Manganese Dissolved µg/L 7 1.2 1.2 1.2 1.2 0.00 0.00 Manganese Total µg/L 7 6.6 10.0 2.0 10.0 4.28 1.62 Nickel Dissolved µg/L 3 2.0 2.0 2.0 2.0 0.00 0.00 Nickel Total µg/L 3 2.0 2.0 2.0 2.0 0.00 0.00

Nitrogen Ammonium NH4 as N Dissolved mg N/L 8 0.0 0.0 0.0 0.0 0.00 0.00 Nitrogen Dissolved mg N/L 8 0.0 0.0 0.0 0.0 0.00 0.00 Nitrogen inorganic Dissolved mg N/L 7 0.0 0.0 0.0 0.0 0.00 0.00

Nitrogen Nitrate NO3 as N Dissolved mg N/L 7 0.0 0.0 0.0 0.0 0.00 0.00

Nitrogen Nitrite NO2 as N Dissolved mg N/L 7 0.0 0.0 0.0 0.0 0.00 0.00

Nitrogen Nitrite NO2 Nitrate NO3 as N Dissolved mg N/L Nitrogen organic Dissolved mg N/L 7 0.0 0.0 0.0 0.0 0.00 0.00 Nitrogen Suspended mg N/L 7 0.0 0.0 0.0 0.0 0.00 0.00 Nitrogen Total mg N/L 8 0.0 0.0 0.0 0.0 0.00 0.00 pH Total sµ 0 Phosphorus Dissolved mg P/L 8 0.0 0.0 0.0 0.0 0.00 0.00 Phosphorus organic as P Dissolved mg P/L 7 0.0 0.0 0.0 0.0 0.00 0.00

Appendix D: Quality Assurance and Quality Control Overview 113 Table D2. Water chemistry detection limits for GRKO 2008-2010 data (continued).

Standard Constituent Units Valid N Mean Median Minimum Maximum Std.Dev. Error Phosphorus orthophosphate as P Dissolved mg/L 8 0.0 0.0 0.0 0.0 0.00 0.00

Phosphorus phosphate PO4 as P Dissolved mg P/L 7 0.0 0.0 0.0 0.0 0.00 0.00 Phosphorus Suspended mg P/L 7 0.0 0.0 0.0 0.0 0.00 0.00 Phosphorus Total mg P/L 8 0.0 0.0 0.0 0.0 0.00 0.00 Potassium Dissolved mg/L 7 0.0 0.0 0.0 0.0 0.00 0.00 Potassium Total mg/L 0 Selenium Dissolved µg/L 7 6.1 10.0 1.0 10.0 4.81 1.82 Selenium Total µg/L 7 6.1 10.0 1.0 10.0 4.81 1.82 Silicon as Si Dissolved mg Si/L 7 0.0 0.0 0.0 0.0 0.00 0.00

Silicon as SiO2 Dissolved mg Si/l 8 0.0 0.0 0.0 0.2 0.08 0.03

Silicon as SiO2 Total mg SiO2/L 3 0.2 0.2 0.2 0.2 0.00 0.00 Silver Dissolved µg/L 3 0.5 0.5 0.5 0.5 0.00 0.00 Silver Total µg/L 3 0.5 0.5 0.5 0.5 0.00 0.00 Sodium Dissolved mg/L 7 0.0 0.0 0.0 0.0 0.00 0.00 Sodium Total mg/L 0 Solids Total Suspended TSS Suspended mg/l 4 3.1 4.0 0.5 4.0 1.78 0.89

Sulfur sulfate SO4 as SO4 Dissolved mg SO4/L 8 0.0 0.0 0.0 0.1 0.04 0.01 Turbidity Total NTU 1 0.3 0.3 0.3 0.3 Zinc Dissolved µg/L 3 40.0 40.0 40.0 40.0 0.00 0.00 Zinc Total µg/L 3 40.0 40.0 40.0 40.0 0.00 0.00

114 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Appendix E: Glossary

AFDW: Ash free dry weight. The weight oxygen uptake by microorganisms in a water (mass) of a material obtained by first drying sample, at 20°C, in the dark, over a five-day the material at 105ºC and then heating the period. material to 500ºC for 1 hour. Chlorophyll-a: The major green pigment Algae: Aquatic organisms that found in the chloroplasts of plants and algae. photosynthesize but lack a vascular system; some are microscopic, others very large. Criteria: A threshold set by well-established regulatory process; may or may not be Aquatic Macroinvertebrate or Benthic ecologically or management relevant for a Macroinvertebrate: An organism found park resource. in waterbodies, frequently associated with stream bottoms, not having a spinal column Density: Quantity of a number per unit area, and which is visible with the naked eye. volume, or mass.

Assessment Point/Reference Value: A state Diatom: Any one of a number of or range of states used to assess or evaluate microscopic algae, which can live as single monitoring data. When this is numeric we cells or in colonies, that are enclosed within refer to it as a reference value. They may two box-like parts or valves (called frustules) be narrative only, but these are less useful made of silica that fit together like the halves in long-term monitoring. They may be an of a Petri dish. ecological threshold, a management-based Desired Future Condition (DFC): threshold, a period of record baseline, or if Describes a desired reference condition there is nothing else available, an arbitrary from a management perspective. Usually value (which is used through time for expressed in a narrative form that makes comparative purposes only). application to monitoring data less direct. Baseflow: The portion of streamflow that The NPS defines the DFC a “park’s natural comes from groundwater and not runoff. and cultural resource conditions that the NPS aspires to achieve and maintain over Baseline: A value or range of values time, and the conditions necessary for calculated from a time series of data visitors to understand, enjoy, and appreciate (from one to many years; if many, the time those resources.” period is often referred to as a “Period of Record”). Baselines may not necessarily be Diel: Involving a 24-hour period that usually an ecological or management threshold, includes a day and the adjoining night. but could be used in a similar way to assess Ecological Threshold: In general, this is monitoring data. a break point(s) or a “break range” that Beneficial Use: A valuable characteristic of describes a shift in an ecological response a stream or river resource that, directly or measure and bounds regions in a response’s indirectly, contributes to human welfare. distribution. Thresholds often describe meaningful changes in the ecology of a Benthic: On or associated with the system. They may involve a non-linear, rapid, sediments or bottom of a body of water. or large response relative to the inputs to a system. They often create a new state from Biomass: The total mass or amount of living which is difficult to recover. or dead organisms in a particular area or volume. Ephemeral Stream: A stream or stream segment that flows only in direct response to Biochemical Oxygen Demand (BOD): precipitation in the immediate watershed or A measure of, as well as a procedure for in response to the melting of a cover of snow determining, how fast oxygen is used up in and ice and whose channel bottom is always water. BOD is usually measured as the rate of above the local water table.

Appendix E: Glossary 115 Eutrophication: The process of enrichment Management-based Threshold: An of a waterbody by nutrients, usually assessment point with explicit management nitrogen- and phosphorus-containing applications or where management-relevant compounds, and the resulting increase changes in a response occurs. Management in primary productivity (algal and plant thresholds may be the same as ecological growth and decay). Some definitions include thresholds, especially in wilderness parks organic enrichment of a waterbody as part of or when there is no or little information eutrophication. about the management relevance of a response (in these cases we generally refer Fixation (Nitrogen Fixation): The process to an ecological threshold). They may be by which nitrogen is taken from its relatively set at some estimated distance from an inert gas form in the atmosphere (N ) and 2 ecological threshold to serve as a warning converted into nitrogen compounds such as (“surveillance threshold”) or to spur nitrate and ammonia. management action (“action threshold”).

Geospatial: Pertaining to the geographic Metric: A characteristic of a biological location and characteristics of natural or assemblage (e.g., fishes, algae) that changes constructed features and boundaries on, in some predictable way with increased above, or below the earth’’ surface; especially human influence. referring to data that are geographic and spatial in nature. Minimally Disturbed Condition (MDC): A reference condition characterized by the Intermittent Stream: A stream or stream absence of significant human disturbance. segment that is below the local water table In practice, even in wilderness parks, there for at least some part of the year and obtains is usually some background human impact its flow from both surface run-off and due to diffuse or indirect factors like climate groundwater discharge. change, nutrient deposition, or residual historic alterations. Least Disturbed Condition (LDC): In some (if not many) cases, the lack of availability Narrative Water Quality Criteria: of a minimally disturbed condition (MDC) Statements codified in state law that or a completely undisturbed reference site describe, in a concise way, a water quality makes it necessary to resort to the concept of condition that must be maintained in order “least-disturbed condition.” Least-disturbed to protect beneficial uses. condition is defined as the best available state within a defined region. The LDC may Nonpoint Source: The source of pollutants include measurable anthropogenic stress which originates from diffuse runoff, and disturbance. The concept of a LDC seepage, drainage, or infiltration. implies that there may be an opportunity Numeric Water Quality Criteria: for management or restoration of sites Quantified expressions of water quality in to a MDC or best attainable state. This state law intended to protect a designated may also be possible with a MDC, but in beneficial use or uses. many cases the set of actions (i.e., reverse climate change) are beyond the scope of Organic Enrichment: From a water management. pollution perspective, the addition of decomposable plant or animal material, or Macrophyte: Macroscopic aquatic vascular their wastes, to a waterbody. plants capable of achieving their generative cycles with all or most of the vegetative parts Perennial Stream: A stream or stream submerged or supported by the water. segment that has flowing water year-round except during extreme drought. Mainstem: The principal river within a given drainage basin, in the case where a Periphyton: The microscopic flora and number of tributaries discharge into a larger fauna that grow or are associated with the watercourse. bottom of a body of water and includes microscopic algae, bacteria, and fungi.

116 Stream Ecological Integrity at Grant-Kohrs Ranch National Historic Site Phytoplankton: Free living, generally Salmonids: Ray-finned fish, whose members microscopic algae commonly found floating include salmon, trout, chars, freshwater or drifting in waterbodies such as the ocean, whitefishes, and graylings. lakes, and streams. Saturation: A state in which the gas Point Source: A discernable, confined, concentration (e.g., oxygen) in a waterbody and discrete conveyance, including but not is in equilibrium with the local partial limited to any pipe, ditch, channel, tunnel, pressure of that gas in the atmosphere. conduit, well, discrete fissure, container, rolling stock, or vessel or other floating Standards (Water Quality Standards): craft, from which pollutants are or may be In a water quality regulatory context, a discharged. term applicable to state waters referring collectively to their designated beneficial Primary Productivity: The production uses, criteria, and the non-degradation of organic compounds from carbon policy, all in Montana law. dioxide, principally through the process of photosynthesis. Strahler Order: A simple hydrology algorithm used to define stream size based Reference Condition(s): The region of a on a hierarchy of tributaries. Streams at the response distribution on the “good side” of a top of the watershed are labeled 1. When (usually ecological) threshold. The reference two order-1 streams join, they create an condition is a distribution as it encompasses order-2 stream. When two order-2 streams natural variation in the response. It is join, they create an order-3 stream, and so somewhat synonymous with a Natural on. If a stream of lower order (e.g., order-2) Range of Variability. Some authors restrict joins a stream of higher order (e.g., order-3), reference condition to describe a biological the order number of the latter does not response, but the ROMN proposes using change. it more loosely to refer to any type of response measure. Reference conditions Wadeable: A stream whose Strahler order is are defined relative to a specific response, first through (at most) sixth (1:100,000 map target population and spatiotemporal scale scale) in which most of the wetted channel (i.e., this is different for a wetland response is wadeable by a person during baseflow in Rocky Mountain National Park than for conditions. the Southern Rockies ecoregion). Based on the dominant landscape condition, reference conditions may reflect various degrees of background disturbance.

Appendix E: Glossary 117

The Department of the Interior protects and manages the nation’s natural resources and cultural heritage; provides scientific and other information about those resources; and honors its special responsibilities to American Indians, Alaska Natives, and affiliated Island Communities.

NPS 451/124876, June 2014 National Park Service U.S. Department of the Interior

Natural Resource Stewardship and Science 1201 Oak Ridge Drive, Suite 150 Fort Collins, Colorado 80525 www.nature.nps.gov

EXPERIENCE YOUR AMERICA™