Assessing Forest Health in the Athabasca Oil Sands Region

Wood Buffalo Environmental Association Technical Reports

T.A. Clair & K.E. Percy (Editors)

WBEA Report # 2015-05-25

Final Report

This report should be cited as:

Clair, T.A. and K.E. Percy (Editors) 2015. Assessing forest health in the Athabasca Oil Sands Region. WBEA Technical Report. 2015-05-25, 180 pp +Appendices

Disclaimer: The contents and opinions expressed by the authors in this report do not necessarily reflect the views of the WBEA or of the WBEA membership.

Dedication

We dedicate this report to the memory of Dr. Sagar V. Krupa

Dr. Krupa attained international recognition as a visionary scientist, science editor, and mentor to many. His long-standing contribution to environmental research and monitoring in Alberta included his formative role in designing and leading WBEA multidisciplinary projects that integrated through receptor modeling the air and land systems in the Athabasca oil sands region.

iii

Acknowledgements

The editors wish to acknowledge the diligent work of Abigale Glashoerster, WBEA, in the desktop publishing and associated tasks in the preparation of this final report. Her effort is greatly appreciated.

The editors and authors of this report wish to sincerely thank the Members of the Wood Buffalo Environmental Association for their long-term and continuing support of essential terrestrial monitoring in the Regional Municipality of Wood Buffalo. Continuity of funding is a prerequisite for successful detection, quantification and interpretation of change in systems that respond to cumulative inputs.

In particular, we wish to commend the multi-stakeholder technical membership of the TEEM Science Subcommittee during the 2006-2007 period, along with the membership of the TEEM Committee for their vision in recognizing the need for significant science enhancement of the monitoring program. The 2008 step change in funding by industry enabled a distinguished team of multidisciplinary scientists to be engaged. This resulted in the first integration of air and land systems, through source to sink measurement and apportionment.

Executive Summary

This report summarizes the most recent results from the Wood Buffalo Environmental Association’s Terrestrial Environmental Effects Monitoring (TEEM) forest health monitoring program. During 1998-2007, WBEA operated the Acid Monitoring Program (AMP) focused on potential adverse effects including acidification of forest soils from sulphur and nitrogen oxides, and detrimental impacts to vegetation. In 1998, TEEM initiated measurement and sampling at a network of 11 jack pine (Pinus banksiana Lamb.) dominated, interior forest stand plots. Additional plots were added between 1999 and 2003, and another cycle of measurement and sampling of soils and vegetation occurred on 13 plots in 2004.

During 2006-2007, the Science Subcommittee (SSC) of TEEM completed a science review of the existing AMP. In 2008 based on science advice, TEEM adopted the source-to-sink approach for effects monitoring by measuring at key points along the emissions, chemical transformation, deposition, and terrestrial receptor pathway. TEEM also adopted the forest health approach to terrestrial monitoring focused on establishing/determining cause-effect relationships between air pollutants and forest ecosystem health in the region. In 2011/12, the enhanced forest health network of 25 plots was sampled.

This report comprises 12 chapters integrating monitoring history, network design, results from air/deposition monitoring, deposition modeling, above- and below-ground biological and chemical measurements made in 2011/12, as well as comparative status of some indicators measured at five plots that were sampled in 1998, 2004 and 2011/12.

The highest passively measured and modeled air concentrations of SO2, NO2, NH3, and HNO3 were reported nearer the industrial operations. Ozone (O3) concentrations, as expected, increased with distance from the sources. Three chapters reporting air quality and deposition of pollutants showed similar patterns of air concentration and spatial deposition for S and N. Two patterns were visible. The first was a west to east pattern resulting in S and N concentrations/ deposition being higher east of the mining and upgrading operations than to the west. There was also a north-south pattern in air concentrations, showing the influence of valley topography on pollutant dispersion. Air quality and deposition measurements indicated that nitrogen and sulphur concentrations/deposition amounts are enhanced within 30 km of the operations, decline with increasing distance from them, and reach background levels ~40-50 km away from main industrial emission sources.

Trace elements in vegetation including heavy metals generally follow the same spatial distribution pattern. Levels of sulphur and nitrogen in jack pine foliage at six plots increased from 1998 to 2012. Nitrogen is being taken up in vegetation, and is not accumulating in mineral soils. Sulphur in soils was correlated with modeled S+N deposition at the LFH, 0-5 and 5-15 and 15-30 cm depths. Neither soil nitrogen nor pH showed any correlation with measured deposition. Soil microflora as well as vascular cover, forb cover and shrub richness were strongly and positively

related to atmospheric deposition of base cations. There was no correlation between ecosystem variables and S and N as acidifiers because of the importance of base cations which neutralizes the acid inputs. The role of atmospheric nitrogen deposition as a nutrient has the potential to increase in relative importance.

Base cation/aluminum (BC: Al) ratios in the LFH and mineral soil layers were not correlated (P=0.94) with modeled sulphur and nitrogen deposition. The BC: Al ratio ranged from 1 to 5 at twenty of the twenty-one plots sampled. The BC:Al trigger set under the CEMA Acid Deposition Management Framework (http://cemaonline.ca/index.php/cema-recommendations/acid deposition) was not exceeded, in 2011/12.

Environmental monitoring must never remain complacent, and must always be innovative, adaptive and responsive as is demonstrated in this report. We hope that state of the art monitoring continues to be responsive, and adaptable to continuing oil sands development, and will be fully supported and adequately funded going forward.

vii

Table of Contents Acknowledgements ...... v Executive Summary ...... vi List of Figures ...... xiv List of Tables ...... xvii Chapter 1: Introduction to the report: Assessing forest health in the Athabasca Oil Sands Region Thomas Clair, Wood Buffalo Environmental Association, Fort McMurray, AB, Mervyn Davies, Stantec Consulting Ltd., Calgary, AB, Keith Puckett, ECOFIN, Waldemar, ON, and Kevin Percy, Wood Buffalo Environmental Association, Fort McMurray, AB 1.1 Industrial atmospheric emissions in the Athabasca Oil Sands Region ...... 1 1.2 History of WBEA forest research in the AOSR ...... 3 1.3 Organization of this report ...... 6 1.4 References ...... 7 Chapter 2: Air pollution and dry deposition of nitrogen and sulphur in the AOSR estimated using passive samplers Yu-Mei Hsu, Wood Buffalo Environmental Association, Fort McMurray, AB and Andrzej Bytnerowicz, US Department of Agriculture, Forest Service, Pacific Southwest Research Station, Riverside, CA, USA 2.1 Introduction ...... 8 2.2 Methods ...... 9

2.2.1 Monitoring, sampling and analysis for NH3, HNO3, SO2, NO2 and O3 ...... 9 2.2.2 Estimating dry deposition flux using the Multi-Layer Inferential Model ...... 10 2.3 Results and discussion ...... 11

2.3.1 Ammonia (NH3) ambient concentrations ...... 11

2.3.2 Nitric acid (HNO3) ambient concentrations…………………………………………………………..13

2.3.3 Nitrogen dioxide (NO2) ambient concentrations ...... 15

2.3.4 Total inorganic gaseous reactive N (Nr sum) ...... 19 2.3.5 Sulphur dioxide ambient concentrations ...... 19 2.3.6 Ozone ambient concentrations ...... 26 2.3.7 Nitrogen and sulphur deposition estimates based on passive sampler data ...... 26 2.3.7.1 Ammonia dry deposition ...... 27

2.3.7.2 HNO3 dry deposition ...... 30

2.3.7.3 NO2 deposition ...... 30

viii

2.3.7.4 Total reactive N deposition ...... 33 2.3.7.5 Sulphur dioxide deposition ...... 33 2.4 Conclusion ...... 36 2.5 References ...... 36 Chapter 3: Wet and dry deposition in the AOSR collected by ion exchange resin samplers Mark Fenn, US Department of Agriculture Forest Service, Pacific Southwest Research Station, Riverside, CA, USA 3.1 Introduction ...... 40 3.2 Methods ...... 40 3.3 Results ...... 41 3.4 Discussion ...... 46 3.5 Conclusions ...... 48 3.6 References ...... 48 Chapter 4: Predicted spatial variations of sulphur and nitrogen compound concentrations and deposition in the AOSR Mervyn Davies, Kanwardeep Bajwa and Reid Person, Stantec Consulting Ltd., Calgary, AB 4.1 Introduction ...... 51 4.2 Model approach ...... 51

4.2.1 Predicted SO2 and NO2 concentrations ...... 52 4.2.2 Predicted sulphur compound deposition ...... 53 4.2.3 Predicted nitrogen compound deposition ...... 53 4.2.4 Acidifying and cation deposition ...... 56 4.2.5 Deposition PAI ...... 58 4.3 Uncertainties ...... 60 4.4 Comments and conclusions ...... 61 4.5 References ...... 62 Chapter 5: Site selection and field methods for the Forest Health study Kenneth Foster, Owl Moon Environmental Inc., Calgary, AB 5.1 Introduction ...... 64 5.2 Site selection and growth of the jack pine monitoring program ...... 66 5.2.1 1998 ...... 66 5.2.2 2001 ...... 68 5.2.3 2004 ...... 69

5.2.4 2011/2012 ...... 69 5.3 Monitoring cycle ...... 72 5.4 Site establishment procedures ...... 72 5.4.1 Vegetation plot ...... 73 5.4.2 Off-plot trees ...... 75 5.4.3 Soil pit–soil characterization ...... 75 5.4.4 Soil plots ...... 75 5.5 Measurements and indicators ...... 76 5.5.1 Vegetation plot sampling–jack pine morphometric measurements ...... 76 5.5.2 Off-plot tree sampling ...... 76 5.5.3 Vegetation community composition ...... 78 5.5.4 Annual forest health assessment ...... 78 5.5.5 Soil sampling ...... 78 5.5.6 Soil sample analyses ...... 81 5.6 Summary and conclusions ...... 82 5.7 References ...... 82 Chapter 6: Plot tree layer description Ellen Macdonald, Department of Renewable Resources, University of Alberta, Edmonton, AB 6.1 Introduction ...... 85 6.2 Methods ...... 85 6.3 Results and discussion ...... 88 6.4 References ...... 93 Chapter 7: Soil biology in the AOSR using phospholipid fatty acid analysis Jacynthe Masse, Carolyn Churchland and Sue J. Grayston, Department of Forest and Conservation Sciences, Faculty of Forestry, University of British Columbia, Vancouver, BC 7.1 Introduction ...... 95 7.2 Methods ...... 96 7.2.1 Experimental design and laboratory analysis ...... 96 7.2.2 Statistical analysis ...... 97 7.3 Results ...... 97 7.3.1 Proximity to the industrial centre and burned sites ...... 97 7.3.2 Soil microbial community spatial structure ...... 98

7.3.3 Soil microbial community relationships with soil chemical characteristics, atmospheric deposition and vegetation ...... 100 7.4 Discussion ...... 106 7.4.1 Spatial structure of microbial communities ...... 106 7.4.2 Impacts of deposition on soil microbial communities ...... 106 7.4.3 Impacts of natural forest fire on soil microbial communities ...... 107 7.5 Conclusions ...... 107 7.6 References ...... 108 Chapter 8: Routine foliar and soil analysis of 15 years monitoring in the AOSR Doug Maynard, Natural Resources Canada, Canadian Forest Service, Victoria, BC 8.1 Introduction ...... 111 8.2 Materials and methods ...... 111 8.2.1 Sites selected ...... 111 8.2.2 Jack pine foliage sampling ...... 111 8.2.3 Soil sampling ...... 112 8.2.4 Foliar chemical analysis ...... 112 8.2.5 Soil chemical analysis ...... 112 8.2.6 Statistical analysis ...... 112 8.3 Results and discussion ...... 114 8.3.1 2011 Foliar chemistry ...... 114 8.3.2 2011 Soil chemistry ...... 120 8.3.2.1 Surface organic horizon (LFH) ...... 121 8.3.2.2 Mineral soil 0-5 cm layer...... 124 8.3.2.3 Mineral soil 5-15 and 15-30 cm layers ...... 125 8.3.2.4 Percent Base saturation (% BS) and Base cation to aluminum (BC:Al) ratio ...... 127 8.4 Summary ...... 131 8.5 References ...... 132 Chapter 9: Responses of understory vegetation to deposition from oil sand processing operations Ellen Macdonald, Department of Renewable Resources, University of Alberta, Edmonton, AB 9.1 Introduction ...... 134 9.2 Methods ...... 135 9.3 Results and discussion ...... 137

9.4 Conclusions ...... 142 9.5 References ...... 143 Chapter 10: Chemical composition and lichen community change in the AOSR Keith Puckett, ECOFIN, Waldemar, ON 10.1 Introduction ...... 145 10.2 Methods ...... 145 10.2.1 Lichen element levels ...... 145 10.2.2 Epiphytic lichen community composition ...... 146 10.2.3 Elemental concentrations in lichen ...... 147 10.2.4 Lichen community structure at sampling sites ...... 152 10.3 Discussion...... 155 10.3.1 Spatial variation in lichen element concentrations ...... 155 10.3.2 Causality - Linkage between changes in vigour, growth, community structure and air quality in the AOSR…………………………………………………………………………………………………158 10.3.3 Causality - Air quality levels known to impair AOSR lichens………………………………159 10.3.4 Causality - Comparison of measured and predicted air pollutant concentration in the AOSR and lichen response ...... 160 10.4 Summary ...... 160 10.5 References ...... 161 Chapter 11: The application of critical loads and estimates of exceedance for sulphur and nitrogen deposition to forests in the AOSR Shaun Watmough, Trent University, Peterborough, ON and Colin Whitfield, Centre for Hydrology, University of Saskatchewan, Saskatoon, SK 11.1 Critical loads and their application to forest ecosystems...... 167 11.2 Critical load calculations ...... 167 11.2.1 The simple mass balance model for acidity ...... 167 11.2.2 The simple mass balance model for nutrient N...... 168 11.3 Previous estimates of critical loads and exceedance for forests in the AOSR ...... 168 11.4 Importance of weathering rates, N dynamics, appropriate chemical limits and deposition ...... 169 11.5 Modified application of critical loads to TEEM plots ...... 170 11.6 Discussion: Implications for forest health/monitoring ...... 171 11.7 References ...... 172

xii

Chapter 12: State of the jack pine forest Thomas Clair and Kevin Percy, Wood Buffalo Environmental Association, Fort McMurray, AB 12.1 Program evolution ...... 175 12.2 Air chemistry and atmospheric deposition implications ...... 176 12.3 Soil and plant interactions with atmospheric deposition and critical loads ...... 178 12.4 Future directions ...... 180 APPENDICES Chapter 1 Appendices Appendix 1.1 Location of current continuous air monitoring sites used for this study 182 Appendix 1.2 Location of Forest Health plots and meteorological measurement equipment ...... 183 Chapter 2 Appendices

Appendix 2.1 Annual NH3 N Dry Deposition at the forest monitoring sites ...... 184

Appendix 2.2 Annual HNO3 as N Dry Deposition at the forest monitoring sites...... 185

Appendix 2.3 NO2 as N Dry Deposition ...... 186 Appendix 2.4 Total N Dry Deposition ...... 187 Chapter 4 Appendices

Appendix 4.1 Predicted annual average SO2 and NO2 concentrations at the forest health assessment sites...... 188 Appendix 4.2 Predicted annual sulphur and nitrogen compound deposition at the forest health assessment sites...... 190 Appendix 4.3 Predicted potential acid input at the forest health assessment sites...... 192 Chapter 5 Appendices Appendix 5.1 Jack Pine Morphometric Measurements – Routine Monitoring Program (Plot and Off-Plot Trees) ...... 194 Appendix 5.2 Jack pine tissue sampling ...... 196 Appendix 5.3 Soil sampling in the TEEM Jack Pine Monitoring Program ...... 199 Appendix 5.4 Soil sample laboratory procedures in the TEEM Jack Pine Monitoring Program ...... 200 Chapter 6 Appendices Appendix 6.1 Stand age and summary information on density, diameter, height, crown depth, and canopy closure for canopy and sub-canopy trees at each site...... 201

xiii

Chapter 8 Appendices Appendix 8.1 Concentrations of jack pine needles at the routine monitoring sites sampled in 2011...... 203 Chapter 9 Appendices Appendix 9.1 Results of regression analyses ...... 204 Appendix 9.2 Results of regression analyses ...... 206 Appendix 9.3 Complete list of species encountered during sampling of the sites in 2011/2012 ...... 208 Chapter 10 Appendices Appendix 10.1 Elements analysed in Hypogymnia physodes in 2011...... 209

List of Figures Figure 1. 1 Historical trend in SO2 and NOx emissions from oil sands extraction activities (from National Pollutant Release Inventory (NPRI)) ...... 2 Figure 1. 2 Location of sampling sites in the AOSR...... 5 -3 Figure 2. 1 Changes in NH3 concentrations (µg m ) at all monitoring sites from May 2005 to September 2013...... 11 Figure 2. 2 NH3 spatial variations in winter 2010, summer 2011, winter 2011 and summer 2012 ...... 12 -3 Figure 2. 3 Changes in HNO3 concentrations (µg m ) in all monitoring sites from May 2005 until September 2013...... 13 Figure 2. 4 HNO3 spatial variations in winter 2010, summer 2011, winter 2011 and summer 2012 ...... 14 Figure 2. 5 Annual fluctuations in NO2 concentrations at sites...... 15 Figure 2. 6 NO2 spatial variations in winter 2010, summer 2011, winter 2011 and summer 2012 ...... 17 Figure 2. 7 Monthly variation in passively measured NO2 concentrations from 2000 to 2013. .. 18 Figure 2. 8 Annual mean NO2 concentrations from CALPUFF and passive samplers (2009 to 2012)...... 19 Figure 2. 9 Reactive gaseous inorganic N spatial variations in winter 2010, summer 2011, winter 2011 and summer 2012 ...... 21 Figure 2. 10 Annual SO2 air concentration value ranges at monitoring sites...... 22 Figure 2. 11 SO2 spatial variations in winter 2010, summer 2011, winter 2011 and summer 2012 ...... 23 Figure 2. 12 Monthly variations in passively measured SO2 concentrations from 2000 to 2013 24 Figure 2. 13 Annual mean SO2 concentrations from CALPUFF and passive samplers (2009 to 2012)...... 25 Figure 2. 14 2013 average O3 concentration in the AOSR ...... 26 Figure 2. 15 Monthly deposition velocities of (a) NH3, (b) HNO3, (c) NO2 and (d) SO2 ...... 27

xiv

Figure 2. 16 NH3-N and HNO3-N deposition average from 2009 to 2012 in the AOSR ...... 28 Figure 2. 17 NH4 as N deposition comparison at 9 sites...... 29 Figure 2. 18 NO2-N and total N deposition average from 2009 to 2012 in the AOSR ...... 31 Figure 2. 19 NO3 as N deposition comparison at 9 sites...... 32 Figure 2. 20 2009-2012 SO2 (as S) annual average deposition ...... 33 -1 Figure 2. 21 SO4 as S deposition (kg ha ) comparison at 9 sites...... 35 - Figure 3. 1 Deposition of dissolved inorganic N and SO4 in forest clearings and in throughfall versus distance ...... 42 Figure 3. 2 Temporal patterns of summer and winter deposition of NH4-N, NO3-N and SO4- S in forest clearings and throughfall under jack pine in the Athabasca Oil Sands Region...... 43 Figure 3. 3 Dissolved inorganic nitrogen, NH4-N, and NO3-N deposition in the AOSR...... 45 Figure 3. 4 Maps of SO4-S deposition and DIN+ SO4-S and sum of base cations in the AOSR. .... 46 Figure 4. 1 Annual average SO2 concentrations in the AOSR due to LAR (CALPUFF) and non-LAR (CMAQ) emissions...... 54 Figure 4. 2 Total sulphur and nitrogen deposition in the AOSR due to LAR and non-LAR emissions...... 55 Figure 4. 3 Predicted total sulphur + nitrogen deposition in the AOSR...... 56 Figure 4. 4 Base cation deposition in the AOSR based on throughfall deposition and bulk measurements ...... 58 Figure 4. 5 Predicted lower bound PAI deposition in the AOSR based on throughfall BC measurement ...... 59 Figure 5. 1 TEEM jack pine monitoring sites – 1998 ...... 67 Figure 5. 2 TEEM jack pine monitoring sites – 2001 ...... 68 Figure 5. 3 TEEM jack pine monitoring sites – 2004 ...... 69 Figure 5. 4 TEEM jack pine monitoring sites – 2011/2012 ...... 71 Figure 5. 5 Typical jack pine monitoring plot layout ...... 73 Figure 5. 6 Schematic jack pine vegetation plot layout ...... 74 Figure 5. 7 Soil plot layouts at jack pine monitoring sites ...... 76 Figure 5. 8 Stemflow, throughfall and freefall soil sampling zones ...... 80 Figure 5. 9 Soil sampling locations (as defined in 2011/2012) ...... 81 Figure 6. 1 Regressions of diameter at 1.3m (breast) height and tree height for canopy trees. . 90 Figure 6. 2 Regressions of branch Internode length and defoliation (both for current year) versus deposition variables ...... 90 Figure 6. 3 Regressions of Stand Age (in 2011) and versus two different deposition variables .. 91 Figure 7. 1 Mean fungal abundance (nmols/g dry weight of soil) over three distances ...... 98 Figure 7. 2 Principal coordinates of Neighbour Matrices on PLFA analysis of microbial communities ...... 99 Figure 7. 3 RDA biplot - scaling 1 showing relationships between sites, abundance of microbial communities and PCNM ...... 100 Figure 7. 4 RDA biplots - scaling 2 showing relationships between abundance of microbial communities and soil chemical characteristics and atmospheric deposition...... 102 Figure 7. 5 RDA triplot - scaling 1 showing relationships between abundance of microbial communities and soil chemical characteristics and atmospheric deposition...... 103

Figure 7. 6 RDA biplots - scaling 2 showing relationships between abundance of microbial communities and soil chemical characteristics and atmospheric deposition...... 104 Figure 7. 7 RDA triplot - scaling 1 showing relationships between abundance of microbial communities and soil chemical characteristics, atmospheric deposition and vegetation coverage...... 105 Figure 8. 1 Total nitrogen and S in current jack pine foliage versus modeled N and S deposition from the CALPUFF model ...... 115 Figure 8. 2 Total nitrogen and S in current jack pine foliage versus modeled potential acid input ...... 117 Figure 8. 3 Total calcium in current jack pine foliage versus modeled base cation deposition . 118 Figure 8. 4 Extractable sulfur and total sulfur in the surface organic horizon (LFH) versus predicted nitrogen and sulfur deposition from the CALPUFF model ...... 121 Figure 8. 5 Total nitrogen in the surface organic horizon (LFH) versus predicted nitrogen and sulfur deposition from the CALPUFF model ...... 122 Figure 8. 6 pH in the surface organic horizon (LFH) versus predicted nitrogen and sulfur deposition from the CALPUFF model ...... 123 Figure 8. 7Extractable Extractable sulfate in the 0-5cm mineral soil layer versus modeled N and S deposition from the CALPUFF model ...... 124 Figure 8. 8 Soil pH in the 0-5 cm mineral layer versus modeled nitrogen and sulfur deposition from the CALPUFF model ...... 125 Figure 8. 9 Ammonium concentration in the 0-5 cm mineral layer versus modeled nitrogen and sulfur deposition from the CALPUFF model ...... 125 Figure 8. 10 Extractable sulfate in the 5-15cm mineral soil layer versus modeled N and S deposition from the CALPUFF model ...... 126 Figure 8. 11 Soil pH in the 5-15 cm mineral layer versus modeled nitrogen and sulfur deposition from the CALPUFF model ...... 126 Figure 8. 12 Extractable sulfate in the 15-30 cm mineral soil layer versus modeled N and S deposition from the CALPUFF model ...... 127 Figure 8. 13 Soil pH in the 15-30 cm mineral layer versus modeled nitrogen and sulfur deposition from the CALPUFF model ...... 127 Figure 8. 14 % Base Saturation (BS) in the LFH and 0-5cm mineral soil layer versus modeled N and S deposition form the CALPUFF model ...... 130 Figure 8. 15 Base cation to aluminum (BC:Al) ratio in the LFH and 0-5cm mineral soil layer versus modeled N and S deposition form the CALPUFF model ...... 131 Figure 9. 1 Total Richness, Cover of Vascular Plants, and Vascular Plant Richness versus deposition ...... 138 Figure 9. 2 Results of constrained ordination (Redundancy Analysis) of understory vegetation composition on atmospheric deposition variables...... 141 Figure 10. 1 Study site locations in the AOSR ...... 146 Figure 10. 2 Al, As, Cr, Co, Fe, Mo, Ni, Si, Na, Ti, and V in H. phylodes as a function of distance from main emission region...... 149 Figure 10. 3 Ca, Cu, Mg, K, Sr and S concentrations in H. physodes as a function of distance from industrial area...... 150

xvi

Figure 10. 4 Ba, Pb, N, and P concentrations in H. physodes as a function of distance from industrial area...... 151 Figure 10. 5 Cd, Mn, Hg, and Z concentrations in H. physodes as a function of distance from industrial area...... 152

List of Tables Table 2. 1 Sen-Theil test of slope significance for NO2 ...... 19 Table 2. 2 Sen-Theil test of slope significance for SO2 ...... 20 Table 2. 3 Sen-Theil test of slope significance for O3 ...... 25 Table 5. 1 Jack Pine Site Selection Criteria ...... 65 Table 5. 2 Jack Pine Monitoring Site Naming Convention as Revised in 2011 ...... 72 Table 6. 1 Deposition; and tree and branch variables that were tested. CALPUFF deposition variables ...... 87 Table 6. 2 Results of regressions analyses of tree variables on the various deposition variables...... 89 Table 7. 1 Two-way analysis of variance of broad-scale PLFA groups. Only statistically significant results and strong trends (α<0.1) were included...... 97 Table 7. 2 Species presenting a significant relationship with PCNM 3 – *** P value<0.01 ...... 100 Table 8. 1 Foliar and soil chemical measurements used in the 2011 sampling year...... 113 Table 8. 2 Range of nutrient concentrations of current jack pine foliage and comparison with other studies in Boreal Plains and Boreal Shield West ...... 114 Table 8. 3 Sulfur concentrations in current foliage of jack pine at monitoring sites established in 1998 with multi-year sampling ...... 118 Table 8. 4 Nitrogen concentrations in current foliage of jack pine at monitoring sites established in 1998 with multi-year sampling...... 119 Table 8. 5 Sulfur and nitrogen concentrations in current foliage of jack pine at monitoring sites established in 2004 and resampled in 2011...... 120 Table 8. 6 Total sulfur and nitrogen and pH of the surface organic horizon (LFH) of burned sites (burned in May 2011 prior to sampling) sampled in 2011 ...... 123 Table 8. 7 The % base saturation (%BS) in soil layers at monitoring sites established in 1998 with multi-year sampling...... 129 Table 9. 1 Results of the best regression model for each of the understory vegetation response variables...... 138 Table 9. 2 Results of constrained ordination of the understory vegetation composition ...... 140 Table 10. 1 R6 Lichen Abundance Scoring Scale...... 147 Table 10. 2 List of lichen impact studies in the Athabasca Oil Sands Region...... 152 Table 11. 1 Base cation weathering rates, critical loads for acidity and N, S, N and BC deposition, and estimated exceedance for 10 TEEM plots...... 171

xvii

Chapter 1: Introduction to the report: Assessing forest health in the Athabasca Oil Sands Region Thomas Clair, Wood Buffalo Environmental Association, Fort McMurray, AB, Mervyn Davies, Stantec Consulting Ltd., Calgary, AB, Keith Puckett, ECOFIN, Waldemar, ON, and Kevin Percy, Wood Buffalo Environmental Association, Fort McMurray, AB

1.1 Industrial atmospheric emissions in the Athabasca Oil Sands Region The Great Canadian Oil Sands operation (now Suncor) was the first commercial-scale bitumen extraction development in the Athabasca Oil Sands Region (AOSR), starting operations in 1967. This was followed by the Syncrude Mildred Lake operation in 1978. Both operations now consist of integrated mine, extraction and upgrading facilities which produce synthetic crude oil as an end-product. Mining was initially undertaken using electrically driven bucket-wheel excavators, draglines, and conveyer belt systems, though is now undertaken using a shovel and truck approach. The associated upgrading facilities were substantive sources of sulphur dioxide (SO2) emissions due to the sulphur content of the bitumen and the associated use of a bitumen product (e.g., coke or coke burner off-gas) as a fuel source. Most of the SO2 emissions in the Regional Municipality of Wood Buffalo (RMWB) however, now result from upgrading operations (>97%) (NPRI 2010, 2011).

The Syncrude Aurora North and the Albian Muskeg River projects began production in 1998 and 2003, respectively, using shovel and truck mining, and extraction facilities to produce diluted bitumen as an end-product. As the associated upgrading takes place offsite, these facilities are not substantive sources of local SO2 emissions. In addition, more recent commercial operations use non-mining (i.e. in situ) methods that inject steam into the bitumen formation to facilitate extraction.

With the commissioning of the Great Canadian Oil Sands operation in 1969, the initial regional -1 SO2 emission rate was 132 t d , with this amount growing with increasing production (Fig. 1, top). -1 In 1981, the RMWB SO2 emission rate increased from 250 to more than 450 t d with the commissioning of the Syncrude operation. The SO2 emission rate from these two facilities peaked in 1995 with a combined emission rate of 478 t d-1. In 1997, Suncor commissioned a flue gas -1 desulphurization system, reducing the SO2 emission rate to 278 t d . Other upgraders were commissioned in 2008 (Nexen Long Lake) and 2009 (CNRL Horizon) again increasing the RMWB SO2 emissions. The in situ projects are also sources of SO2 emissions from combustion of produced gases containing trace levels of reduced sulphur compounds (RSC).

In contrast to SO2 emissions, NOX emissions (Fig. 1, bottom) are not driven by a single source type. Regional sources broadly include the upgrader stacks, mine fleet exhausts, in situ steam

generator stacks, other oil and gas production activity stacks, and non-industry sources. The latter include community heating and traffic, and highway traffic emissions. In addition, NOX emissions are not as well documented as the SO2 emissions and historical data are not available for the earlier years.

Figure 1. 1 Historical trend in SO2 (top) and NOx (bottom) emissions from oil sands extraction activities (from National Pollutant Release Inventory (NPRI)). WB, Wood Buffalo.

-1 -1 NOX emissions show a general increase from 108 t d in 1996 to a maximum of 258 t d in 2011. The lower emissions in 2002 may reflect the uncertainty associated with the first year of NPRI reporting. Unlike the SO2 emissions, the NOX emissions are likely to increase in the future as more facilities are approved and become operational.

1.2 History of WBEA forest research in the AOSR The AOSR bitumen deposits lie beneath the Boreal Plains Ecozone which is dominated by upland jack pine, aspen, mixed forest, wetland bogs and poor fens. Due to the obvious potential effects of air pollutants from oil extraction operations on regional natural ecosystems, research on air pollutant effects on northern Alberta forests was undertaken to varying degrees since the first oil sands operation began in the late 1960’s (Percy et al., 2012).

Industrial development in the AOSR had led to concerns about the emissions to the atmosphere, focusing on the associated potential impacts of the industrial emissions to the surrounding environment. To address the concern, the Wood Buffalo Environmental Association initiated the Terrestrial Ecosystem Effects Monitoring (TEEM) Program in the mid 1990’s. The TEEM Acid Deposition Monitoring Program (AMP) was established in 1998 to determine if anthropogenic emissions of acidifying compounds such as SO2 and NOx gases were having long-term adverse effects on regional terrestrial environment, and to assess what was the magnitude of the impact.

In 1998 TEEM initiated measurement and sampling at a network of 10 jack pine (Pinus banksiana Lamb.) dominated interior forest stand plots (AMEC, 2000). Five additional plots were added between 1999 and 2003, and two were lost to development. Another cycle of measurement and sampling of soils and vegetation occurred at 13 plots in 2004 (8 were previously sampled in 1998 and one in 2001).

Monitoring of foliar vigour and stand condition in 2004 revealed no emissions-related effects on either needle retention or condition, and no anomalous damage/health issues in any of the 13 study sites (Jones and Associates, 2007). Analysis of foliar chemistry data showed that the effect of local industrial emissions was evident in increased concentrations of total sulphur, inorganic sulphur, iron and nickel, all of which are known to be components of oil sands emissions. However despite evidence of increasing elemental concentrations in foliage with increasing predicted deposition levels, there was no demonstrated evidence of a negative effect on forest productivity.

There was nevertheless, a positive association between estimated potential acid input and soluble and exchangeable Mg and K contents and EC. This was unexpected, and was likely indicative of site differences that were unrelated to the impacts of industrial air emissions. This then required that a better understanding of these differences was necessary to properly identify potential cause-effect relationships which may be occurring in the region.

Importantly, Jones and Associates (2007) revealed that the 13 sites measured in 2004 were imperfect ecological analogues as differences in site factors confounded interpretations of the potential effects of local industrial emissions. It was concluded that continued future sampling at these sites would have little chance to relate air emissions to forest ecosystem change. It was, therefore, apparent that the existing approach to assessing these parameters, while founded on accepted existing designs and concepts available in 1997 required review.

In 2006-2007, the Science Subcommittee (SSC) of TEEM undertook a review of preliminary results of the AMP 2004 sampling. The review was conducted by TEEM science advisors Legge, Percy, and Maynard along with external experts contributing. The reviewers considered concerns/recommendations made in Jones and Associates (2007) against the backdrop of published research and monitoring experience, including retrospective reviews of other monitoring programs (Percy and Ferretti, 2004). Recommendations for a significant science enhancement were presented to SSC (Percy, 2007) and ultimately approved by TEEM. In late 2007, the WBEA Board approved the new program design for implementation beginning 2008.

The 2008-2012 TEEM program enhancement is described in Percy, Maynard and Legge (2010). Key improvements made included moving to an integrated assessment by measuring at key points along the air pollutant pathway, from source to sink. TEEM adopted the forest health approach to monitoring as it has been shown to be successful (Percy, 2002):  When scales of stressors have been considered  When monitoring has been succeeded by process-oriented research  When appropriate indicators and endpoints were measured  When investigations on physical/chemical cycles were coupled with biological cycles and  When there was continuity in investigation

In order to "...implement an approach for establishing/determining cause-effect relationships between air pollutants and forest ecosystem health in the Oil Sands Region", six key elements of the new design were to: 1. Adopt the forest health approach 2. Change the conceptual sampling design 3. Relax the stand area restriction 4. Maintain adaptive capacity under continuous industrial development 5. Incorporate ecologically analogous sites 6. Co-measure inputs (predictors) and responses (indicators) in space and time

A more complete discussion on theory and practice of forest health monitoring can be found in Percy et al. (2012). Attributes of the current TEEM forest health system are compared with those of the Europe and the U.S. in Tkacz et al. (2013).

Figure 1. 2 Location of sampling sites in the AOSR. Light blue hexagons are air monitoring stations, red triangles are passive monitoring sites, green squares are TEEM Forest Health plots, yellow dots are meteorological towers and dark blue spots are the forest health edge plots.

Currently, the forest health network which has been developed by the TEEM program comprises 25 interior stand plots, 25 forest edge plots for early warning, meteorological towers, and passive/active monitoring analyzers which provide independent variables to determine cause- effect change with dependent soil and vegetation indicators (Fig. 1.2).

1.3 Organization of this report The main objectives of the TEEM Program are to detect, characterize and quantify the effect that air emissions have had or may have, on terrestrial ecosystems so that WBEA members, regulators, and other stakeholders can make informed decisions on airshed management. Activities in TEEM including the forest health program also directly address WBEA Strategic Goals #1 (To monitor to meet eight objectives), #2 (To establish cause-effect relationships), and #3 (To maintain high quality measurements and data sets in support of user needs) (WBEA, 2011).

There are three main themes to the chapters of this technical report. The first includes three papers describing concentration and deposition patterns of S, N and ozone (O3) in the region, which were assumed to be the main anthropogenic stressors to the forests of the AOSR. As it is impossible to provide good regional coverage of pollutant deposition due to the lack of electrical infrastructure at isolated sites, indirect, passive methods were used to estimate S and N concentrations and deposition, as well as O3 concentrations at the sampling sites. Chapter 2 (Y.- M. Hsu and A. Bytnerowicz) provides estimates of air concentrations and estimated deposition of S and N, as well as O3 in AOSR forests based on the use passive air samplers and deposition models. M. Fenn (Chapter 3) used a resin sampling approach to measure pollutant deposition at forest sites to try to estimate the deposition of chemicals in the region. The third chapter of this series (M. Davies et al., Chapter 4) uses a combined atmospheric-transport modelling approach to estimate air concentrations and deposition in the region.

The second group of papers describes the location and descriptions of the TEEM forest sampling sites. K. Foster provides the rationale for the site selection and to the changes in the network composition over time (Chapter 5). Chapter 6 (E. MacDonald) describes the general plot vegetation, including tree ages and sizes measured in the 2011/12 sampling.

The third theme is populated with 5 chapters which evaluate various ecosystem components and how they react (or not) with atmospheric chemistry. S. Grayston describes soil microbial ecology in relation to changing air chemistry (Chapter 7) while D. Maynard reports on soil and tree needle chemistry in relation to sulphur and nitrogen (Chapter 8). E. MacDonald assesses the relationship between the forest understory and air quality (Chapter 9), while K. Puckett does the same for lichens (Chapter 10). Chapter 11 (S. Watmough and C. Whitfield) relates the deposition estimates to critical load calculations based on the plant community and soils to predict if deleterious effects to forest soils can be expected. These critical loads are then compared to the measured values.

The final section (Chapter 12, Clair and Percy) summarizes the main findings of the various indicators and integrates them together to produce an overall picture of forest health in the AOSR.

1.4 References AMEC. 2000. Monitoring long-term effects of acid emissions in northeast Alberta – 1990 annual report. Wood Buffalo Environmental Association, Fort McMurray, AB.

Jones, C.E., Associates 2007. Terrestrial environmental effects monitoring – acidification monitoring program: 2004 sampling event report for soils, lichen, understory vegetation and forest health and productivity. Prepared for Wood Buffalo Environmental Association, Fort McMurray, AB, Canada.

NPRI 2010, 2011. National Pollutant Release Inventory (NPRI). http://www.ec.gc.ca/inrp-npri. [accessed July, 3 2014].

Percy, K.E. 2002. Is air pollution an important factor in international forest health? pp 23-42. IN R.C. Szaro, A. Bytnerowicz and J. Oszlanyi (Eds.) Effects of Air Pollution on Forest Health and Biodiversity in Forests of the Carpathian Mountains., IOS Press, Amsterdam, NL.

Percy, K. 2007. WBEA TEEM Monitoring in the Athabasca Oil Sands (AOS): "Going Forward". Presetation to TEEM SSC, March 7, Edmonton. Available at www.wbea.org.

Percy, K.E. and Ferretti, M. 2004. Air pollution and forest health: Towards new monitoring concepts. Environmental Pollution 130: 113-126.

Percy, K.E., Maynard, D., and Legge, A.H. 2010. Going forward: Enhancing the WBEA Terrestrial Effects Monitoring Program. Extended Abstract 2010-A-1009-AWMA. Proceedings Air and Waste Management Association Conference, Calgary, Alberta, June 22-25, 2010. Available at www.wbea.org.

Percy, K.E., Maynard, D.G., and Legge, A.H. 2012. Applying the forest health approach to monitoring boreal ecosystems in the Athabasca oil sands region. pp 193-218. IN K.E. Percy (Ed.) Alberta Oil Sands:Energy, Industry and the Environment. Elsevier, Oxford, UK.

Tkacz, B., Riiters, K., Percy, K.E. 2013. Forest monitoring methods in the US and Canada-an overview. pp 43-73. IN Ferretti. M, Fischer, R. (Eds.) Forest Monitoring Terrestrial Methods in Europe with Outlook to North America and Asia. Elsevier, Oxford, UK.

WBEA 2011. Strategic Plan Wood Buffalo Environmental Association 2011-2015. Available at www.wbea.org.

Chapter 2: Air pollution and dry deposition of nitrogen and sulphur in the AOSR estimated using passive samplers

Yu-Mei Hsu, Wood Buffalo Environmental Association, Fort McMurray, AB and Andrzej Bytnerowicz, US Department of Agriculture, Forest Service, Pacific Southwest Research Station, Riverside, CA, USA

2.1 Introduction

Ammonia (NH3), nitrogen dioxide (NO2) and nitric acid vapor (HNO3) are important components of reactive atmospheric nitrogen (Nr) and major drivers of atmospheric nitrogen (N) dry deposition to forests and other ecosystems (Hanson and Lindberg, 1991; Lovett, 1994; Magnani et al., 2007). Nitrogen dioxide is produced during oxidation of nitric oxide (NO) emitted during fossil fuel combustion, forest fires, and other natural and anthropogenic processes (Finlayson- Pitts and Pitts, 2000). Nitric acid is a final product of complex photochemical reactions between - - NO, NO2, O and OH radicals (Seinfeld and Pandis, 2006). Ammonia emissions are caused by agricultural activities, biological decay processes, catalytic converters, smoldering phase of fires, and other activities (Krupa, 2003).

Sulphur dioxide (SO2) may be released during natural processes (volcanic emissions, forest fires) and from anthropogenic activities such as combustion of fossil fuels, refining and smelting of sulphide ores, and other industrial processes (Legge et al., 1998). Ambient ozone (O3) is an important air pollutant which adversely affects human and forest health worldwide (Ainsworth et al., 2012). It is produced during complex photochemical reactions between nitrogen oxides (NOx), volatile organic compounds (VOCs) and carbon monoxide (CO), with reaction rates depending on ratios of the reactants and ambient conditions such as radiation and temperature of thermal inversions (Seinfeld and Pandis, 2006). All these pollutants are potentially phytotoxic at very high ambient concentrations, and their toxicity may be enhanced at low temperatures when rates of the biochemical detoxification mechanisms in plants are low (Bytnerowicz et al., 1998).

Passive samplers for NH3, NO2, HNO3, SO2 and O3 are useful for providing data describing ambient air conditions relevant to the health of forest plants. Passive samplers are simple to use, inexpensive, do not require electricity or air conditioned shelters and thus can be used in remote locations (Krupa and Legge, 2000; Bytnerowicz et al., 2005). As such, passive samplers allow for spatial coverage of the areas of interest providing data that can be used for the generation of geostatistical maps of air pollutants distribution. The information gathered supports the better understanding of the atmospheric chemistry, attribution, deposition, regional trends, deposition,

regional imports and exports for those chemical species. The objectives of this chapter are (1) to characterize the spatial and temporal variations of NH3, NO2, HNO3, SO2 and O3 concentrations in the Athabasca Oil Sands Region (AOSR); (2) to estimate the annual dry depositions of NH3, NO2, HNO3 and SO2; (3) to compare the dry deposition flux with IER (Chapter 3) and CALPUFF model (Chapter 4).

2.2 Methods

2.2.1 Monitoring, sampling and analysis for NH3, HNO3, SO2, NO2 and O3

Passive samplers to collect NH3, HNO3, SO2, NO2 and O3 were exposed to ambient air from towers above the tree canopy, where they were attached to cables on a pulley system, or on 2 m high posts in the industrial areas of the AOSR and in wetlands. The samplers were collected every month in the summer and every 2 months in winter.

Ambient concentrations of HNO3 and NH3 were monitored with passive samplers near the mining and industrial operations and in remote areas of the AOSR from 2005 until 2013. The monitoring network for NH3 and HNO3 consisted of 25 sites in 2005, gradually grew to 38 sites in 2009 and was reduced to 26 sites by 2010 - 2013. Passive samplers with two replicate filters coated with citric acid of the Ogawa design (Roadman et al., 2003) were used for NH3 monitoring. Ammonia reacts with citric acid on the filters producing ammonium citrate. After water extraction, + ammonium (NH4 ) concentrations in filter extracts were determined colorimetrically on a TRAACS 2000 Autoanalyzer, and ambient NH3 concentrations were calculated based on a comparison of passive samplers against the collocated annular denuder systems (Koutrakis et al., 1993). Three replicate HNO3 samplers of the US Forest Service design (Bytnerowicz et al., 2005) were used at each monitoring site. In these samplers, ambient air passes through a Teflon membrane and - gaseous HNO3 is absorbed on a Nylasorb nylon filter as nitrate (NO3 ). Nylon filters from the – samplers were extracted in nano-pure water, and NO3 concentrations in filter extracts were analyzed by ion chromatography (Dionex ICS 2000 LCD). Average ambient HNO3 concentrations were calculated using calibration curves developed by comparing the passive samplers against the collocated annular denuder systems (Koutrakis et al., 1993). In the field trials performed in Riverside, California, the samplers showed high accuracy (relative standard deviation of three replicate readings of ~5%). Exposed HNO3 and NH3 samplers were shipped to the USDA FS chemical laboratory in Riverside, California, where samplers’ filters were extracted. The HNO3 - sampler extracts were analyzed for NO3 using ion exchange chromatography (Dionex ICS 2000 + LCD), and the NH3 sampler extracts for NH4 were done colorimetrically with the TRAACS 2000 Autoanalyzer.

WBEA has operated a passive program for monitoring NO2, SO2 and O3 at six sites (AH3, AH7, JP101, JP102, JP104 and JP107) since 2000. The all-season SO2 passive sampling system (SPSS) (Tang et. al, 1997), the NO2 passive sampling system (NPSS) (Tang et al., 1999) and the O3 passive sampling system (OPSS) (Tang and Lau, 2000) were used to estimate these parameters. After collection, SO2, O3 and NO2 samples were extracted and analyzed by ion chromatography (DX- 2- - - 120, Dionex Corp., US) for SO4 , NO3 , and NO2 concentrations by Maxxam Analytic Inc. (Hsu,

2013).

Air pollutant distribution maps based on the passive sampling results for NH3, HNO3, NO2 and SO2 were developed with inverse distance weighted (IDW) methods using the Geostatistical Analyst, an extension of the ESRI ArcGIS software. This method was selected because for the given data, characterized by relatively low number of samples, high spatial variation, and lack of trend or spatial pattern, it was providing the most accurate results (Johnstone et al., 2001). Examples of distribution of NH3, HNO3, NO2, reactive N and SO2 were provided for winters (from November to April) 2010 and 2011 and summers (from May to October) 2011 and 2012. For O3 concentration (Fig. 2.14) and deposition flux (NH3-N, HNO3-N, NO2, total N and SO2), kriging was used which provides an optimal interpolation estimate for a given coordinate location, as well as a variance estimate for the interpolation value.

2.2.2 Estimating dry deposition flux using the Multi-Layer Inferential Model The National Oceanic and Atmospheric Administration (NOAA) multilayer inferential model (referred to as MLM) is used by the U.S. Environmental Protection Agency in the CASTNet 2- monitoring program. CASTNet monitors ambient ozone (O3), sulphur compounds (SO2, and SO4 - + ), and nitrogen compounds (HNO3, NO3 , and NH4 ) throughout the United States (Clarke et al., 1997). MLM is used to estimate dry deposition flux for these chemical species. A brief description of MLM and how it is used is provided below. More complete descriptions are offered by Meyers et al. (1998) and Cooter and Schwede (2000).

Chemical species dry deposition velocity (Vd) is estimated in MLM using an inference approach based upon an assumption that deposition flux can be estimated as the linear product of ambient concentration (C) and deposition velocity (Vd) (Meyers et al., 1998; Wesely and Hicks, 2000):

Flux = C × Vd Equation 2.1

Where C = chemical species concentration, and Vd = chemical species deposition velocity. The MLM is based on a resistance model framework analogous to Ohm’s Law (Meyers et al., 1998):

Vd = 1/(Ra+Rc) Equation 2.2

Where Ra = aerodynamic resistance between some height (a shallow sub-layer within the atmospheric constant flux layer, as a function of atmospheric turbulence and stability, and surface characteristics) above local ground and the canopy height, and Rc = total canopy resistance.

Deposition variations of the 2009-2012 annual average for NH3 as N, HNO3 as N, NO2 as N, total N, and SO2 as S were produced by using Surfer (Golden Software Inc.) for creating the spatial grids of data using Kriging (default options) from the point measurements and isopleths and ArcGIS for the maps.

2.3 Results and discussion

2.3.1 Ammonia (NH3) ambient concentrations

Since the beginning of monitoring in May 2005, NH3 concentrations were quite variable in time, without clear patterns, with peak values approaching 25 µg m-3 during the summer of 2011 (from May to October), though most of other values were generally <10 µg m-3 (Fig. 2.1). Ammonia was much higher in summers 2011 and 2012 than in winters (from November to April) 2010 and 2011 (Fig. 2.2) with the highest values in the centre and south western portion of the AOSR.

High NH3 values in the S and SW portion of the monitored area in summer probably not related to oil processing. Most likely the high NH3 values result from a dispersion were NH3 emissions from the nearby agricultural lands north of Edmonton and a long-range transport of the pollutant from the Grand Prairie/Peace River farmland with prevailing westerly winds in summer (http://www1.agric.gov.ab.ca/$department/deptdocs.nsf/all/sag6452/$FILE/onl_s_19_twp_an nual_normals_19712000.gif) into the AOSR. In dry summer conditions, NH3 can be transported far away from the pollution-source areas as has been shown for the eastern parts of the Sierra Nevada affected by agricultural emissions from the Central Valley of California (Bytnerowicz et al., 2014). This phenomenon was not observed in winter in absence of agricultural activities, and highest winter NH3 values were limited to the industrial center of the AOSR. The highest values that occurred in summer 2011, about two- fold higher than in summer 2012, were most likely caused by extensive fires in northern Alberta during that period.

-3 Figure 2. 1 Changes in NH3 concentrations (µg m ) at all monitoring sites from May 2005 to September 2013.

Figure 2. 2 NH3 spatial variations in winter 2010, summer 2011, winter 2011 and summer 2012 (summer: from May to October; winter: from November to April) (Inverse Distance Weighted, IDW).

2.3.2 Nitric acid (HNO3) ambient concentrations

From 2005 to 2013, HNO3 concentrations were quite variable. Concentrations of this secondary pollutant were generally low in winter with no significant differences between the sites. In summer, HNO3 concentrations were higher due to the photochemical nature of this pollutant. The highest values approaching 25 µg m-3 were measured during the 2011 summer season. The Richardson fire was an additional high source of HNO3 in the NE, which contributed to the overall higher levels of HNO3 experienced during the summer of 2011 shown during EXP1106 through EXP1109 (Fig. 2.3). In summer 2012 the highest HNO3 concentrations occurred in the northern and south-western portions of the AOSR (Fig. 2.4). Long-range transport of HNO3 produced during photochemical reactions in the urban-agricultural area of Edmonton could be responsible for elevated levels of the pollutant in SW portion of the AOSR. This phenomenon could resemble that seen in the San Bernardino Mountains of southern California affected by the Los Angeles photochemical air plume and the W slopes of the Sierra Nevada receiving polluted air masses from the Central Valley of California and the San Francisco Bay Area (Bytnerowicz et al., 2014).

-3 Figure 2. 3 Changes in HNO3 concentrations (µg m ) in all monitoring sites from May 2005 until September 2013.

Figure 2. 4 HNO3 spatial variations in winter 2010, summer 2011, winter 2011 and summer 2012 (summer: from May to October; winter: from November to April) (Inverse Distance Weighted, IDW).

2.3.3 Nitrogen dioxide (NO2) ambient concentrations

The annual NO2 concentrations from 2000 to 2013 (Fig.2.5) show that of the six sites sampled, NO2 concentrations at JP102 and JP104 were higher than those at AH3, AH7, and JP101 and that concentration at JP104 varied significantly (Hsu, 2013). From 2000 to 2013, the highest NO2 median concentrations occurred in 2009.

12 12 (A) AH3 (B) AH7 10 10

8 8

(ppb) (ppb) 6 6

NO 4 4 95%

2 2 90%

0 0 75% 2000 2002 2004 2006 2008 2010 2012 2000 2002 2004 2006 2008 2010 2012 12 12 (C) JP101 Year (D) JP102 Year median 10 10 25% 8 8 10%

(ppb) 6 (ppb) 6 5%

2 2

NO NO 4 4

2 2

0 0 2000 2002 2004 2006 2008 2010 2012 2000 2002 2004 2006 2008 2010 2012 12 12 (E) JP104 Year (F) JP107 Year 10 10

8 8

(ppb) 6 (ppb) 6

2 2

NO NO 4 4

2 2

0 0 2000 2002 2004 2006 2008 2010 2012 2000 2002 2004 2006 2008 2010 2012 Year Year

Figure 2. 5 Annual fluctuations in NO2 concentrations at sites (A) AH3, (B) AH7, (C) JP101, (D) JP102, (E) JP104, and (F) JP107. From Hsu (2013).

The spatial variations of NO2 concentration for winter 2010, summer 2011, winter 2011 and summer 2012 (Fig 2.6) show that the elevated NO2 concentrations were observed near the major industrial emissions and residential areas for both summer and winter. The NO2 concentrations in winter (2010 and 2011) were higher than those in summer (2011 and 2012). In the summer, the elevated NO2 concentration is localized and limited to a small area close to the major emission sources. In the winter, the elevated NO2 concentration is distributed more widely and to the north, due to the Athabasca Valley influence (prevailing wind direction). In winter months, meteorological condition (e.g., low temperature, stable atmosphere, less sunlight, low mixing height and ice surface) did not favor NO2 dilution, NO2 photochemical reaction, and NO2 deposition in the atmosphere thus resulting in longer life time for NO2.

Monthly NO2 concentration variations from 2000 to 2013 (Fig 2.7) at the six sites show a strong seasonal variation: high in the winter and low in the summer months. The resulting higher NO2 emissions in the winter included more NO2 emitted to the atmosphere from both on-road and off-road vehicles at both idling and cold-start conditions and residential heating in the winter. At 57°N latitude, winter day length is very short (seven hours of daylight on Dec. 21st), and temperature is extremely low (average January high -13°C), and these factors necessarily result in slow photochemical reaction rates. Among the six sites, NO2 concentrations were highest in the winter months at JP104 (median 7.2 ppb in December) and JP102 (median 4.2 ppb in December) as these sites are closer to two main emission sources with the NO2 concentrations at AH3, AH7 and JP101 showing a similar pattern though with lower concentrations. However, JP107, 94 km north of oil sands operations, which is influenced by the Athabasca Valley orography, also had higher NO2 concentrations during winter months. Comparing the annual average passive NO2 concentrations from 2009 to 2012 with NO2 predictions from CALPUFF model (Davies et al., Chapter 4, Appendix 2.3) show that the NO2 concentrations from the model were generally higher than those from the passive measurements, especially at JP102, JP212, AH7, NE11 and WF4 (Fig. 2.8) though the pattern of the spatial distribution between the two approaches was similar.

A non-parametric approach, the Sen-Thiel trend analysis of time series (Hsu, 2013), has been used on the 14-year monthly NO2 data from AH3, AH7, JP101, JP102, JP104 and JP107. Table 2.1 lists the Sen slope (Sen) and the corresponding Sen-Theil 95% confidence interval (Slow, Shigh). The results show the NO2 concentrations at JP104 near the major emission sources to have increased during the past 14 years (Fig. 2.5) though AH3, AH7, JP101, JP102 and JP107 show no statistically significant differences.

µg N/m3 0.03-0.30 0.30-0.61 0.61-0.91 0.91-1.22 1.22-1.52 1.52-1.83 1.83-2.13 2.13-2.43 0.01-0.30 2.43-2.74 0.30-0.61 2.74-3.04 0.61-0.91 3.04-3.88 0.91-1.20

0.01-0.30 0.30-0.61 0.61-0.91 0.91-1.22 1.22-1.52 1.52-1.83 0.02-0.30 1.83-2.13 0.30-0.61 2.13-2.43 0.61-0.91 2.43-2.74 0.91-1.22 2.74-3.04 1.22-1.52 3.04-4.60 1.52-1.82

Figure 2. 6 NO2 spatial variations in winter 2010, summer 2011, winter 2011 and summer 2012 (summer: from May to October; winter: from November to April) (Inverse Distance Weighted, IDW).

95% 90% 75%

median 25% 10%

Figure 2. 7 Monthly variation in passively measured NO2 concentrations from 2000 to 2013, (A) AH3, (B) AH7, (C) JP101, (D) JP102, (E) JP104, and (F) JP107. From Hsu (2013).

20 NO2 concentration

15 3

g/m m 5

NE7

AH3

AH7

SM7

SM8

WF4

BM7

NE10

NE11

JP101

JP102

JP104

JP107

JP205

JP210

JP212

JP213

BM10

BM11 AH8-R CALPUFF 2009 2010 2011 2012

Figure 2. 8 Annual mean NO2 concentrations from CALPUFF and passive samplers (2009 to 2012).

Table 2. 1 Sen-Theil test of slope significance for NO2 Slow Sen Shigh SLOPE* AH3 –0.001286 0.000353 0.002562 No AH7 –0.002273 0.000000 0.002404 No JP101 –0.002740 –0.001087 0.000086 No JP102 –0.004291 –0.000048 0.003704 No JP104 0.009375 0.014778 0.021795 Yes JP107 –0.001047 0.001187 0.004110 No SLOPE*. No: no statistically significant slope (trend) at 95% confidence level. Yes: statistically significant positive slope (trend) at 95% confidence level.

2.3.4 Total inorganic gaseous reactive N (Nr sum)

Combining the HNO3, NH3 and NO2 concentration data on a network of 23 sites where all types of samplers were collocated, allowed calculation of inorganic gaseous reactive nitrogen (Nr sum) (Bytnerowicz et al., 2010). Highest “Nr sum” levels were driven in winter by NO2 concentrations and in summer mainly by NH3 with a significant contribution of NO2 and HNO3. Geostatistical maps show that the areas in the vicinity of Fort McKay and Fort McMurray had the highest levels of Nr both in winter and summer. The highest Nr concentrations were observed in summer 2011 due to the wildfire emissions.

During that period the most affected regions were the central and southwestern portion of the AOSR. In summer 2012 highest levels of pollution were in the center as well as western and southern areas of the AOSR domain (Fig. 2.9, Appendix 2.4). These high levels of total reactive N in the S and SW portions of the AOSR could also be caused by a long-range transport of NH3 and HNO3 downwind of Edmonton and other urban and agricultural centers in central Alberta.

2.3.5 Sulphur dioxide ambient concentrations

Annual SO2 concentrations from 2000 to 2013 at the six sampling sites are displayed Figure 2.10. SO2 concentrations at JP102, AH7, and JP104 were higher than those from the other sites. The

spatial variations of SO2 concentration for winter 2010, summer 2011, winter 2011 and summer 2012 are showed in Figure 2.10. The SO2 spatial variations in summer 2011 and summer 2012 were similar and the elevated SO2 concentrations were observed near the major industrial emissions, e.g., stacks. After being emitted from sources, SO2 undergoes dilution, chemical reaction, dry deposition to surfaces, and the concentration decreases quickly.

The SO2 distribution is localized and limited to a small area close to the major emitters in summer months. The SO2 concentrations in winter were generally higher than those in summer. SO2 spatial variations in winter 2010 and winter 2011 followed a similar pattern with the highest concentrations observed near the major emission sources. The lower mixing height and stable atmosphere, wind speed and wind direction are the major factors for the controlling SO2 distribution in winter months. The annual average SO2 concentrations from CALPUFF (Davies et al., Chapter 4, Appendix 4.1) and passive samplers (Hsu, 2013) were very close at JP102, JP210, JP212, NE10 and SM8; the SO2 concentrations from CALPUFF model were lower at few other sites (e.g., JP101, JP107, JP205, AH8-R) (Fig. 2.8). However, the SO2 concentrations (Fig.2.10 and Fig.4.1) from two methods show the similar spatial pattern that the elevated SO2 concentrations in the west of Fort McMurray and the east of Fort McKay were reported by both passive measurement and CALPUFF. Over the 14-year sampling period, the temporal trends of SO2 concentration at the 6 sampling sites (Table 2.2), showed there was a decrease in air concentration at JP102 and JP104, perhaps a decrease at JP101 and that there was no statistically significant change at AH3, AH7 and JP107.

Table 2. 2 Sen-Theil test of slope significance for SO2 Slow Sen Shigh SLOPE* AH3 -0.002353 -0.000725 0.000450 No AH7 -0.004645 -0.001387 0.001389 No JP101 -0.003571 -0.001613 0.000000 Perhaps decrease JP102 -0.005712 -0.003305 -0.001042 Yes JP104 -0.006111 -0.003724 -0.001756 Yes JP107 -0.000727 0.000779 0.002273 No * No: no statistically significant slope (trend) .Yes: statistically significant positive slope (trend) at 95% confidence level.

Figure 2. 9 Reactive gaseous inorganic N spatial variations in winter 2010, summer 2011, winter 2011 and summer 2012 (summer: from May to October; winter: from November to April) (Inverse Distance Weighted, IDW).

95% 90% 75%

median 25% 10%

Figure 2. 10 Annual SO2 air concentration value ranges at monitoring sites. From Hsu (2013).

Figure 2. 11 SO2 spatial variations in winter 2010, summer 2011, winter 2011 and summer 2012 (summer: from May to October; winter: from November to April) (Inverse Distance Weighted, IDW).

95% 90% 75%

median 25% 10%

Figure 2. 12 Monthly variations in passively measured SO2 concentrations from 2000 to 2013, (A) AH3, (B) AH7, (C) JP101, (D) JP102, (E) JP104, and (F) JP107. From Hsu (2013).

7 6 SO2 concentration 5 3 4

g/m 3 m 2 1

NE7

AH3 AH7

SM7 SM8

WF4

BM7

NE10 NE11

JP101 JP102 JP104 JP107 JP205 JP210 JP212 JP213

BM10 BM11 AH8-R CALPUFF 2009 2010 2011 2012

Figure 2. 13 Annual mean SO2 concentrations from CALPUFF and passive samplers (2009 to 2012).

Three types of seasonal SO2 concentration patterns were found using monthly SO2 concentration variations from 2000 to 2013 (Fig. 2.12). No clear trend was observed at JP104. At sites AH3, JP101 and JP107, SO2 concentrations had a seasonal pattern which was low in summer and high in winter. The lower mixing height, stable atmosphere and slow photochemical reactions are probably the major reasons for the relatively high SO2 concentrations in the winter months (Hsu, 2013). In summer, SO2 reactions including dry deposition and heterogeneous reactions with aerosols (Seinfeld and Pandis, 2006), are usually faster due to greater surface areas exposed to water, higher concentrations of oxidants and higher temperature. The distance (or time) between the emission sources and these three sampling sites (AH3, JP101 and JP107) allowed SO2 to undergo direct deposition, heterogeneous reaction, and within plume dilution. AH7 and JP102, close to emission sources, showed similar patterns with the highest SO2 median concentration occurring in February and March and the lowest observed September to November. It is possible that these two sites were sometimes directly impacted by emissions.

Table 2. 3 Sen-Theil test of slope significance for O3 Slow Sen Shigh SLOPE* JP101 -0.041 -0.020 0.000 No JP102 -0.042 -0.023 -0.005 Yes JP104 -0.055 -0.035 -0.014 Yes JP107 -0.035 -0.013 0.007 No AH3 -0.036 -0.013 0.005 No AH7 -0.044 -0.025 -0.003 Yes SLOPE *, No: no statistically significant slope (trend) at 95% confidence level. Yes: statistically significant positive slope (trend) at 95% confidence level.

2.3.6 Ozone ambient concentrations

In 2013, the O3 passive sampling program was conducted at 28 sites, including 23 remote sites (AH3, AH7, AH8-R, BM7, BM10, BM11, JP101, JP102, JP104, JP107, JP108, JP205, JP210, JP212, JP213, JP316, NE7, NE10, NE11, SM7, SM8, R2, and WF4) and 5 air monitoring stations (AMS 1, AMS 2, AMS 6, AMS 8 and AMS 14). The 2013 annual average O3 concentration was low near the major NOx emission sources, including stacks and mine fleets, and increased as the distance increased (Fig. 2.14). The lowest 2013 annual average concentration was 16.0 ppb at R2 south of Fort McKay and the highest concentration was 28.7 ppb at JP316 and AMS 8 (Fort Chipewyan). The low O3 concentration in the industrial area was likely due to O3 titration. Nitric oxide emitted from sources (stack and mine fleets) reacts with available ambient O3 to form NO2 and O2. The O3 spatial variation suggests titration effect Figure 2. 14 2013 average O3 concentration (ppb) in the (low O3 concentration) is limited to the area AOSR (Kriging). close to industrial activity.

The passive dataset comprises a monthly O3 time series covering the time period from January 2000 to December 2013 (Table 2.3) and Sen-Thiel trend analyses were done for the 14 year data set from AH3, AH7, JP101, JP102, JP104 and JP107. The data show that the O3 concentrations from 2000 to 2013 have decreased at JP102, JP104 and AH7, sites which are all close to major oil sands activities. As indicated in section 2.3.3, the NO2 concentration at JP104 has increased from 2000 to 2013 so that O3 titration could be the reason to explain the decrease in O3 concentration and the increase in NO2 concentration at JP104. No statistically significant temporal changes in O3 concentration were detected for three sites, JP101, JP107 and AH3 (Hsu, 2013). 2.3.7 Nitrogen and sulphur deposition estimates based on passive sampler data

Monthly Deposition Velocities (Vd) for NH3, HNO3, NO2 and SO2 from 2000 to 2012 at AMS 1 are generally higher in summer than in winter (Fig. 2.15). The maximum Vd occurred in July for both NH3 and NO2, in summer months (May, June and July) for HNO3, and in March and July for SO2. Since Vd for NH3, HNO3, NO2 and SO2 is influenced by both ground surface and aerodynamic resistance, higher Vd reflects the more foliated and water ground cover (compared to snow and dust) with the combination of higher wind speeds in spring. Hence, the Vd is generally higher in summer months. Estimated Vd values of these species are within a reasonable range compared

to earlier studies (Brook et al., 1999; Krupa, 2003; Schwede et al., 2011; Sickles and Shadwick, 2007a; Sickles and Shadwick, 2007b; Zhang et al., 2009). In North America, two models (MLM and Big Leaf Model) are used for long-term monitoring network (including CASTNet and CAPMoN) to calculate the Vd and dry deposition. Schewede et al. (2011) have compared these two models and concluded the Vd of SO2 from MLM was lower than that from BLM but it is unclear how accurate the models are or which is the best for the AOSR. The MLM approach was used, taking this uncertainty in to account. Further calculations using BLM are recommended to determine which model might be more suitable for the AOSR.

95% 90% 75%

median 25% 10%

Figure 2. 15 Monthly deposition velocities of (a) NH3, (b) HNO3, (c) NO2 and (d) SO2 at AMS 1 (from 2000 to 2012) calculated by MLM.

2.3.7.1 Ammonia dry deposition Ammonia deposition values for 23 monitoring sites were calculated for individual years during 2006 – 2012 period and showed values ranging from 0.38 kg N ha-1 yr-1 at JP205 to 2.35 kg N ha- 1 yr-1 at SM7 in 2011 (Appendix 2.1). A map of average distribution for the 2009 – 2012 period (Fig. 2.16, left) shows two areas of high NH3 deposition – one in the center of the AOSR between Fort McMurray and Fort McKay; and the second larger area in the southern portion of the monitoring area. While the first area can reflect deposition caused by elevated NH3 emissions and concentrations during oil sands operations at the AOSR, the second most likely is caused by dispersion of NH3 emissions from the agricultural lands north of Edmonton and a long-range transport of the pollutant from the Grand Prairie/Peace River farmlands with the prevailing 27

westerly winds in summer. (http://www1.agric.gov.ab.ca/$department/ deptdocs. nsf / all / s ag6452/$FILE/onl_s_19_twp_annual_normals_19712000.gif). In general, the determined 4-year -1 -1 average NH3 deposition ranged between 0.7 to 1.25 kg N ha yr . These values were higher than the dry deposition values for HNO3 or NO2.

Passive samplers were collocated with IER samples (Fenn, Chapter 3) at nine sites including AMS 1, AMS 14, JP101, JP102, JP104, JP107, JP210, JP212 and JP213 from the summer in 2008. The + + NH4 dry deposition from passive data with MLM and the NH4 deposition from IER + open/throughfall samples (wet and dry) were similar (Fig. 2.17). All results show that the NH4 deposition in summer was higher than that in winter.

Figure 2. 16 NH3-N (left) and HNO3-N (right) deposition average from 2009 to 2012 in the AOSR in kg N ha-1 yr-1 (Kriging).

2.0 5

(A) AMS 1 open (B) AMS 14 open

)

1 - - throughfall 4 throughfall 1.5 MLM MLM 3 1.0

as N (kg ha

as N (kg ha 4

4 0.5 NH NH 1

0.0 0 1.0 4

) )

1 1 - - 0.8 throughfall throughfall 3 MLM MLM 0.6 2

0.4

as N (kg ha as N (kg ha 4

4 1 NH NH 0.2

0.0 0 3 3

(E) JP104 open (F) JP107 open

)

1 - throughfall 1

- throughfall

2 MLM 2 MLM

as N (kg ha as N (kg ha

4 1

NH NH

0 0 1.0 3

(G) JP210 open (H) JP212 open

)

1 - 0.8 throughfall - throughfall MLM 2 MLM 0.6

0.4

as N (kg ha

as N as (kg ha 4

4 1 NH 0.2 NH

0.0 0 3

(I) JP213 open

) 1

- throughfall 2 MLM -1 Figure 2. 17 NH4 as N deposition (kg ha )

comparison at 9 sites. (2008 S: 2008 summer, as N as (kg ha

4 1 from May 2008 to October 2008; 2008 W:

NH 2008 winter, from November 2008 to April 0 2009). Open and throughfall from: Fenn (Chapter 3).

2.3.7.2 HNO3 dry deposition -1 -1 Deposition of HNO3 during the 2006 – 2012 period ranged between 0.16 kg ha yr at AH8 in -1 -1 2007 and NE11 in 2009 to 2.70 kg ha yr at BM7 in 2012 (Appendix 2.2). Similarly to NH3, the highest HNO3 deposition values were determined in the central area of the AOSR and in its southern portion (Fig. 2.16, right). The deposition distribution pattern was similar to that of the HNO3 summer concentrations (Fig. 2.4). The 4-year average deposition values for this pollutant -1 -1 ranged between 0.3 and 0.9 kg N ha yr , and were slightly lower than the NH3 deposition values (Fig. 2.17). These values were in the similar range as those for NO2 deposition (Fig. 2.18, left).

2.3.7.3 NO2 deposition

NO2 deposition is a component of total N deposition for the AOSR. Nitrogen dioxide dry deposition was estimated at 23 sites from 2009 to 2012 (Appendix 2.3) with an annual average -1 -1 -1 -1 NO2 -N deposition ranging from 0.03 kg ha yr in BM7 to 0.98 kg ha yr at AMS 6 (Fig. 2.18,(left)). After NOx emitting from the sources, NOx undergoes dilution and photochemical reactions. The major NO2 removal mechanism from the atmosphere is a photochemical reaction, reacting with OH radicals during the daytime and NO3 radicals at night. The life time of NO2 is around one day (Seinfeld and Pandis, 2006) during the summer condition in the AOSR and with the short life time, the major NO2 deposition is localized.

The greatest NO2 deposition (Fig. 2.18, (left)) occurred in places where major oil sand operation activities took place and where communities existed. It is likely that elevated NO2 deposition resulted from industrial emissions (e.g., stacks and mine fleet) at JP104 and from local community activities (e.g., personal vehicles, transport trucks) at AMS 6. The dry deposition results from the passive NO2 and HNO3 data with MLM calculation at nine sites including AMS 1, AMS 14, JP101, JP102, JP104, JP107, JP210, JP212 and JP213 are in the range of IER bulk (open) and throughfall results (Fig. 2.19).

Figure 2. 18 NO2 -N (left) and total N (right) deposition average from 2009 to 2012 in the AOSR in kg N ha-1 yr-1 (Kriging).

3 3

(A) AMS 1 open (B) AMS 14 open

)

1 - - throughfall throughfall

2 MLM 2 MLM

as N (kg as N (kg ha

as N (kg as N (kg ha 3

3 1 1

NO NO

0 0 1.0 1.0

)

1 - 0.8 throughfall - 0.8 throughfall MLM MLM 0.6 0.6

0.4

as N (kg as N (kg ha 0.4

as N (kg as N (kg ha

3 NO 0.2 NO 0.2

0.0 0.0 3 1.0

(E) JP104 open (F) JP107 open

)

1 - throughfall - 0.8 throughfall 2 MLM MLM 0.6

as N (kg as N (kg ha 0.4

as N (kg as N (kg ha 3

1 3 NO NO 0.2

0 0.0 1.0 3

(G) JP210 open (H) JP212 open

) )

1 1 - - 0.8 throughfall throughfall MLM 2 MLM 0.6

0.4

as N (kg (kg N as ha as N (kg (kg N as ha 3

3 1 NO NO 0.2

0.0 0 1.0

(I) JP213 open

) 1

- 0.8 throughfall MLM 0.6 -1 Figure 2. 19 NO3 as N deposition (kg ha )

0.4 comparison at 9 sites. (2008 S: 2008 as N (kg (kg N as ha 3 summer, from May 2008 to October 2008; NO 0.2 2008 W: 2008 winter, from November 0.0 2008 to April 2009). Open and throughfall from: Fenn (Chapter 3).

2.3.7.4 Total reactive N deposition Total deposition of reactive inorganic gaseous N during the 2006 – 2012 period ranged between 0.78 kg N ha-1 yr-1 at AH8 in 2007 to 4.97 kg N ha-1 yr-1 at AMS14 in 2012 (Appendix 2.4). A map of the average N distribution for the 2009 – 2012 period (Fig. 2.18, (right)) shows two areas of high N deposition – one in the center of the AOSR between Fort McMurray and Fort McKay with values reaching 2.5 kg N ha-1 yr-1, and the second in the southern portion of the monitoring area with values approaching 2 kg N ha-1 yr-1. Distribution patterns for total deposition of inorganic gaseous N are similar to those of the NH3 deposition indicating relative importance of that pollutant. Nitric acid had a strong contribution to elevated N deposition in the southern AOSR, while NO2 contributed mainly to the elevated levels between Fort Mc Murray and Fort McKay.

While higher deposition in the center of the AOSR is mainly related to the industrial emissions of NO2 and NH3, while the elevated deposition values in the southern part of the AOSR could be caused by agricultural activities in central and western Alberta and long-range transport of NH3 (see the above discussion for NH3 and HNO3). It can be shown that nitrogen deposition from inorganic reactive gaseous pollutants is an important component of total (wet and dry) N deposition. While values of total wet and dry deposition calculated with the modeled CALPUFF are usually higher, the CALPUFF results were surprisingly lower at a few sites (BM10, BM7, and SM7). There is a need to examine performances of the multilayer and CALPUFF models.

2.3.7.5 Sulphur dioxide deposition

Sulphur dioxide (SO2) in the ambient air is removed by dry and wet deposition and oxidized to sulphate. The SO2 residence times are 60, 100 and 80 hours for removing by dry deposition, wet deposition and chemical reaction, respectively, and the overall residence time is 25 hours when three mechanisms are combined (Seinfeld and Pandis, 2006). All three removal mechanisms are important for ambient SO2 concentration. The SO2 as S dry deposition was calculated at 13 sites from 2006 to 2008 and 23 sites from 2009 to 2012. Annual averages ranged from 0.26 kg ha-1 yr-1 in BM7 in 2011 to 2.04 kg ha-1 yr-1 at AMS 1 in 2006. From 2006 to 2008, AMS 1 had the highest SO2-S deposition with values of 2.04, 1.91 and 1.61 kg ha-1 yr-1 in 2006, 2007 and 2008,

Figure 2. 20 2009-2012 SO2 (as S) annual average respectively. From 2009 to 2012, JP104 deposition (kg ha-1 yr-1) in the AOSR (Kriging). had the higher SO2 deposition than

-1 -1 AMS 1 (Fig. 2.20). The lowest SO2 as S deposition was 0.32 kg ha yr in BM7, 80 km north of Fort McKay. JP104, AMS 1, AMS 6, and NE7 had higher SO2 deposition rate, 1.24, 1.19, 1.12 and 1.05 kg ha-1 yr-1, respectively.

The elevated SO2 deposition was observed near the major SO2 emission sources, i.e., stacks (Fig. 2.20). Based on Environment Canada’s National Pollution Release Inventory (NPRI), there were two major SO2 emission sources (stacks) in the oil sand operation area and the highest deposition occurred in the same area. SO2 deposition trended towards the east and south most likely due to the Athabasca Valley influence. The increased SO2 deposition was also reported in Fort McMurray, an urban transportation hub.

The SO2 dry deposition estimated using passive SO2 concentration with MLM was compared to the sulphate deposition from IER (Chapter 3) as shown in figure 2.21 and CALPUFF model (Chapter 4) and the MLM estimate was considerably lower than the bulk/throughfall and CALPUFF model values. MLM and Big-Leaf Model (BLM) are used by USEPA and Environmental Canada for dry deposition estimates of SO2, NH3, NO2, and HNO3. Schewede et al. (2011) conducted an intercomparison of MLM and BLM and concluded that the Vd of SO2 from MLM and BLM were uncorrelated and the median Vd of SO2 from MLM was 49% lower than that from BLM. Sickles and Shadwick (2007) estimated the S and N deposition in the eastern U.S. and also identified that the Vd of SO2 might be underestimated for forested canopies. The MLM deposition was compared with CALPUFF deposition to understand if the MLM could be a reliable tool for deposition calculation in the AOSR although the CALPUFF deposition consists of both dry and wet depositions. It is expected that the deposition from both MLM and CALPUFF had similar patterns and the dry deposition should be 20%-60% of total deposition. Both MLM and CALPUFF conclude that the elevated SO2 deposition was identified in the west of Fort McMurray, the east of Fort McKay and Anzac (Fig. 2.20 and Fig. 4.2). IER bulk (open) sampler mainly collects wet deposition and throughfall sample includes wet deposition and dry deposition, at least partially, as integrated by the natural surface area of the forest canopy. Assuming the difference of SO4 as S deposition from IER bulk (open) and throughfall sample is SO4 as S dry deposition. The SO2 as S dry deposition from MLM is still lower than that from IER. It is possible that the MLM underestimated the SO2 as S dry deposition for near source calculation. Further investigation should be conducted to improve the model calculation or to apply the BLM for S deposition calculation.

20 1.0 20 1.0

(A) AMS 1 open (B) AMS 14 open

)

1 -

- throughfall 0.8 throughfall 0.8

)

) 1

15 1 15 - MLM - MLM 0.6 0.6

10 10 IER (kg (kg IERha

IER (kg IER (kg ha 0.4 0.4 MLM (kg MLM (kg ha 5 5 MLM (kg ha 0.2 0.2

0 0.0 0 0.0 10 1.0 20 1.0

)

1 -

- 8 throughfall 0.8 throughfall 0.8

) )

15 1

1 - MLM - MLM 6 0.6 0.6

10 IER (kg (kg ha IER

IER (kg (kg IERha 4 0.4 0.4 MLM (kg MLM(kg ha MLM (kg ha 5 2 0.2 0.2

0 0.0 0 0.0 20 1.0 10 1.0

(E) JP104 open (F) JP107 open

)

1 1 - 8 throughfall 0.8

- throughfall 0.8

)

) 1

15 1 - MLM - MLM 0.6 6 0.6

10 IER (kg (kg IERha

IER (kg (kg ha IER 0.4 4 0.4 MLM (kg ha 5 MLM(kg ha 0.2 2 0.2

0 0.0 0 0.0 10 0.6 20 1.0

(G) JP210 open (H) JP212 open

)

1 -

8 throughfall - throughfall 0.8

)

) 1

15 1 - MLM 0.4 MLM - 6 0.6

10 IER (kg (kg ha IER 4 (kg ha IER 0.4

0.2 MLM (kg MLM(kg ha 5 MLM(kg ha 2 0.2

0 0.0 0 0.0 10 0.6

(I) JP213 open

) 1

- 8 throughfall

) 1 MLM 0.4 - -1 6 Figure 2. 21 SO4 as S deposition (kg ha ) comparison at 9 sites. (2008 S: 2008 IER (kg (kg ha IER 4 0.2 summer, from May 2008 to October 2 MLM(kg ha 2008; 2008 W: 2008 winter, from November 2008 to April 2009). Open and 0 0.0 throughfall from: Fenn (Chapter 3).

2.4 Conclusion

NO and SO2 are the primary pollutants produced by oil sands facilities. The major emission sources are the upgrader stacks for SO2 and stacks, mine fleets and off-, on-road vehicles for NOx. After emitting from the sources, NO2 and SO2 undergo dilution and chemical reaction. As shown in figure 2.6, figure 2.11, figure 2.18 (left), and figure 2.20, both enhanced concentration and deposition were limited within a certain range (~30 km) from the emission sources. The highest NO2 and SO2 deposition occurred in major industrial operation areas and Fort McMurray. However, concentrations of NH3 and HNO3 were also greater outside of the AOSR industrial activities, mainly in its southern part. Ammonia increased levels were possibly also caused by forest fires and agricultural activities in central Alberta. Nitric acid, as the secondary pollutant produced via photochemical reactions, was found at distances from the main NOx emission sources both in the northern and southern directions. Consequently, deposition of these pollutants as well as deposition of total gaseous inorganic N was also greater in those areas.

The comparison of the N and S deposition results calculated with the MLM with the CALPUFF modeled values (Chapter 4) found similar patterns for dry deposition estimates. The major difference between the two approaches was the total S (wet and dry) from CALPUFF (Fig. 4.2) was significantly higher than the SO2 deposition (dry) calculated with the MLM especially near emission sources. The MLM might underestimate the SO2 dry deposition due to its default setting. The comparison for total N deposition shows a good agreement (and a good spatial trending) between the two methods for 17 sampling sites (Appendix 2.4) except AH7, JP102, JP104 and JP212. These four sites are very close to emission sources where the CALPUFF model might overestimate the total N deposition.

+ The NH4 deposition calculated from the passive sampler data with the MLM and determined with the IER open/throughfall sampling (Chapter 3) had similar patterns at 9 sampling sites. The - NO3 deposition results from two methods were similar. However, SO2 dry deposition calculated with the MLM was much lower than that from the IER sampling. Clearly a further study should be carried out for better SO2 dry deposition estimation.

2.5 References Ainsworth, E. E., Yendrek, C. R., Sitch, S., Collins, W. J., Emberson, L. D. 2012. The effects of tropospheric ozone on net primary productivity and implications for climate change. Annu. Rev. Plant Biol., 63, 637-661.

Brook, J.R., Zhang, L.M., Li, Y.F., Johnson, D., 1999. Description and Evaluation of a Model of Deposition Velocities for Routine Estimates of Dry Deposition over North America. Part II: Review of Past Measurements and Model Results. Atmos Environ 33, 5053-5070.

Bytnerowicz, A., Dueck, T., Godzik, S. 1998. Nitric oxide, nitrogen dioxide, nitric acid vapor and ammonia. In: R. Flagler (ed.) Recognition of Air Pollution Injury to Vegetation: a Pictorial Atlas, Air & Waste Management Association, Pittsburgh, PA, 5-1 through 5-17.

Bytnerowicz, A., Sanz, M. J., Arbaugh, M. J., Padgett, P. E., Jones, D. P., Davila A. 2005. Passive sampler for monitoring ambient nitric acid (HNO3) and nitrous acid (HNO2) concentrations. Atmospheric Environment, 39, 2655-2660.

Bytnerowicz, A., Schilling, S., Alexander, D., Fraczek, W., Hansen, M. 2010. Passive monitoring to estimate N (NO2, HNO3, NH3) exposure in remote areas and geospatial analysis to optimize monitoring networks in the Athabasca Oil Sands Region, 103 Annual Conference and Exhibition of the Air & Waste Management Association, Calgary Canada, June 22-25, 2010, Extended Abstract 2010-A-563-AWMA.

Bytnerowicz, A., Fenn, M., Allen, E., Cisneros, R. 2014. Ecologically relevant air chemistry. In: Ecosystems of California: a Source Book, H. Mooney and E. Zavaleta, Editors, University of California Press (in print).

Clarke, J.F., Edgerton, E.S., Martin, B.E., 1997. Dry Deposition Calculations for the Clean Air Status and Trends Network. Atmos Environ 31, 3667-3678.

Cooter, E.J., Schwede, D.B., 2000. Sensitivity of the National Oceanic and Atmospheric Administration Multilayer Model to Instrument Error and Parameterization Uncertainty. J Geophys Res-Atmos 105, 6695-6707.

Finlayson-Pitts, B. J., Pitts, J. N. Jr. 2000. Chemistry of the Upper and Lower Atmosphere. Academic Press, San Diego, USA.

Hanson, P. J., and S. E. Lindberg. 1991. Dry deposition of reactive nitrogen compounds: a review of leaf, canopy and non-foliar measurements. Atmospheric Environment, 25A, 1615-1634.

Hsu, Y.-M., 2013. Trends in Passively-Measured Ozone, Nitrogen Dioxide and Sulfur Dioxide Concentrations in the Athabasca Oil Sands Region of Alberta, Canada. Aerosol and Air Quality Research. pp 1-16.

Johnstone, K., Ver Hoef, J., Krivoruchko, K., Lucas, n. 2001. Using ArcGIS Geostatistical Analyst. ESRI, Redlands, California.

Koutrakis, P., Sioutas, C., Ferguson, S. T., Wolfson, J. M., Mulik, J. D., and Burton, R. M. 1993. Development and evaluation of a glass honeycomb denuder/filter pack system to collect atmospheric gases and particles. Environmental Science & Technology, 27, 2497-2501.

Krupa, S.V., 2003. Effects of Atmospheric Ammonia (NH3) on Terrestrial Vegetation: A Review. Environ Pollut 124, 179-221.

Krupa, S. V., and Legge, A. H. 2000. Passive sampling of ambient, gaseous air pollutants: an assessment from an ecological perspective. Environmental Pollution. 107, 31-45.

Legge, A. H., Jäger, H.–J., Krupa, S. V. 1998. Sulfur dioxide. In: R. B. Flagler (ed.) Recognition of Air Pollution Injury to Vegetation – a Pictorial Atlas. Air & Waste Management Association, Pittsburg, PA, 3-1 through 3-42.

Lovett, G. M. 1994. Atmospheric deposition of nutrients and pollutants in North America: an ecological perspective. Ecological Applications, 4, 629-650.

Magnani, F., Mencuccini, M., Borghetti, M., et al. (2007). The human footprint in the carbon cycle of temperate and boreal forests. Nature. 447: 848-450.

Meyers, T.P., Finkelstein, P., Clarke, J., Ellestad, T.G., Sims, P.F., 1998. A Multilayer Model for Inferring Dry Deposition Using Standard Meteorological Measurements. J Geophys Res-Atmos 103, 22645-22661.

Roadman, M. J., Scudlark, J. R., Meisinger, J. J., Ullman, W. J. 2003. Validation of Ogawa passive samplers for the determination of gaseous ammonia concentrations in agricultural settings. Atmospheric Environment, 37, 2317-2325.

Schwede, D., Zhang, L.M., Vet, R., Lear, G., 2011. An Intercomparison of the Deposition Models Used in the Castnet and Capmon Networks. Atmos Environ 45, 1337-1346.

Seinfeld, J.H., Pandis, S.N., 2006. Atmospheric Chemistry and Physics: From Air Pollution to Climate Change, 2nd ed. Wiley, Hoboken, N.J.

Sickles, J., Shadwick, D.S., 2007a. Changes in Air Quality and Atmospheric Deposition in the Eastern United States: 1990-2004. J Geophys Res-Atmos 112.

Sickles, J.E., Shadwick, D.S., 2007b. Seasonal and Regional Air Quality and Atmospheric Deposition in the Eastern United States. J Geophys Res-Atmos 112.

Tang, H., Lau, T., Brassard, B., Cool, W., 1999. A New All-Season Passive Sampling System for Monitoring No2 in Air. Field Anal Chem Tech 3, 338-345.

Tang, H.M., Brassard, B., Brassard, R., Peake, E., 1997. A New Passive Sampling System for Monitoring SO2 in the Atmosphere. Field Anal Chem Tech 1, 307-314.

Tang, H.M., Lau, T., 2000. A New All Season Passive Sampling System for Monitoring Ozone in Air. Environ Monit Assess 65, 129-137.

US EPA 1998. Emission Facts: Idling Vehicle Emissions. U.S. Environmental Protection Agency Washington, DC.

Wesely, M.L., Hicks, B.B., 2000. A Review of the Current Status of Knowledge on Dry Deposition. Atmos Environ 34, 2261-2282.

Zhang, L., Vet, R., O'Brien, J.M., Mihele, C., Liang, Z., Wiebe, A., 2009. Dry Deposition of Individual Nitrogen Species at Eight Canadian Rural Sites. J Geophys Res-Atmos 114. DOI: 10.1029/2008JD010640., D02301 (13 pp).

Chapter 3: Wet and dry deposition in the AOSR collected by ion exchange resin samplers Mark Fenn, US Department of Agriculture Forest Service, Pacific Southwest Research Station, Riverside, CA, USA

3.1 Introduction Atmospheric deposition of nitrogen (N), sulfur (S), and base cations was measured across the network of jack pine sites using ion exchange resin (IER) collectors as described in Fenn et al. 2015. Deposition was measured in forest clearings (bulk deposition) and under jack pine canopies (throughfall). Bulk deposition consists primarily of wet deposition, with a minor dry deposition component that collects onto the funnel collectors during dry periods. Throughfall deposition refers to the hydrologic flux of ions and other compounds washed from the tree canopies by precipitation or snowmelt and deposited in solution to the forest floor (Parker, 1983). Measurement of nutrient deposition in throughfall is a widely used method for estimating atmospheric deposition inputs to forest ecosystems (Bleeker et al., 2003; Thimonier, 1998).

A passive sampling method for measuring bulk deposition in forest clearings or throughfall deposition based on the deployment of ion exchange resin (IER) columns connected to a sampler funnel has been successively used in many field studies (Fenn and Poth, 2004; Root et al., 2013). The IER sampling method is very useful for deposition monitoring in remote regions because of the infrequent sampling requirements.

The objective of this study was to measure atmospheric deposition levels and to characterize spatial gradients and seasonal trends of deposition to the forest surrounding and downwind of the major industrial zone in the AOSR. A major question to be addressed by this research was to describe the spatial extent of the zone of influence resulting from atmospheric emissions of N and S from the industrial processes within the AOSR. Base cation deposition was also measured for one year, allowing a comparison of atmospheric inputs to the forest of acidifying N and S deposition versus acid-neutralizing base cation deposition. It is important to quantify base cation deposition in the AOSR because of their potential to neutralize the acidifying potential of N and S deposition. Nitrogen deposition also has the potential to cause eutrophication or other N- excess effects.

3.2 Methods Throughfall and bulk precipitation samples were collected with “passive” throughfall collectors based on a mixed bed (cation and anion resin) ion exchange resin (IER) column (Fenn and Poth, 2004; Fenn et al., 2009). Precipitation or throughfall samples are collected by a polyethylene funnel or snow tube and channeled through the resin column, where ions are retained by the ion exchange resin. The major advantage of the IER method is that sample collection continues in the field without the need for repeated field trips to collect liquid samples or the need for 40

repeated sample analyses from each collector. This is highly advantageous in the AOSR study region where many sites are only accessible by helicopter. Deposition samples were collected seasonally. The IER columns for the summer exposures were installed in May and changed out in October. Winter exposures were from October to May.

The IER columns were prepared by pouring 25 grams of the ion exchange resin beads into PVC tubes (20 cm in length and 1.25 cm I.D.) as aqueous slurry and then further rinsed with distilled water. After the field exposure periods the IER columns were extracted with 75 ml of 1N KI, followed by a second extraction with 75 ml of 1N KI. Nitrate and sulfate concentrations in the column extracts were analyzed by ion chromatography (Dionex DX-1600, Sunnyvale, CA) using a procedure modified from Simkin et al. (2004). Ammonium concentrations in the KI extracts were determined colorimetrically (Technicon TRAACS autoanalyser). Samples designated for base cation analysis received an additional extraction of 200 ml KCl. The KI and KCl extracts were proportionately combined and analyzed for Ca, Mg and Na by inductively coupled plasma atomic emission spectroscopy (ICP-AES). Atmospheric deposition fluxes were determined by extrapolating from the area of the collector opening and the amounts of inorganic N and S or base cations that were extracted from the IER columns (Fenn and Poth, 2004; Fenn et al., 2009, 2015). Spatial patterns of deposition amounts with increasing distance from the industrial center of the oil sands were determined by plotting deposition fluxes versus distance and curve fitting this relationship.

IER deposition maps were created using inverse distance weighting (IDW) with a smoothing option, using 21 data points, except for base cations in which case the smoothed IDW was performed using 19 data points. Base cation deposition was not sampled at all 21 sites where N and S were sampled. Because base cation deposition was not sampled at all the same sites in winter 2009 and summer 2010, base cation deposition was estimated for some sites. The seasonal DIN+S distance regression was used to estimate the missing winter 2009 or summer 2010 base cation deposition in throughfall. The annual value was then the sum of the estimated and the actual data. If both winter and summer throughfall base cation data were missing (e.g. Peat Pond and W1 sites), the seasonal estimates were summed to obtain an annual estimate. For the eight sites with data for both the summer and winter seasons, the actual (not estimated) data were used (Fenn et al., 2015).

3.3 Results Atmospheric deposition of N, S and base cations all showed a similar pattern of rapidly decreasing values with distance from the industrial center, particularly in throughfall (Fig. 3.1). Throughfall deposition of S and N (NH4-N + NO3-N) decreased from maximum values of 24 and 22 kg ha-1yr-1 at 3 km from the emissions sources to predicted values of 11 and 3 kg ha-1yr-1 at 20 km from the industrial center–a decrease of 56% and 88%, respectively. Nitrogen in throughfall was already reduced by 75% at only 10.5 km from the industrial center. In contrast, moderately elevated fluxes of S in throughfall extended further away from the emissions sources than N deposition, with S deposition in throughfall reduced by 80% at 84 km from the industrial center (Fig. 3.1). 41

12 30 Open May'08-May'09 Throughfall May'08-May'09 May'09-May'10 May'09-May'10 10 May'10-May'11 25 May'11-May'12 May'10-May'11 Regr May'11-May'12 Regr ) 95% CI 8 ) -1 20 -1 95% CI

-1

-1 6 15

DIN (kg ha (kg DIN 4

DIN (kg ha (kg DIN 10

2 5

0 0 0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140 Distance (km) Distance (km)

16 50

Open May'08-May'09 Throughfall May'08-May'09 14 May'09-May'10 May'09-May'10 May'10-May'11 40 May'10-May'11 May'11-May'12 12 May'11-May'12 Regr

) Regr )

-1 95% CI -1 95% CI

yr 10 1 30

-1

-S (kg -S ha- (kg 20

-S (kg -S ha (kg

4 4 6

SO 4 10 2

0 0 0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140 Distance (km) Distance (km)

2500 7000 Open DIN+S May'08-May'09 Throughfall DIN+S May'08-May'09 DIN+S May'09-May'10 6000 DIN+S May'09-May'10 DIN+S May'10-May'11 2000 DIN+S May'10--May'11 DIN+S May'11-May'12 DIN+S May'11-May'12 Base Cation Oct'09-Oct'10 5000 Base Cations Oct'09-Oct'10 DIN+S Regr

DIN+S Regr )

)

-1 -1 1500 95% CI 95% CI 4000

-1

3000 1000

(eq (eq ha

(eq ha (eq

2000

500 1000

0 0 0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140 Distance (km) Distance (km) -1 -1 - Figure 3. 1 Deposition (kg ha yr ) of dissolved inorganic N (DIN; sum of NO3-N + NH4-N) and SO4 in forest clearings and in throughfall versus distance from the industrial center of the Athabasca Oil Sands Region. -1 -1 Also shown are four years of DIN + SO4-S deposition (eq ha yr ) in forest clearings and in throughfall versus distance. On these latter two graphs, data points (as + symbols) are shown (but not a separate regression line) for one year of base cation deposition (Na + Mg + Ca).

Annual deposition of NH4-N in throughfall across the monitoring sites in the AOSR ranged widely -1 -1 from 0.8 – 14.7 kg ha compared to a range of 0.3 – 6.7 kg ha for throughfall NO3-N. Deposition -1 -1 of SO4-S in throughfall ranged also range widely from 2.5 – 23.7 kg ha yr . Deposition of NH4-N, NO3-N and SO4-S in bulk precipitation (in open, canopy-free areas) across the network ranged from 0.8 – 2.8, 0.4 – 1.6 and 1.0 – 6.8 kg ha-1 yr-1, respectively. The sum of measured base cation (Na, Mg and Ca) inputs in bulk deposition and throughfall were generally similar to or greater than inputs of acidic deposition in the form of DIN + S (Fig. 3.1). In bulk deposition, Ca was by far the most prevalent cation (57-80% of the total), followed by Mg (14-34%), and lastly Na (4-20%).

The average NH4-N:NO3-N ratios for annual bulk deposition and throughfall, considering all years and sampling locations were 1.8 and 2.0, respectively, indicating that over the monitoring network deposition of NH4-N was approximately double that of NO3-N (Fig. 3.2). The highest NH4- N: NO3-N ratios in bulk deposition during summer were found at the Peat Pond and W1 sites located 6 and 8 km. from the industrial center; winter ratios at these sites were also among the highest. However ratios in bulk deposition in winter only occasionally exceeded 2.0.

8 2.5 11 SO -S: Open NH : Open NO -N: Open 4 4 3 10 7 9 2.0 6 8

5 7 1.5

-1

-1 6 3

kg ha

kg ha 5

kg ha 1.0 4 2 3

0.5 1 2

0 0.0 0 S08 S09 S10 S11 W08 W09 W10 W11 S08 S09 S10 S11 W08 W09 W10 W11 S08 S09 S10 S11 W08 W09 W10 W11

AMS 1 AMS 14 JP101 JP107 JP304 AMS 5 AMS 15 JP102 JP201 JP309 AMS 9 Lysimeter JP104 JP210 JE310 AMS 10 Peat Pond R2 JP212 JE311 JE312 AMS 13 W1 JP106 JP213 10 30 15 SO -S: Throughfall NO3-N: Throughfall 4 NH4-N: Throughfall 14 9 25 13 6

12 5 20 7

-1

-1 6 4 -1 15

kg ha

kg ha 5 kg ha 3 4 10

3 2

2 5 1 1

0 0 0 S08 S09 S10 S11 W08 W09 W10 W11 S08 S09 S10 S11 W08 W09 W10 W11 S08 S09 S10 S11 W08 W09 W10 W11

Figure 3. 2 Temporal patterns of summer and winter deposition of NH4-N, NO3-N and SO4- S in (a) forest clearings (open) and (b) throughfall under jack pine in the Athabasca Oil Sands Region.

The IER deposition maps illustrate the rapid drop-off of NH4-N, NO3-N and their sum (DIN; Figs. 3.3) with distance from the industrial center, which is near the Lysimeter site shown on the maps. Sulfate deposition shows a larger footprint (Fig. 3.4), although sites with >12 kg S ha-1 yr-1 are generally restricted to a zone within approximately 30 km of the Lysimeter site. Of particular interest is a comparison of the deposition maps showing the combined deposition of DIN and SO4-S and base cation deposition (Fig. 3.4). It can readily be seen that estimated base cation deposition is greater than the sum of DIN + SO4-S deposition throughout the study region. However, it should be emphasized that this data set may be substandard as the base cation deposition values in throughfall are for only one year (Oct 2009 to October 2010) and base cation data were not collected at as many sites as for N and S. Estimated values, based on distance/deposition relationships (see Fig. 3.1) were used to provide missing data base cation throughfall deposition for some combinations of monitoring sites and seasons (summer or winter).

Figure 3. 3 Dissolved inorganic nitrogen (DIN; sum of NO3-N and NH4-N), NH4-N, and NO3-N deposition (kg ha-1 yr-1) in the AOSR. Maps were developed from ion exchange resin (IER) data collected from October 2008 to October 2012 from 21 monitoring sites using inverse distance weighting.

-1 -1 Figure 3. 4 Maps of SO4-S deposition in kg ha yr and DIN+ SO4-S and sum of base cations in keq ha-1 yr-1 in the AOSR.

3.4 Discussion

Deposition in throughfall of NO3-N and NH4-N were reduced by 80 and 91% at 20 km from the industrial center demonstrating the rapid drop-off with distance from the source area. In contrast throughfall deposition of base cations and SO4-S only decreased by 72 and 56% at 20 km and by 75% at 25 and 53 km, illustrating the larger footprint of SO4-S, and to a lesser degree, of base cation deposition in the AOSR.

Historical emphasis on N monitoring, modeling and effects in the AOSR has focused almost exclusively on oxidized forms of N as evidenced by N and S deposition modeling work that includes only NOx and SOx (Davies, 2012). However, our deposition data clearly show that atmospheric inputs of N in reduced forms are greater than oxidized forms. Atmospheric concentrations of gaseous NOy and NH3 in the AOSR are enhanced; however, much of the

reduced N input in the region is likely in particulate form as evidenced by stack emissions (Wang et al., 2012; Watson et al., 2011) and the steeply declining patterns of deposition with distance from the source zone.

Although deposition of NH4-N is on average double that of NO3-N deposition in the AOSR, emissions inventories of the National Pollutant Release Inventory (NPRI, 2010, 2011) report that NH3 emissions from stationary sources were only 3% as large as NO2 emissions in 2010 and 2011. Furthermore, the NPRI inventory does not include NOx emissions from mobile sources, which in 2008 made up 40% of the anthropogenic NOx emissions in the AOSR (Davies, 2012). The discrepancy between the reported dominance of NOx emissions and deposition data showing that NH4-N deposition is double that of NO3-N deposition is apparently due to unreported elevated emissions of particulate NH4 and likely also because of underestimates in reported NH3 emissions. Emissions of particulate matter mass is reported in the emissions inventory (NPRI, 2010, 2011), but chemical characterization of particulate matter emissions is not.

In summary, publicly available N emissions data for the AOSR are inadequate to address quantitatively what sources and processes are responsible for the observed atmospheric deposition in throughfall and bulk deposition. However, it is clear that emissions and atmospheric + deposition of reduced forms of N (NH3 and NH4 ) are much greater than previous understanding and available emissions inventories indicate.

3.4.1 Potential Ecological Effects of Air Pollution in the AOSR Concentrations of N in foliage of P. banksiana (jack pine) and in the lichen species Evernia mesomorpha and Hypogymnia physodes were positively correlated with atmospheric concentrations of NO2 (Laxton et al., 2010), indicating that N deposition in the AOSR enriches the N status of vegetation and lichens (Davies, 2012) in the more polluted portions of the AOSR. Likewise, a 2004 survey of the TEEM plots revealed that S concentrations in foliage of jack pine (total S and inorganic S) and of the forest floor, and both N and S levels in lichen tissue, increased with increasing atmospheric deposition (C.E. Jones & Associates Ltd., 2007). Further work is needed to evaluate possible biological responses to this N and S enrichment and its spatial extent. These findings suggest that the areas of highest potential risk of N and S deposition effects are limited to within 20-30 km of the main industrial zone of the AOSR—and possibly even a smaller zone for N effects.

Although soils in jack pine stands are naturally acidic with low base cation saturation, the consensus of previous studies in the AOSR is that there is limited potential for acidification of soils or lakes under current conditions (Hazewinkel et al., 2008; Jung et al., 2013; Whitfield et al., 2009; Watmough, this report). A field study at four forest sites in the AOSR found that soil pH increased from 2005 to 2010. The authors proposed that likely mechanisms for the soil pH increase were decreased H+ input as a result of decreasing S deposition and increased base cation deposition (Jung et al., 2013). The results of our study also indicate that base cation deposition closely tracks acidic deposition in the form of N and S deposition. Watmough et al. (2014)

concluded that despite extremely low soil base cation weathering rates in the region, the risk of soil acidification is mitigated to a large extent by high base cation deposition as shown in Fig. 4.1.

3.5 Conclusions

As noted previously for other pollutants, throughfall deposition of NO3-N and NH4-N decreased exponentially within a 20-25 km zone surrounding the industrial center (Fenn et al., 2015). At a distance of 20 km from the industrial center N deposition decreased by 88%, while S and base cations decreased by 56 and 72% respectively, showing a greater footprint. Deposition of NH4-N in the AOSR was on average double that of NO3-N. Higher deposition of reduced forms of N compared to oxidized forms had not previously been well documented except for data showing elevated atmospheric concentrations of NH3 in the AOSR (Bytnerowicz et al., 2010a,b). More studies are needed to better understand the emissions sources and chemical forms of reduced N that contribute to NH4-N deposition in the region.

Results of this study support the hypothesis that eutrophication effects to sensitive organisms such as epiphytic lichens may be of greater concern than acidification because acidic deposition is matched by equivalent amounts of buffering base cation deposition (Watmough et al., 2014). However, the zone at risk of excess N effects is limited in size as indicated by the steep decrease in N deposition with distance from the source areas.

3.6 References Bleeker, A., Draaijers, G., van der Veen, D., Erisman, J.W., Mols, H., Fonteijn, P. and Geusebroek, M. 2003. Field intercomparison of throughfall measurements performed within the framework of the Pan European intensive monitoring program of EU/ICP Forest. Environ. Pollut. 125: 123- 138.

Bytnerowicz, A., Fraczek, W., Schilling, S., and Alexander, D. 2010a. Spatial and temporal distribution of ambient nitric acid and ammonia in the Athabasca Oil Sands Region, Alberta. J. Limnol., 69(Suppl. 1): 11-21, DOI: 10.3274/JL10-69-S1-03.

Bytnerowicz, A., Schilling, S., Alexander, D., Fraczek, W., and Hansen, M. 2010b. Passive monitoring to estimate N (NO2, HNO3, NH3) exposure in remote areas and geospatial analysis to optimize monitoring networks in the Athabasca Oil Sands Region. Extended Abstract 2010-A-563- AWMA. In Proceedings of The Air and Waste Management Association 103rd Annual Conference and Exhibition, Calgary, Alberta, Canada, June 22-25, 2010.

C.E. Jones and Associates Ltd. 2007. Terrestrial Environmental Effects Monitoring: Acidification Monitoring Program. 2004 Sampling Event Report for Soils, Lichen, Understory Vegetation and Forest Health and Productivity. Prepared for: Wood Buffalo Environmental Association, Terrestrial Environmental Effects Monitoring Committee. Fort McMurray, Alberta: 198 pp.

Davies, M.J.E. 2012. Air quality modeling in the Athabasca Oil Sands Region. In Developments in Environmental Science. pp 267-309. IN K.E. Percy (Ed.) Alberta Oil Sands:Energy, Industry and the Environment. Elsevier, Oxford, UK.

Fenn, M.E., and Poth, M.A. 2004. Monitoring nitrogen deposition in throughfall using ion exchange resin columns: A field test in the San Bernardino Mountains. J. Environ. Qual. 33: 2007- 2014.

Fenn, M.E., Sickman, J.O., Bytnerowicz, A., Clow, D.W., Molotch, N.P., Pleim, J.E., Tonnesen, G.S., Weathers, K.C., Padgett, P.E., and Campbell., D.H. 2009. Methods for measuring atmospheric nitrogen deposition inputs in arid and montane ecosystems of western North America. pp 179- 228 IN A.H. Legge (ed.), Developments in Environmental Science, Vol. 9: Air Quality and Ecological Impacts: Relating Sources to Effects. Elsevier, Amsterdam.

Fenn, M.E., Bytnerowicz, A., Schilling, S.L., and Ross, C.S. 2015. Atmospheric deposition of nitrogen, sulfur and base cations in jack pine stands in the Athabasca Oil Sands Region, Alberta, Canada. Environ. Pollut. 196: 497-510.

Hazewinkel, R.R.O., Wolfe, A.P., Pla, S., Curtis, C. and Hadley, K. 2008. Have atmospheric emissions from the Athabasca Oil Sands impacted lakes in northeastern Alberta, Canada? Can. J. Fish. Aquat. Sci. 65: 1554-1567.

Jung, K., Chang, S.X., Ok, Y.S., and Arshad, M.A. 2013. Critical loads and H+ budgets of forest soils affected by air pollution from oil sands mining in Alberta, Canada. Atmos. Environ. 69: 56-64.

Laxton, D.L., Watmough, S.A., Aherne, J., and Straker, J. 2010. An assessment of nitrogen saturation in Pinus banksiana plots in the Athabasca Oil Sands Region, Alberta. Journal of Limnology, 69(Suppl. 1): 171–180.

National Pollutant Release Inventory (NPRI). 2010, 2011. Environment Canada. http://www.ec.gc.ca/inrp-npri.

Parker, G.G. 1983. Throughfall and stemflow in the forest nutrient cycle. Adv. Ecol. Res. 13: 57- 133.

Root, H.T., Geiser, L.H., Fenn, M.E., Jovan, S., Hutten, M.A., Ahuja, S., Dillman, K., Schirokauer, D., Berryman, S., and McMurray, J.A. 2013. A simple tool for estimating throughfall nitrogen deposition in forests of western North America using lichens. For. Ecol. Manage. 306: 1-8.

Simkin, S.M., Lewis, D.N., Weathers, K.C., Lovett, G.M., and Schwarz, K. 2004. Determination of sulfate, nitrate, and chloride in throughfall using ion-exchange resins. Water, Air, and Soil Pollut. 153: 343-354.

Thimonier, A. 1998. Measurement of atmospheric deposition under forest canopies: some recommendations for equipment and sampling design. Environ. Monitor. Assess. 52: 353-387.

Wang, X.L., Watson, J.G., Chow, J.C., Kohl, S.D., Chen, L.-W.A., Sodeman, D.A., Legge, A.H., and Percy, K.E. 2012. Measurement of real-world stack emissions with a dilution sampling system. pp 171-192. IN K.E. Percy (Ed.) Alberta Oil Sands: Energy, Industry and the Environment. Elsevier, Oxford, UK.

Watmough, S.A., Whitfield, C.J., and Fenn, M.E. 2014. The importance of atmospheric base cation deposition for preventing soil acidification in the Athabasca Oil Sands Region of Canada. Science of the Total Environment 493: 1-11.

Watson, J.G., Chow, J.C., Wang, X., Kohl, S.D., Gronstal, S., and Zielinska, B. 2011. Winter Stack Emissions Measured with a Dilution Sampling System. Report to the Wood Buffalo Environmental Association, Ft. McMurray, Alberta, Canada. Desert Research Institute Contract Number: T113- 10.

Whitfield, C.J., Aherne, J., and Watmough, S.A. 2009. Modeling soil acidification in the Athabasca Oil Sands region, Alberta, Canada. Environ. Sci. Technol. 43: 5844-5850.

Chapter 4: Predicted spatial variations of sulphur and nitrogen compound concentrations and deposition in the AOSR Mervyn Davies, Kanwardeep Bajwa and Reid Person, Stantec Consulting Ltd., Calgary, AB

4.1 Introduction To determine the linkage between air emissions from oil sands operations and changes in forest health, an understanding of the magnitude and the spatial distribution of ambient concentrations and deposition is required. The CALMET/CALPUFF dispersion model system was used to predict: ambient annual SO2 and NO2 concentration patterns; as well as total sulphur, total nitrogen, and potential acid input deposition patterns in the Athabasca Oil Sands Region (AOSR). The CALMET/CALPUFF model predictions were supplemented by provincial scale model predictions to include the contribution from emissions sources located outside the AOSR, and by regional ambient measurements to include compounds not simulated by the model application. The combined spatial concentrations and depositions are provided for the individual forest health monitoring sites and as contour plots superimposed over an AOSR base map.

4.2 Model approach The CALMET/CALPUFF model system application accounts for existing industrial and non- industrial emission sources located in the Lower Athabasca Region (LAR) of Alberta, of which AOSR is a subset. This model system was applied for a nominal 700 km (north-south) by 290 km (east-west) area which includes conventional oil sands mining, extraction and upgrading operations north of Fort McMurray, and in situ bitumen extraction operations as far south as Cold Lake. The CALMET model was applied using Weather Research and Forecasting (WRF) mesoscale meteorological model data for the 2002 to 2006 period complemented by concurrent surface data from three regional airports (i.e., Fort Chipewyan, Fort McMurray and Cold Lake) and nine stations from the Wood Buffalo Environmental Association (WBEA) ambient air quality monitoring network. The CALMET model accounts for terrain and land cover properties (the latter varying with land cover type and time of year).

The model simulation is based on a SO2 emission rate of 364 t/d from the LAR, with 302 t/d being from the region north of Fort McMurray. Virtually all SO2 emissions are from industrial stacks. The associated NOX emission rate from the LAR is 310 t/d, with 178 t/d being from the region north of Fort McMurray. The NOX emissions result mainly from industrial stacks and mine fleets. While emission sources from the LAR are used, the CALPUFF model was applied to predict ambient concentrations and deposition for a smaller region centered over the oil sands area north of Fort McMurray. This region is a 3 latitude (56 to 59 N) by 4 longitude (109 to 113 W) area with a north-south extent of 334 km and an east-west extent of 240 km. Further details

with respect to the emission, meteorological, and model assumptions are provided in Teck and SilverBirch (2011).

Emission sources located outside the LAR also contribute to ambient concentrations and deposition in the AOSR. As these non-LAR sources are not included in the CALMET/CALPUFF model application, the Community Multi-scale Air Quality (CMAQ) model predictions undertaken for Cumulative Environmental Management Association (CEMA) (Vijayaraghavan et al., 2013) are used to account for these sources. Specifically, the CMAQ model was applied to a 4,536 by 6,804 km area that includes all of Alberta, most of British Columbia and Saskatchewan, and five northwestern U.S. states. All emission sources in this area, excluding the LAR emissions, are included in the CMAQ simulation. The sum of the CALPUFF and CMAQ model predictions therefore includes contributions from LAR and non-LAR emission sources. Both model simulations include industrial and non-industrial emissions. The CMAQ application also includes wildfire emissions that correspond to the selected emission inventory year. The CALPUFF application does not include wildfire emissions.

The CALPUFF and CMAQ model simulations do not include all chemical compounds that are of interest for this assessment. For these compounds, ambient concentration and deposition measurements from the AOSR are added to the model predictions. Specifically, the CALPUFF application does not include reduced nitrogen compound (e.g., ammonia and ammonium) deposition; therefore ammonium deposition derived from throughfall measurements at 21 AOSR sites (Fenn, 2014) are added to the model predictions to account for this contribution. Potential acid input (PAI) is defined as the sum of acidifying sulphur and nitrogen compound contributions minus the neutralizing contribution of base cations (BC). The associated BC deposition values were also obtained from 17 AOSR throughfall measurement sites (Fenn, 2014). The SURFER objective analysis and gridding software was used to provide associated reduced nitrogen and BC deposition fields for the AOSR.

4.2.1 Predicted SO2 and NO2 concentrations

Figure 4.1 (left) shows the predicted annual average SO2 concentration contours that include LAR and non-LAR contributions. The non-LAR contribution ranges from 0.11 µg/m3 along the southern 3 boundary to 0.05 µg/m along the northern one. The SO2 concentrations predicted near the oil sands operations are dominated by the oil sands sources, with maxima greater than 10 µg/m3 occurring primarily within the respective development areas. The 20 µg/m3 contour is the annual Alberta Ambient Air Quality Objective (AAAQO) for SO2.

The predicted annual average NO2 concentration contours (Fig. 4.1, right) include LAR and non- LAR contributions. The non-LAR contribution ranges from 0.6 µg/m3 along the southern boundary 3 to 0.1 µg/m along the northern one. The NO2 concentrations predicted near the oil sands operations are again dominated by the oil sands sources, with maxima greater than 45 µg/m3 occurring primarily within the respective development areas, with 45 µg/m3 being the annual AAAQO for NO2.

The CALPUFF model simulations were undertaken using five years of hourly meteorological data and annual average concentrations for each simulation year are calculated. Appendix A4.1 provides the minimum, average and maximum annual SO2 and NO2 concentrations predicted at each of the forest health assessment sites. The average predicted SO2 concentrations range from being less than 1 μg/m3 at the more distant sites to being greater than 4 μg/m3 at the sites closer to the oil sands operations. The average predicted NO2 concentrations range from less than 1 μg/m3 at the more distant sites to greater than 10 μg/m3 at the sites closer to the oil sands operations.

4.2.2 Predicted sulphur compound deposition

The wet and dry deposition (i.e., total deposition) of sulphur compounds (i.e., SO2 and sulphate) are of interest. Figure 4.2 (left) shows the total sulphur compound deposition contours that include LAR and non-LAR contributions. The non-LAR contribution ranges from 1.3 kg S/ha/yr along the southern boundary to 0.6 kg S/ha/yr along the northern one. The sulphur compound deposition predicted near the oil sands operations are dominated by oil sands sources, with maxima of greater than 20 kg S/ha/yr occurring primarily within the respective development areas. Outside the plant sites, the predicted deposition ranges from being less than 1.5 kg S/ha/yr at the more distant sites to being greater than 5 kg S/ha/yr at the sites closer to oil sands operations. The 1 and 2 kg S/ha/yr predicted deposition values near Lake Athabasca are due to wildfire contributions. Sulphur deposition estimates for individual plot site locations are provided in Appendix Table A4.2.

4.2.3 Predicted nitrogen compound deposition - + Figure 4.2 (right) shows the total nitrogen compound (NO, NO2, HNO3, NO3 , NH4 ) deposition contours that include both LAR and non-LAR contributions. The non-LAR contribution ranges from 1.4 kg N/ha/yr along the southern boundary to 0.6 kg N/ha/yr in the northern region. The nitrogen compound deposition predicted near the oil sands operations are dominated by oil sands sources, with maxima greater than 20 kg N/ha/yr occurring primarily within the respective development areas. Outside these intensive deposition areas, the deposition ranges from being less than 1.5 kg N/ha/yr at the more distant sites to being greater than 5 kg N/ha/yr at the sites closer to the oil sands operations. Nitrogen deposition estimates for individual plot site locations are provided in Appendix Table 4.2.

Figure 4. 1 Annual average SO2 concentrations (left) in the AOSR due to LAR (CALPUFF) and non-LAR (CMAQ) emissions. Contour intervals = 0.5, 0.7, 1, 2, 3, 5, 10 and 20 (red contour) µg/m3 and annual average NO2 concentrations (right). Contour intervals = 0.5, 0.7, 1, 2, 3, 5, 10, 15, 25 and 45 (red contour) µg/m3.

Figure 4. 2 Total sulphur (left) and nitrogen (right) deposition in the AOSR due to LAR and non-LAR emissions. Contour intervals = 1, 2, 3, 4, 5, 7, 10 and 20 (red contour) kg S and 1, 2, 3, 4, 5, 7, 10, 15 and 20 (red contour) N/ha/yr.

4.2.4 Acidifying and cation deposition Total acidifying compound (S + N) deposition ranges from 0.3 keq/ha/yr along the southern boundary to 0.15 keq/ha/yr in the north (Fig. 4.3). The S + N deposition predicted near the oil sands operations are dominated by the oil sands sources, with maxima greater than 3 keq/ha/yr occurring primarily within the respective development areas.

Figure 4. 3 Predicted total sulphur + nitrogen deposition in the AOSR. Contour intervals = 0.2, 0.3, 0.4, 0.5, 0.7, 1, 2 and 3 (red contour) keq/ha/yr.

As BC deposition is not included in the model simulations, it was estimated from bulk and throughfall sample measurements collected in the AOSR for a one-year period October 2009 to October 2010. BC depositions from bulk and throughfall measurements are used to represent the lower and upper bounds of the associated BC deposition, respectively. The bulk deposition measurements represent deposition to open areas and include wet deposition and a portion of the dry deposition. Other studies have consistently found that Ca+2, Mg+2 and Na+ deposition

derived from bulk samples are typically about 1.8, 1.5 and 1.5 times, respectively, greater than those associated with wet deposition (Staelens et al., 2005). Figure 4.4 (left) shows the BC (i.e., Ca+2, Mg+2 and Na+) deposition contours based on throughfall measurements, which are assumed to represent an upper bound BC deposition. The minimum deposition of 0.251 keq/ha/yr is based on the lowest value derived from the throughfall measurements (Appendix Table 4.3). The BC deposition predicted near the oil sands operations are dominated by oil sands sources, with maxima greater than 5 keq/ha/yr. The BC deposition decreases to less than 1 keq/ha/yr distant from the oil sands sources.

Throughfall measurements represent deposition to forest canopies and include wet deposition, dry deposition, and internal leaching from the canopy. It has been common practice to adjust the throughfall Ca+2 and Mg+2 using the Na+ deposition that is not subject to internal leaching (e.g., Houle et al., 1999; Ferm, 2000). This adjustment approach, however, was not applied to the AOSR deposition measurements as the assumptions associated with the approach do not appear to be met for measurements near emission sources. Therefore, for the purposes of this assessment, the throughfall BC deposition was assumed to equal the upper bound of BC deposition to a forest canopy.

Figure 4.4 (right) shows the associated contours based on bulk measurements, which are assumed to represent a lower bound BC deposition. The minimum BC deposition of 0.081 keq/ha/yr is based on the lowest value derived from the bulk measurements (Table A4.3). The BC deposition near the oil sands operations are dominated by oil sands sources, with maxima greater than 3 keq/ha/yr. The BC deposition is typically less than 1 keq/ha/yr distant from the oil sands sources.

The high BC depositions near the oil sands activities are likely due to windblown dust emissions from land disturbances (i.e., mine haul roads and tailings pond beach areas), and these high values are consistent with a source apportionment study that found fugitive dust sources to be major contributors of metals in lichens (Landis et al., 2012).

Figure 4. 4 Base cation (BC) deposition in the AOSR based on throughfall deposition (left) and bulk (right) measurements. Contour intervals for throughfall (Upper bound BC estimate) are 0.251, 0.5, 1, 1.5, 2, 3, 4 and 5 (red contour) keq/ha/yr. Contour intervals for bulk (Lower bound BC estimate) are 0.081, 0.1, 0.2, 0.5, 1, 1.5, 2 and 3 (red contour) keq/ha/yr.

4.2.5 Deposition PAI Figure 4.5 (left) shows the total PAI deposition contours based on the model predictions and the BC throughfall measurements. Since throughfall BC deposition measurements represent a higher bound, the associated PAI deposition represents a lower bound. The PAI depositions predicted near the oil sands operations are dominated by the BC deposition and are thus basic. The area shown in pink shading is associated with predicted positive PAI values which is acidifying. There are locations on the lease site where the predicted PAI deposition is acidifying, as well as along the eastern boundary of the model area.

Figure 4.5 (right) shows the total PAI deposition contours based on the model predictions and the BC bulk measurements. As with the use of throughfall estimates, predicted PAI deposition near the oil sands operations are dominated by the BC deposition. There are locations on the lease sites where the predicted PAI deposition is acidifying as well as areas along the northern and eastern boundaries, and to the south of the oil sands area.

Appendix 4.3 shows the total sulphur, nitrogen, BC, and PAI deposition at each of the forest health assessment sites. The BC deposition and associated PAI are shown for two cases: one assuming BC based on throughfall and the other assuming BC based on bulk measurements. The PAI deposition is positive at one of the 38 plots assuming throughfall BC deposition, and positive at 12 of the 38 plots assuming bulk BC deposition.

Figure 4. 5 Predicted lower bound PAI deposition in the AOSR based on throughfall BC measurements (left); contour intervals = -2 (red contour), -1.5, -1, -0.5, -0.1, and 0 keq/ha/yr. Predicted upper bound PAI deposition in the AOSR based on bulk BC deposition measurements (right) ; contour intervals = -1.5 (red contour), -1, -0.5, -0.1, 0, 0.17, 0.22, 0.25 and 0.35 keq/ha/yr. Positive PAI values are in areas shaded pink.

4.3 Uncertainties The determination of the magnitude and spatial distribution of deposition involves several analytical steps with associated uncertainties. The first step is the application of the dispersion models that depend on the emission inventory, meteorological parameters, and land cover properties. The AOSR SO2 and NOX emission rates used in the CALPUFF simulation are estimated to be biased 9% and 1% higher than average respectively. The uncertainty associated with the year-to-year meteorological data variation is of the order 10% (from Appendix 4.1).

Explicit comparisons between maximum annual predicted and measured SO2 concentrations indicate average model over-predictions of 31% at the 13 WBEA continuous monitoring sites, and over-predictions of 14% at the 31 WBEA passive monitoring sites (Teck, 2013). These comparisons are based on an assumed background value of 1 µg/m3 being added to the model output. Without this background adjustment, there is an average over-prediction of 20% at the continuous stations and an average under-prediction of 35% at the passive sites.

Explicit comparisons between maximum annual predicted model and measured NO2 concentrations indicate average over-predictions of 57% at the nine WBEA continuous sites and average over-predictions of 169% at the 31 WBEA passive monitoring sites (Teck, 2013). These comparisons are based on an assumed background value of 2 µg/m3 being added to the model output. Without this background adjustment, there is an average over-prediction of 39% at the continuous stations, and an average over-prediction of 56% at the passive sites.

The second step is associated with model deposition assumptions. The land cover properties were derived for 4 by 4 km grid cells and are based on the dominant land cover within each grid cell. The model predictions therefore are on a broader landscape scale than on an individual forest plot basis. The model predictions do not include enhanced deposition that typically occurs along the edges of tree canopies.

The third step relates to the incorporation of ammonium and BC depositions derived from the bulk and throughfall deposition measurements. Sampling is often accompanied by uncertainties. For example, bulk samplers are in open areas and may not capture all the precipitation during high wind speeds (referred to as “undercatch”). Undercatch is not expected for throughfall as the samplers are sheltered by the tree canopy from high wind speed influences. The throughfall measurements may overstate BC from emission sources due to canopy leaching (Draaijers et al., 1997; Ferm et al., 2000). For this reason, the PAI was calculated assuming both open and throughfall BC measurements.

Lastly, the extrapolation of the ammonium and BC measurements to the AOSR as depicted in the figures has limitations due to the relatively few number of measurement sites. While it is clear that there is an oil sands contribution to both, the uncertainty associated with extrapolating the deposition rates becomes greater with increasing distance from the oil sands facilities. Fortunately, there is more confidence near the oil sands operations where the highest deposition of acid forming compounds occur.

4.4 Comments and conclusions The CALMET/CALPUFF model system was applied to predict annual average ambient concentrations and depositions of sulphur and nitrogen compounds in the Athabasca Oil Sands Region. The CALMET/ CALPUFF model system predictions were supplemented by larger scale model predictions to include the contribution from emissions sources located outside the region, and by ambient measurements to include compounds not simulated by the model application.

While high concentrations and depositions near the oil sands operations are dominated by oil sands facility emissions, the background values due to non-LAR sources become a more important contributor with increasing distance from the oil sands operations. The predicted annual SO2 and NO2 concentrations are slightly over-predicted when compared to WBEA ambient measurements. The predicted sulphur and nitrogen compound depositions are similar to other predictions (e.g., Davies, 2012).

Until this assessment, one large uncertainty associated with calculating PAI in the AOSR was the absence of regional BC cation deposition measurements. Regional BC deposition values based on bulk and throughfall measurements (Fenn, 2014) indicate significantly greater BC deposition in the AOSR than indicated by provincial scale measurements (e.g., Chaikowsky, 2001). Lower and upper bound PAI plots based on throughfall and bulk PC deposition measurements have been presented. The lower bound PAI predictions based on throughfall BC values are more likely applicable to closed forest canopies, and the upper bound PAI deposition plots based on bulk BC values are more likely applicable open areas. Further study is recommended to confirm the BC measurements that have been used for this assessment.

The resulting PAI deposition patterns are therefore quite different from those presented by others, indicating that acid forming emission contributions are completely neutralized by BC emissions, especially near oil sands mining operations. Watmough et al. (2014) in their review of the same BC data also concluded that the risk of soil acidification in the AOSR is mitigated to a large extent by high BC deposition, and that soil base saturation and pH may be expected to increase. The historical concern about regional acidification in the AOSR therefore appears to be diminished.

In conclusion, monitoring and modelling are complementary air quality management tools. Monitoring may be limited to a small number of sampling sites for a limited sampling duration, and often to a limited suite of chemical compounds. In contrast, modelling can be used to interpolate and extrapolate monitoring site data, can identify source receptor relationships, and assess future emission scenarios.

Acknowledgements The authors would like to thank Teck Resources for providing model information that were used for this assessment, to CEMA for providing CMAQ model predictions for non-LAR emission sources, and to thank Stantec for their support to this study. 61

4.5 References Chaikowsky, C.L.A. 2001. Base Cation Deposition in Western Canada (1982–1998). Prepared by Environmental Physical Sciences, University of Alberta for the Science and Technology Branch, Environmental Sciences Division, Alberta Environment. Pub. No: T/605. Edmonton Alberta. 59 pp.

Davies, M.J.E. 2012. Air Quality Modelling in the Athabasca Oil Sands Region. pp 267 - 309. IN K.E. Percy (Ed.) Alberta Oil Sands:Energy, Industry and the Environment. Elsevier, Oxford, UK.

Draaijers, G.P.J., Erisman, J.W., Van Leewen, N.F.M., Romer, F.G., Te Winkel, B.H., Veltkamp, A.C., Vermeulen, A.T., and Wyers, G.P. 1997. The impact of canopy exchange on differences observed between atmospheric deposition and throughfall fluxes. Atmospheric Environment 31. 387 -397.

Fenn, M. 2014. Chapter 3: Wet and dry deposition in the AOSR collected by ion exchange resin samplers, pp 40-50, this report.

Ferm, M., Westling, O., and Hultberg, H. 2000. Atmospheric deposition of base cations, nitrogen and sulphur in coniferous forests in Sweden – a test of a new surrogate surface. Boreal Environment Research 5. 197-207.

Houle, D., Ouimet, R., Paquin, R., and Laflamme, J. 1999. Interactions of atmospheric deposition with a mixed hardwood and a coniferous forest canopy at the Lake Clair Watershed (Duchesnay, Quebec). Can. J. For. Res. 19. 1944 – 1957.

Landis, M., Pancras, J.P., Graney, J.R., Stevens, R.K., Percy, K.E., and Krupa, S. 2012. Receptor modeling of epiphetic lichens to elucidate the sources and spatial distribution of inorganic air pollution in the Athabasca oil sands region. pp 427- 467. IN K.E. Percy (Ed.) Alberta Oil Sands:Energy, Industry and the Environment. Elsevier, Oxford, UK.

Staelens, J., De Shrivjer, A., Wan Avermaet, P., Genouw, G., and Verhoest, N. 2005. A comparison of bulk and wet-only deposition at two adjacent sites in Melle (Belgium). Atmospheric Environment 19. 7-15.

Teck Resources Limited and SilverBirch Energy Corporation. 2011. Application for Approval of the Frontier Oil Sands Mine Project (Frontier Project). Submitted to Alberta Environment, Energy Resources Conservation Board and Canadian Environmental Assessment Agency. Volume 4. Acoustics and Air. Section3. Air Quality. Pp 3-1 to 3-255. With Appendices A to F.

Teck Resources Limited. 2013. Frontier Oil Sands Mine Project Integrated Application. Supplemental Information Request, Round 2. Appendix 7a.1. Air Quality Revision (Volume 4, Section 3). Pp 3-1 to 3-78. With Attachments 7a1.A to 7a.1F.

Vijayaraghavan, K., Jung, J., Sakulyanontvittaya T., Shah, T., Hahn, R., Morris, R., Davies, M., and Person, R. 2013. CMAQ Modelling for the OMF, ADMF and NEP. Prepared for the Cumulative Environmental Management Association - Air Working Group by ENVIRON International Corporation and Stantec Consulting Ltd. 196 pp plus appendices.

Watmough, S.A, Whitfield, C.J., and Fenn, M.E. 2014. The importance of atmospheric base cation deposition for preventing soil acidification in the Athabasca Oil sands Region of Canada. Science of the Total Environment. 493. 1-11.

Chapter 5: Site selection and field methods for the Forest Health study Kenneth Foster, Owl Moon Environmental Inc., Calgary, AB

5.1 Introduction Following a review of information regarding ecological responses to acidic deposition by the TEEM committee in the period of 1995 to 1996, jack pine/lichen forests growing in sandy soils were identified as the most sensitive ecological systems in the region to acidic deposition. On this basis, the TEEM program was initiated in 1996 to establish a network of jack pine/lichen monitoring locations at varying distances and directions from the primary regional sources of acidifying emissions. The premise of the study was that soil chemistry would first be altered by acidic inputs, and then altered vegetative growth and/or health status would follow. Because chemical changes of soils in response to acidic deposition occur slowly, thus leading to delayed expression, a 6-year intensive monitoring cycle was established.

Three intensive monitoring cycles have now been completed – 1998 (AMEC, 2000), 2004 (C.E. Jones, 2006) and 2011/2012 (Table 5.1). Each cycle was preceded by activities and programs based on outcomes of the preceding cycle(s), designed to support the subsequent intensive monitoring program. Thus, the TEEM program has included activities between intensive monitoring cycles, and data relevant to the understanding of ecological responses to atmospheric emissions has been acquired in years before and after each intensive monitoring program. For clarity and simplicity, the intensive monitoring programs discussed herein are defined as occurring in 1998, 2001, 2004 and 2011/2012, however, the following activities must be considered in the interpretation of data acquired in these years:  1998: site selection and preliminary sample collection and analyses completed in 1996  2001: selection and establishment of one site (JP212; AMEC, 2002), from a pool of sites evaluated and sampled in 2000-2001 (AMEC, 2001)  2004: selection and establishment of several sites, with consideration of the site selection and evaluation program in 2000-2001  2011/2012: selection and establishment of sites based on efforts from 2007 to 2010, with the intensive monitoring program spread out over two years due to schedule adjustments required in response to a major forest fire in May to July, 2011. Some site selection and establishment efforts extended into 2013

A combination of an examination of vegetative (stand) characteristics, and preliminary evaluations of the soils on which these stands were growing, was used to select the sites. The selection of suitable stands incorporated evaluations of soils and vegetation in an iterative process. In recognition of the flow from air to soil to vegetation, the monitoring program is presented in the order of soil monitoring followed by the vegetation monitoring. However, these systems are integrated, and logistically, both systems are sampled and measured concurrently,

so the response order is not necessarily mirrored in the order of execution of the site selection and sampling programs.

Detailed procedures for sampling and laboratory analyses of samples collected in the program are presented in Foster et al. (2015). A summary of these procedures follows.

Table 5. 1 Jack Pine Site Selection Criteria Intensive Sampling Cycle Ecological 1998, 2001 and 2004 2011/2012 Component Topography  Landscape with xeric to subxeric moisture regime  Soil surface ≥2m above the level of nearby  Mid to upper slope (<5%) position bog/fen complex Soil  Mildred soil series  Mildred soil series  Poor nutrient regime  Parent materials of shallow calcareous bedrock,  Coarse glaciofluvial parent material aeolian deposits, or sand-textured glaciofluvial  Well to rapid soil drainage regime deposits Vegetation  Minimum of 30 ha stand size  Cover by characteristic species:  Jack pine trees: o 32.2% Pinus banksiana (overstory) o averaging 60 to 70 years old o 0.1% Pinus banksiana (seedlings) o without dwarf mistletoe o 3.5% Arctostaphylos uva-ursi o without excessive dead tops o 3.4% Vaccinium myrtilloides o with minimal physical damage o 52.5% Cladina mitis o of a density of 40 live trees per 400m2  Cover by differential species: o approximate 14m canopy height o 9.4% Pleurozium schreberi o approximate 9m crown length o 2.3% Vaccinium vitis-idaea o approximate 16cm DBH o 2.0% Dicranum polysetum  Cover by indicator species: o 0.3% Sheperdia canadensis o 20% Pinus banksiana (canopy cover >90%) o 0.2% Oryzopsis pungens o 0% to 20% Arctostaphylos uva-ursi o 0.1% of each of: Alnus crispa, Cornus o 2% to 20% Vaccinium myrtilloides canadensis, Elymus innovates o 20% Cladina rangiferina o 0% of each of: Amelanchier alnifolia, o <2% Hudsonia tomentosa Anemone mulifida, Cladonia gracilis, Ledum o 5% to 10% Vaccinium vitis-idaea groenlandicum, Linnaea borealis,  Cover by characteristic species: Lycopodium complanatum, Oryzopsis o 5% to 10% Pleurozium schreberi asperfolium, Picea glauca, P. mariana, o <2% to 5% Linnaea borealis Populus tremuloides, Salix spp., Sheperdia o <2% to 10% Polytricium piliferum canadensis  Cover by selected species:  Cover by common species: o ≤8% Maianthemum canadense o 5.2% Cladina stellaris o ≤1% Ledum groenlandicum o 2.5% Maianthemum canadense o ≤2% Equisetum spp. o 0.5% Rosa acicularis o 0.4% Polytrichum juniperinum Other  Sites geographically dispersed  Near a jack pine – bog/fen transition, with the  ≥500m from developments edge of the stand oriented towards regional  Not influenced by local emission sources emission sources  5 sites each in areas predicted to be receiving higher and lower levels of acidic deposition (≥0.3 and ≤0.2 kmol H+/ha/year, respectively)

5.2 Site selection and growth of the jack pine monitoring program The criteria for site selection (qualification of a jack pine stand as suitable for inclusion in the monitoring program), and the processes for establishment of plots for long-term monitoring, have evolved during the period of monitoring program operation, with updated criteria and procedures being implemented as knowledge and understanding of the jack pine systems in the region increased. The fundamental criteria that were used to guide the selection of jack pine stands for inclusion in the TEEM program are presented in Table 5.1. Initiation of the TEEM program was strongly influenced by the then Canadian Forestry Service national ARNEWS (Acid Rain Network Early Warning System) program, and several of the site-selection criteria applied within ARNEWS were incorporated into the initial criteria for selection of the TEEM jack pine stands (BOVAR, 1996). These criteria were applied from the initiation of the program in 1996 through to the completion of the second intensive sampling cycle in 2004. In 2006, the TEEM program underwent a formal evaluation, one outcome of which was an initiative to more precisely define site selection criteria according to an Ecological Analogue concept. This concept led to the definition of a Type 3 Analogue based on ecological criteria that were adopted for the selection of additional jack pine sites in 2011 through 2012.

There are a number of common criteria in the two sets of site selection criteria, although they may have been expressed differently (e.g., well drained sandy soils and soils of Aeolian genesis are expressions of the same criterion). The Type 3 Ecological Analogue criteria are generally more stringent in terms of vegetative parameters (which themselves are expressions of physiographic conditions). Concurrent with the adoption of the Type 3 criteria was a relaxation of the stand size criterion, allowing for the selection of jack pine stands into the program providing that sufficient buffer between the stand and adjacent ecological types (wetlands, other forest types) was present to properly construct the plot monitoring system. The site selection processes, and the sites selected, during each of the major jack pine monitoring initiatives are reviewed below.

5.2.1 1998 From 1995 through 1996, a review of information was completed to identify candidate jack pine stands in the region for long-term monitoring of responses to acidic deposition. Air photographs were examined to determine the location of 10 candidate stands in each of the high (≥0.3 kmol H+/ha/year) and low deposition (≤0.2 kmol H+/ha/year) zones as predicted by ADEPT2 modelling. Each of these 20 stands was surveyed and sampled, with the results of field measurements and laboratory analyses of soil, vegetation and lichen samples being used to select the initial set of 10 jack pine stands – five in each deposition zone BOVAR et al., 1996). Many of the measurements and samples collected during site selection and site suitability confirmation were adopted from the ARNEWS program. These sites were assigned labels in the form of JPHx and JPLx, where “H” and “L” indicated the location of the stand within the ADEPT2-predicted high and low deposition zones (respectively), and “x” was a number from 1 to 10 that uniquely identified each stand (Figure 5.1). In 1998, the first intensive monitoring of these 10 stands was completed.

Figure 5. 1 TEEM Jack Pine Monitoring Sites – 1998

5.2.2 2001 Recognizing that a number of sites established in 1996 to 1998 had been lost to development, and that additional sites would likely be impacted or lost due to natural (fire, insect infestation) or human causes (industrial development, road building), a limited site selection process was conducted in 2000, resulting in the recommendation for the addition of five sites to the program (AMEC Earth & Environmental, 2001). One of the recommended sites, JP212, was selected and established (AMEC Earth & Environmental, 2002). The criteria used to select the original 10 sites (1995 to 1996) were applied in 2000 and 2001. The distribution of sites in the program as of the end of the 2001 program is presented in Figure 5.2.

Figure 5. 2 TEEM Jack Pine Monitoring Sites – 2001

5.2.3 2004 To compensate for the loss of a site to industrial development, and to provide for greater regional monitoring coverage, the four remaining sites recommended in 2000 for addition to the program and a fifth recommended but not established in 1998 were added to the program in 2004. These new sites were included on the basis of their consistency with the site selection criteria applied to date. This brought the total number of sites in the program to 13 (Figure 5.3), all of which were intensively monitored in 2004.

Figure 5. 3 TEEM Jack Pine Monitoring Sites – 2004

5.2.4 2011/2012 Between 2004 and 2010, the 13 sites were examined for their ecological consistency. The sites were divided into ecological analogue groupings, based on common ecological (soil, tree, vegetation community) characteristics. It was determined that the number of sites sharing the desired ecological analogue grouping (“Type 3”) was lower than desirable. A significant effort was undertaken in 2011 to establish monitoring sites in new jack pine stands meeting the Type 3 Ecological Analogue criteria, at distances at all directions and distances from regional emission

sources. In this effort, some of the earlier criteria were revised, in particular, the restriction on stand size was removed. Providing that the monitoring plots could be established in a suitable stand in a manner that provided for a buffer of a distance equivalent to three tree heights between the plots and the stand edge (transition to a different forest type, wetland), the stand was considered to be of an acceptable size. The 25 sites in the program as of the end of 2012 are listed in Table 5.2, and their distribution shown in Figure 5.4.

Two of the recently established sites, JP308 and JP310, have been designated as “Investigative Sites” (Table 5.2). These stands are relatively small, and although they may have met the Ecological Analogue Type 3 criteria, the plots within them encroach on each other, and are not separated from stand edges by the required distances. However, Ecological Analogue Type 3 jack pine stands are sparsely distributed in the western reaches of the study area, and JP308 and JP310 are the only ones available between the emission sources and JP201. Their “Investigative Site” designation identifies them as being different from the other sites in the program; data will be examined carefully to ensure that the data are interpreted properly, so that site-specific conditions and/or plot placement decisions are considered in the analyses.

In 2011, the site naming convention was changed, the primary intent being to remove a possible perception or introduction of bias in the program associated with the pre-determination of sites being located in high or low acidic deposition zones. The replacement naming convention is based on the chronology of the site selection and establishment process, with all sites now being labelled in a consistent format that does not indicate deposition history, distance from emission sources, or any other expression of a causative factor.  the JP100 series of sites are those that were established 1996, and intensively monitored in 1998  the JP200 series of sites are those that were established in 2001 and 2004  the JP300 series of sites are those that were established from 2011 to 2013 Subsequent series (JP400 to JP900) are reserved for sites that may be established and included in the program in the future.

From mid-May to mid-July 2011, a major forest fire burned across the north and east reaches of the region, affecting a number of monitoring sites. Some sites were severely burned, some were burned primarily in the canopy, while others suffered the majority of damage at the ground level. Because fire is a natural process in the Boreal forest, the decision was made to recover these sites, and to continue monitoring on the planned cycle. Delays imposed by the fire resulted in some of the planned 2011 activities being postponed into 2012. Thus, the 2011/2012 program was a combination of site selection and site recovery, coupled with the significant intensive monitoring program.

Figure 5. 4 TEEM Jack Pine Monitoring Sites – 2011/2012

Table 5. 2 Jack Pine Monitoring Site Naming Convention as Revised in 2011 Intensive Monitoring Cycles Original New Site 2011- Comment Site Label Label 1998 2001 2004 2012 JPL1 JP101 X X X JPH2 JP102 X X X JPH7 JP103 X X X 3rd monitoring cycle after 2011 forest fire JPH4 JP104 X X X JPH6 JP105 X Lost to Development prior to 2004 JPL6 JP106 X X X 3rd cycle after 2011 forest fire JPL7 JP107 X X X 3rd cycle after 2011 forest fire JPL8 JP108 X X JPL9 JP109 X X ? Burned in 2011 – unclear if it was monitored JPH1 X Lost to development between 1998 and 2004 JP201 JP201 X X JP205 JP205 X X 2nd cycle after 2011 forest fire JP210 JP210 X X JP212 JP212 X X X JP213 JP213 X X 2nd cycle after 2011 forest fire JP303 JP303 X JP304 JP304 X JP307 JP307 X JP308 JP308 X Early warning edge site JP310 JP310 X Early warning edge site JP311 JP311 X JP312 JP312 X JP313 JP313 X JP315 JP315 X JP316 JP316 X JP317 JP317 X JP318 JP318 X

5.3 Monitoring cycle The TEEM program monitoring cycle is based on a 6-year period, with the interior stand monitoring plots being intensively sampled every sixth year. The last monitoring effort (2011/2012) was initiated seven years after the previous effort (2004), to provide an additional year in which to complete plans for a substantial expansion of the program. Additionally, the majority of the last effort was spread out over two years, due to delays imposed on the program by the extensive 2011 forest fire.

5.4 Site establishment procedures A typical jack pine monitoring site includes a vegetation plot around which are arrayed four soil plots and an off-plot tree area, and a reference stake (Figure 5.5). The reference stake, a 1.2m length of rebar pounded approximately 0.75m into the ground, allows for reconstruction of the plots at the site in the event that it burns or animals remove plot corner stakes using measurements (distances and bearings) taken at the time of site establishment.

Figure 5. 5 Typical Jack Pine Monitoring Plot Layout

5.4.1 Vegetation plot Although vegetative growth and health is dependent on soil chemistry (fertility, responses to acidic input), and selection of soil plots would appear to be a logical first step, placement of the vegetation plot containing trees that best represented the stand as a whole was considered a priority. Soils plots were placed around the vegetation plot. This prioritization of plot placements was established to ensure that minor variations in topography and other stand characteristics wouldn’t lead to substantial variation among vegetation plot trees, so as to improve the ability to detect changes in tree condition and growth.

Figure 5. 6 Schematic Jack Pine Vegetation Plot Layout

A vegetation plot measuring 10m x 40m has been established within each selected jack pine stand. The vegetation plot is established with a minimum of three tree heights (approximately 50m) distance from the stand edge, seismic lines, roads, and other disturbances. The plot area chosen is representative of the overall stand, including trees of similar age and structure as are present in the stand as a whole, at a density that is representative of the stand. The longer (40m; Y axis) of the two centre lines in the vegetation plot is designated as the “Reference Line”. The intersection of this line and the shorter (10m; X axis) of the two centre lines at the plot centre is designated as the plot origin, and assigned the coordinates 0,0 (Figure 5.6). This divides the vegetation plot into four, 20m x 5m quadrants. The quadrant that represents best the southwest quadrant is assigned coordinates in negative numbers (-y, -x), and in clockwise rotation, the NW quadrant is assigned coordinates in +y, -x format, the NE quadrant in +y, +x format and the SE quadrant in -y, +x format. Each tree within the plot is numbered using an aluminum tag, loosely attached to the tree with plastic-coated 18 or 24 gauge copper wire. Using the coordinate system, each numbered tree was mapped within the plot to the nearest 0.1m.

In 2004, an understory plant community component was added to the jack pine monitoring component of the program. Assessments of plant community composition was made using a series of 10 small subplots on the inside perimeter of the vegetation plot, two medium sized subplots also on the inside plot perimeter, and one large subplot in the centre of the vegetation plot (Figure 5.6).

5.4.2 Off-plot trees Trees of the same growth form and in an area of similar density as those in the vegetation plot are used for destructive sampling of branches and needles. Use of these off-plot trees preserves the plot trees, ensuring that the effects of branch harvest do not influence the health of trees that comprise the core of the jack pine monitoring program.

An area within the site but outside of the boundaries of the soil and vegetation plots by a minimum of 5m (tree crowns must not touch those within the vegetation plot), within which 20 or more trees of similar height, morphological structure and insect/disease infestation to those within the vegetation plot, was selected at each site (Figure 5.5). Within this off-plot tree area, 10 trees are randomly selected and tagged. In the 1998, 2001 and 2011/2012 sampling cycles, the tagged trees within the off-plot tree area were used for destructive sampling and branch measurements. A departure from this method occurred in 2004, to focus on trees maximally exposed to air emissions. In this cycle, five dominant and co-dominant trees were chosen from the stand, and these trees sampled.

5.4.3 Soil pit – soil characterization In the year that the site is established, the soil at the site is classified on the basis of the soil pedon exposed in and samples taken from a pit of approximate 1m x 1m dimensions, dug to the depth of the C horizon. The pit is dug a minimum of 10m from the vegetation plot, and 5m from any of the soil plots. The location for the soil pit is determined at the completion of plot layout and staking, ensuring that it does not interfere with trees within the vegetation or soil plots. The pit location should avoid hummocks, depressions or other unique site characteristics.

Samples of the LFH horizon are taken prior to digging the soil pit. Samples by pedogenic horizon are taken from the faces of the soil pit, combining material from the same horizon sampled from three of the four pit faces. The soil pedon exposed in the soil pit is then described in sufficient detail that, together with the results of the laboratory analysis of pit samples, the soil can be classified into the appropriate subgroup of the Canadian System of Soil Classification (SCWG, 1998), and to the appropriate soil map unit. The soil pit is refilled after sampling, and pit location is clearly noted on the site drawings.

5.4.4 Soil plots Around the vegetation plot at distances of no less than 10m, four soil plots were established. While a 40m x 10m plot is preferred, where site conditions (site non-uniformity, slope, drainage features, etc.) negate the ability to establish a plot(s) with this configuration, one of two acceptable alternate layouts is used (Figure 5.7). Soil plots of 10m x 30m and 7.5m x 30m (not shown on Figure 5.7) are acceptable when the standard plots cannot be established due to non- uniformity within the stand (e.g., depressions, proximity to stand edges). Plots of these smaller dimensions were only established in the 2011/2012 cycle, a reflection of the relaxation of the stand size criterion that had been applied in the earlier sampling periods.

5.5 Measurements and indicators 5.5.1 Vegetation plot sampling – jack pine morphometric measurements Vegetation monitoring procedures in the 6-year cycle component apply in the year of plot establishment, and in every subsequent cycle. Morphometric data (Appendix 5.1) are collected from each numbered tree within the vegetation plot during each monitoring cycle. Optimally, routine vegetation monitoring activities are conducted in the latter half of August, allowing for the health assessment and sampling of current annual growth (CAG), while reducing the potential for interfering physiological responses relating to the onset of autumn conditions. Vegetation monitoring activities are completed by September 15.

5.5.2 Off-plot tree sampling The use of off-plot trees that are morphologically and physiologically similar to the trees within the vegetation plot permits destructive sampling, examination of branch and needle growth and condition, and sampling and analysis of foliar tissues from trees representative of the stand, without affecting the health and condition of trees in the marked vegetation plot used for growth measurements and health assessments.

Figure 5. 7 Soil Plot Layouts at Jack Pine Monitoring Sites

For the 1998 monitoring cycle, in the establishment of JP212 in 2011, and for the 2011/2012 monitoring cycle, 10 off-plot trees representative of the trees within the stand were selected in

a loose grouping to one side of the vegetation plot (Figure 5.5). Selected trees were tagged using numbered aluminum tags attached to the trunk at about eye level using plastic-coated copper wire.

A different approach to selection of off-plot trees was applied in 2004. Height and DBH were measured on the four largest-diameter, undamaged, dominant/co-dominant trees per plot (vegetation plot, four soils plots); a sample size of 20 trees. These trees were marked and tagged, and branches sampled from the upper third of the canopy on the side facing the regional emission sources.

Branches were excised from the upper third of the canopy of the off-plot trees, on the side facing the regional emissions sources. Measurements of internode growth and needle retention were made for the current annual growth (CAG), and for each preceding year to the internode that is at least 7-years-old (older internodes are measured if available). In the first monitoring cycle (1998) and in the 2001 sampling of JP212, a branch was obtained from each of the 10 off-plot trees. Needle retention and condition assessments and internode measurements (to 0.1cm) for the current year internode and at least the preceding four (to the 4-year-old age class) were completed, and to the 7-year-old age class if possible. Following these assessments and measurements, needles were sampled for chemical analyses. Needle sampling was conducted between August 15 and September 15, by which time current annual growth (CAG) will be complete, physiological responses to the onset of autumn conditions (night frost) will be minimal, and the extent and magnitude of insect or pathogenic microorganisms will be less than expected later in the year.

Each soil plot is divided into four subplots measuring 10m x 10m (Figure 5.7), 7.5m x 10m, or 7.5m x 7.5m depending on main plot size. Soil plots are numbered from 1 to 4 (i.e., S1 to S4), and the four subplots within each soil plot are numbered (e.g., S1-1 to S1-4 for subplots in soil plot S1). The four plots, each divided into four subplots, provide for sufficient sample numbers to support the statistical analysis of the chemical analysis data at a level of sensitivity that allows for a detection of deposition effect(s). This is consistent with the requirements of the Acid Deposition Management Framework (CEMA, 2004).

In 1998, samples of needle age classes (CAG, Age-1, Age-2) from each of 10 trees were collected combined into a single, composite sample (Appendix 5.2). In 2001 (site JP212 establishment), needles from each of the age classes from each branch were separately sampled. In 2004, a branch from each of five trees was excised, and needles from the same age class (CAG, Age-1, Age-2) being combined from all branches to create three composite samples, one for each age class, per site. In 2011/2012, CAG, Age-1 and Age-2 age classes from each of the five off-plot trees were separately sampled (no compositing). Similarly, laboratory analyses have varied among sample cycles, both in the analyses performed and the instruments used. Within each sampling cycle a number of duplicate needle samples were obtained during field sampling. These varied from cycle to cycle, with the 2011/2012 having the most rigorously defined criteria for duplicate sampling.

Within each sampling program, a chain-of-custody form was completed in the field. Each party receiving a sample set retained a copy of this form, and one copy remained with each sample set. This process ensured that all sample sets were shipped and delivered without loss of samples, and any concerns relating to an anomalous analysis (or analyses) could be examined in the context of the treatment each sample set received (sampling, storage, shipping, processing, etc.).

5.5.3 Vegetation community composition Changes in soil chemistry and subsequent changes in vegetation growth and health may result in changes to the relative competitive ability of species currently growing at the jack pine monitoring sites, resulting in changes in species composition over time. In 2004, an understory plant composition component was initiated. The percent cover of vascular and non-vascular in each of 10 small, two medium and one large subplots within the vegetation plot at each site was estimated. In the 2011/2012 program, the same small, medium and large subplots were assessed. In 2011/2012, the percent cover assessment was supplemented with the Daubenmire (1959) and Coulloudon et al. (1996) methods of cover class estimation.

In addition to the quantitative cover data collected within each of the vegetation subplots, a “standard random walk” was conducted in 2011/2012 at each site to identify the presence of species not observed within the subplots. This “standard random walk” is conducted prior to activities at the site that may cause substantive disturbance to ground cover vegetation (e.g., branch excision). The “standard” component of the walk is measured in time as the walk is to be conducted for a period of 30 minutes. The “random” component of the walk relates to the absence of a specific survey pattern or trail that must be followed. The walk is to cover the entire jack pine monitoring site, avoiding stand edges where transition from the jack pine ecological analogue forest type gives way to other vegetation types and soil conditions. This is a presence/absence survey only; quantitative cover or abundance data are not required.

5.5.4 Annual forest health assessment In the early stages of the Forest Health Monitoring Program, an assessment of the health of each of the jack pine monitoring sites was made annually (AMEC 2000a, 2000b, 2002a, 2003). The purpose of this annual assessment program was to identify and record the disease, insect and mechanical damage within a plot, recent tree mortality, identification of the cause of death, and general forest health in proximity to the monitoring plots. These data were collected in support of the understanding of stresses and their causes that may underlie effects observed or measured during a 6-year monitoring cycle. The annual forest health assessment survey program was suspended in 2002. In 2011, the annual evaluation of the health of the forests at each jack pine site recommenced, with AESRD forest specialist(s) conducting the assessments.

5.5.5 Soil sampling Soil sampling has evolved from a stratified design to one based on random sampling over the course of the study. In 1998, 2001 and 2004, soil sampling locations were defined by their position under the tree canopy. For each subplot within each soils plot, a random coordinate pair

was derived, and the tree in the subplot closest to the coordinates was identified. From that tree, working outward from the trunk in the direction of the emission sources, a sampling location in each of three precipitation zones was selected (Figure 5.8), and at this location samples were taken by depth:  Stemflow zone (SF): o the area receiving precipitation flowing down the trunk from the upper canopy o 0 to 0.2m from the trunk o LFH, 0-5cm and 5-15cm depths sampled  Throughfall zone (TF): o the area receiving precipitation interacting with the canopy o 0.2 to 1.5m from the trunk o LFH, 0-5cm, 5-15cm, 15-30cm and 30-50cm depths sampled  Freefall zone (FF): o receiving precipitation unaffected by interaction with the canopy o in an area with no overhanging canopy o LFH, 0-5cm and 5-15cm depths sampled The distance required to achieve sampling in an FF zone, and occasionally in the TF zone, meant that some samples were acquired from outside of the subplot and/or plot boundaries.

This stratified sampling procedure was used in 2004, except that:  samples associated only with subplots 1 and 3 in each main soils plot were sampled  the SF zone was eliminated from the program  in the FF zone at sites established in 1998 the 5-15cm sampling depth was eliminated (this depth was sampled at newly established sites)

Prior to the 2011/2012 sampling campaign, soil chemistry data acquired from 1998 through 2004 were analysed to determine if an effect of precipitation zone on soil chemistry was apparent. None was found. For the 2011/2012 soil monitoring cycle, a sampling procedure based on random selection of sampling points contained within the defined subplots was implemented, ensuring that samples were taken within defined plots, and to ensure that samples were taken under varying degrees of canopy cover, thus representing the effect of varying canopy cover within the stand. Soils were sampled (by depth) at one randomly selected location of nine defined locations within each soil subplot (six defined locations in 10m x 7.5m and 7.5m x 7.5m subplots), with specific allowable adjustments to accommodate the interference imposed by a tree, stump, deadfall or prior disturbance. The sampling locations are illustrated in Figure 5.9 for each of the plot sizes.

No Sampling within 1 m of Canopy Perimeter

Freefall Zone Throughfall Zone

20 cm Stemﬂow Trunk Zone

Tree Canopy

Figure 5. 8 Stemflow, Throughfall and Freefall Soil Sampling Zones

Figure 5. 9 Soil Sampling Locations (as defined in 2011/2012)

Compositing of individual samples into a single sample for analysis results in significantly reduced laboratory costs at the expense of lower statistical power. On the other hand, analysis of individual replicates is more expensive but results in data permitting an evaluation of within-site variability, a more statistically robust approach. The soil sampling program has evolved from a heavy reliance on compositing (1998) to one supportive of more robust statistical analyses by eliminating compositing altogether (2011/2012). The number of samples acquired, the compositing approaches used, and the final number of samples available for laboratory analyses, are presented in Appendix 5.3.

Within each sampling cycle a number of duplicate soil samples were obtained during field sampling. These varied from cycle to cycle, with the 2011/2012 having the most rigorously defined criteria for duplicate sampling. In this recent cycle, one full set of duplicate samples was obtained from one randomly selected subplot per site. Duplicate sampling and the use of the data from lab analyses of duplicate samples are discussed within the Soils chapter.

5.5.6 Soil sample analyses As with sampling intensity, and the evolution from composting samples for analyses to the analysis of individual samples, laboratory analyses have evolved through the program. This 81

laboratory evolution reflects improvements driven by increased understanding of the soils and their responses to acidic inputs, as well as improvements in individual analyses predominantly driven by technological advances.

5.6 Summary and conclusions The WBEA TEEM program began in 1996 with the evaluation of a number of jack pine stands for their suitability for monitoring the chemical and biological effects of acid deposition. In 1998, 10 suitable stands were chosen, and a standard set of plots and tree-sampling locations were established within each stand, and a series of soil samples, jack pine tissue samples, lichen samples and tree growth and morphological measurements were taken. These procedures were largely based on the then-operational national Acid Rain Network Early Warning System monitoring program. In 2001, another monitoring site was selected and established, and in 2004, the program grew to encompass 13 monitoring sites.

A review of the ecological similarity and differences among the 13 monitoring sites led to the identification of a set of criteria for the selection of additional sites. These criteria, defining a “Type 3 Ecological Analogue” (Jaques and Legge, 2012), were used to guide the selection and establishment of 12 additional monitoring sites distributed around and at varying distances from the industrial emissions sources. The third intensive sampling and measurement program at these 25 sites was conducted primarily in 2011, however, due to a large forest fire in the region that caused delays in program execution, some sampling and measurement was extended into 2012.

Sampling and monitoring procedures have been adjusted in each cycle to reflect advances in technical methods, and to adapt sampling strategies in response to the results of the previous sample and measurement programs. In 2004 a community-level assessment of vegetation composition was initiated and continued into 2011/2012. In addition to the core sampling programs (soil, jack pine needles, lichens), several investigative programs have been conducted, including those examining the potential responses of soil microbiology to acid inputs. With the completion of three cycles of sampling and measurement, it is now possible to conduct trend analyses related to ecological responses to acid input, a primary objective of the TEEM program.

5.7 References AMEC Earth & Environmental Ltd. 1998. Vegetation Stress Survey in the Vicinity of the Syncrude and Surrounding Oil Sands Leases, August 1996. Prepared for Syncrude Canada Ltd., Fort McMurray, AB. 34 pp.

AMEC Earth & Environmental Limited. 2000. Monitoring the Long-term Effects of Acid Emissions on Soil and Vegetation in Jack Pine and Aspen Forest of Northeast Alberta – 1998 Annual Report. Submitted to the Wood Buffalo Environmental Association.

AMEC Earth & Environmental (2000a). 1999. Forest Health Assessment of the Jack Pine Soil Acidification Monitoring Plots. Submitted to the Wood Buffalo Environmental Association, Fort McMurray, AB. 6 pp.

AMEC Earth & Environmental (2000b). 2000. Forest Health Assessment of the Jack Pine Soil Acidification.

AMEC Earth & Environmental Limited. 2001. Jack Pine Acid Deposition Monitoring Network. Site Selection 2000. Prepared for the Wood Buffalo Environmental Association.

AMEC Earth & Environmental Limited. 2002. Establishment of Site JP212 as a TEEM Jack Pine Acid Deposition Monitoring Site. Submitted to the Wood Buffalo Environmental Association Terrestrial Environmental Effects Monitoring Committee.

AMEC Earth & Environmental Ltd. (2002a). 2001. Forest Health Assessment of the Jack Pine Soil Acidification Monitoring Plots. Submitted to the Wood Buffalo Environmental Association, Fort McMurray, AB. 31 pp.

AMEC Earth & Environmental Ltd. (2002b). Vegetation Stress Survey in the Vicinity of the Syncrude and Surrounding Oil Sands Leases, August 2001. Submitted to the Wood Buffalo Environmental Association, Fort McMurray, AB. 50 pp.

AMEC Earth & Environmental Ltd. (2003). 2002. Forest Health Assessment of the Jack Pine Soil Acidification Monitoring Plots. Submitted to the Wood Buffalo Environmental Association, Fort McMurray, AB. 37 pp.

BOVAR Environmental, Landcare Research & Consulting Inc., AGRA Earth and Environmental. 1996. Environmental Effects of Oil Sand Plant Emissions in Northeastern Alberta. Regional Effects of Acidifying Emissions. 1996 Annual Report. Prepared for the Environmental Effects Subcommittee of the Wood Buffalo Regional Air Quality Coordinating Committee.

CE Jones and Associates Ltd., AMEC Earth & Environmental Ltd., Gentian Botanical Research. 2007. Terrestrial Environmental Effects Monitoring Acidification Monitoring Program – 2004 Sampling Event Report for Soils, Lichen, Understory Vegetation and Forest Health and Productivity. Prepared for the Wood Buffalo Environmental Association Terrestrial Environmental Effects Monitoring Committee.

CEMA. 2004. Recommendations for the Acid Deposition Management Framework for the Oil Sands Region of North-Eastern Alberta. Prepared by the Cumulative Environmental Management Association, NOx/SOx Management Working Group. 39 pp.

Coulloudon B., Eshelman K., Gianola J., Habich N., Hughes L., Johnson C., Pellant M., Podborny P., Rasmussen A., Robles B., Shaver P., Spehar J., Willoughby J. 1996. Sampling Vegetation Attributes. Interagency Technical Reference, Cooperative Extension Service, U.S. Department of Agriculture, U.S. Department of the Interior. Revised in 1997 and 1999. Technical Reference 1734-4.

Daubenmire R. 1959. A canopy-coverage method of vegetational analysis. Northwest Sci. 33:43- 64.

Foster K.R., Baines, D., Percy K., Legge, A., Maynard D., Chisholm V. 2011. Forest Health Monitoring Program, 2011 Procedures Manual. Prepared for the Wood Buffalo Environmental Association Terrestrial Environmental Effects Monitoring Program, Fort McMurray, AB. 288 pp.

Jaques, D.R. and Legge, A.H. 2012. Ecological analogues for biomonitoring industrial sulfur emission in the Athabasca oil sands region, Alberta, Canada. pp 219-241. IN K.E. Percy (Ed.) Alberta Oil Sands:Energy, Industry and the Environment. Elsevier, Oxford, UK.

Soil Classification Working Group. 1998. The Canadian System of Soil Classification (third edition). http://sis.agr.gc.ca/cansis/references/1998sc_a.html.

Chapter 6: Plot tree layer description Ellen Macdonald, Department of Renewable Resources, University of Alberta, Edmonton, AB

6.1 Introduction Atmospheric nitrogen (N) and sulphur (S) deposition can cause acidification - reducing forest growth and health through direct damage to trees or indirect impacts through mobilization of toxic elements, such as aluminum, or imbalances in soil nutrient availability (Sogn and Abrahamsen, 1998; Savva and Berninger, 2010; Greaver et al., 2012). Sites where weathering of base cations is not sufficient to counterbalance acidification effects of N and S deposition will be most sensitive to the effects of such deposition (Greaver et al., 2012). Alternatively, N and S deposition can result in a fertilization effect causing increased tree growth, particularly in nitrogen-limited systems (Derome et al., 2009; Solberg et al., 2004; 2009; Greaver et al., 2012), and potentially increasing carbon sequestration (but see Gundale et al., 2013).

For the purposes of monitoring impacts of atmospheric deposition associated with oil sands activities on forests, the TEEM (Terrestrial Environmental Effects Monitoring) program chose to focus on dry, nutrient poor jack pine (Pinus banksiana) sites, as these were expected to be the mostly highly sensitive to nutrient inputs and acidification associated with these depositions (Foster, Chapter 5, this report). Locations of sites were selected across the anticipated zones of deposition and indeed this has now been verified (Davies et al., Fenn, Chapters 3 and 4 in this report).

Herein we summarize information on the tree component of the vegetation at the TEEM monitoring sites, including examination of possible relationships to zones of atmospheric deposition.

6.2 Methods The sites were all classified as the a1.1 Pj/bearberry/lichen ecosite phase (Beckingham and Archibald, 1996) and were typical of this forest type. The understory vegetation was heavily dominated by ground lichens – primarily Cladina mitis – with low cover and diversity of vascular plants. See Macdonald (Chapter 9, this report) for further information on understory vegetation.

During the 2011/12 monitoring program tree data were collected from 25 sites. These included six sites which burned in the 2011 wildfire (see Foster this report Table 5.2). For JP106 and JP213 the fire damage was less severe and thus it was possible to collect a full set of tree and branch data for them; JP103, JP107, JP109, JP205 were severely burned and thus only data on tree density and height were obtained for them and the trees were not categorized as to dominance class.

Data on tree density, size (height, diameter at 1.3m (breast) height), crown depth (length from top of tree to base of the crown), and crown closure (in classes: 0 = not touching adjacent trees; 1 = touching on 1 side, 2 = touching on 2 sides, 3 = touching on 3 sides, 4 = touching on 4 sides) were collected for all trees in the 10m x 40m vegetation plot (see Foster, Chapter 5, this report ). Each tree was classified as to its position in the canopy (1 = above the general canopy height; 2 = at the general canopy height; 3 = just below canopy height; 4 = entirely below canopy height). For analyses, trees in dominance classes 1 and 2 were combined and called ‘canopy trees’ while those in dominance classes 3 and 4 were combined and called ‘sub-canopy trees’.

Data on branch growth and defoliation were collected from 10 off-plot trees. For one clipped branch from each of these 10 trees per site, internode length was measured for each year separately from the current year to 4- or 5-years old. Visual estimates of percent defoliation (in classes: 0 = no defoliation, 1: ≤ 25%, 2: ≤ 50%, 3: ≤ 75%, 4: ≤ 100% defoliation) were made on the branches, for each year segment separately.

The 10 off-plot trees were cored on 16 plots (JP106, JP201, JP210, JP213, all JP300-series plots) and a few off-plot trees were cored in JP102, JP108 in the 2011/2012 sampling. Trees at the other sites had been cored in previous sampling campaigns (Amec Earth & Environmental Ltd., 2000; Jones & Associates, 2007). Trees were cored at 1.3 m (breast) height. Cores were dried, mounted and carefully sanded prior to being measured. Ring counts from these cores – with 10 years added to account for growth to 1.3m height - were used as the basis for an estimate of stand age for each site.

Summary statistics were calculated to describe tree density, size, crown depth and canopy closure (for canopy trees and sub-canopy trees separately) and branch growth and defoliation (overall for the current year to 4-years old and also for each year separately) for each site. I then examined relationships of these variables with a set of atmospheric deposition variables of two different types: 1) predictions based on source emissions as derived from the CALMET/CALPUFF dispersion model (See Davies et al., Chapter 4 this report); and 2) predictions modelled on the basis of ion exchange resin data collected at a network of sites in the region (see Fenn, Chapter 3 this report). Both types of predictor variables were examined because they are based on different data and analyses and they were not always strongly correlated with one another. Prior to analyses we did not know which might be more strongly related to the response variables of interest. I also examined possible relationships with distance from the source of emissions (from Fenn, Chapter 3 this report). Because there was a high degree of correlation among the different CALPUFF variables and among the different IER variables, a subset of each for these analyses. The subset was chosen by examining correlations among the deposition variables and from groups of highly correlated variables the one most strongly correlated with most response variables was chosen for subsequent analyses. See Table 6.1 for a complete list of tree and branch and deposition variables.

Individual linear regressions were conducted for each of the 14 tree and branch variables plus stand age versus each of the 12 deposition variables. For each of these, residuals were examined

for normality, homogeneity of variance and to look for evidence of non-linearity. No transformations were required to meet these assumptions. All regression analyses were conducted in R. version 3.1.1 using the “lm” procedure (R Core Development Team, 2014). This analysis was exploratory in that testing was for any indication of a relationship between the deposition variables and the measured forest structure attributes. For four of the “response” variables (canopy and sub-canopy height, diameter at breast height, age) a real cause effect relationship, was not considered with the deposition variables since these attributes are the product of periods of growth that preceded any industrial development in the region. Rather, we investigated whether there was any confounding of variation in these aspects of forest structure with the deposition gradient; these could, in turn, influence other terrestrial response variables. For these reasons, any single regression for which p < 0.05 was considered to be of potential interest, despite the probability for inflation of Type I statistical error due to 12 different regressions being conducted for each response variable.

Table 6. 1 (A) Deposition; and (B) tree and branch variables that were tested. CALPUFF deposition variables are from Davies et al. (Chapter 4, this report); IER deposition variables and distance from source are from Fenn (Chapter 3 this report). (A) Deposition variables Abbreviation in Table 6.2 3 1 CALPUFF SO2 concentration (µg/m ) CALPUFF S conc 3 1 CALPUFF NO2 concentration (µg/m ) CALPUFF N conc CALPUFF S deposition (keq/ha/yr)2 CALPUFF S deposition CALPUFF N deposition (keq/ha/yr)2 CALPUFF N deposition CALPUFF Base Cations (throughfall) (keq/ha/yr)3 CALPUFF BC (TF) CALPUFF PAI (throughfall BC) (keq/ha/yr)3 CALPUFF PAI (TF BC) CALPUFF Base Cations (bulk) (keq/ha/yr)4 CALPUFF BC (bulk) CALPUFF PAI (bulk BC) (keq/ha/yr)4 CALPUFF PAI (bulk BC) Distance from source (km) Distance from source IER SO4-S deposition (bulk) (kg/ha/yr)5 IER S deposition (bulk) IER DIN deposition (bulk) (kg/ha/yr)5,6 IER DIN deposition (bulk) IER Base Cation deposition (bulk) (keq/ha/yr)5 IER Base Cations (bulk) (B) Tree and Branch variables Density (# trees per ha) canopy trees (dominance class 1 or 2) sub-canopy trees(dominance class 3 or 4) Diameter at 1.3 m(breast) height (cm) canopy trees (dominance class 1 or 2) sub-canopy trees(dominance class 3 or 4) Height (m) canopy trees (dominance class 1 or 2) sub-canopy trees(dominance class 3 or 4) Crown depth (m) canopy trees (dominance class 1 or 2) sub-canopy trees(dominance class 3 or 4) Crown closure canopy trees (dominance class 1 or 2) (0 = none; touching on 1, 2, 3, or 4 sides) sub-canopy trees(dominance class 3 or 4) Internode length (cm) Overall for current yr to 4 yr-old separate by age cohort: current yr, 1-, 2-, 3-, 4-yr old Defoliation in branch segment overall for current yr to 4 yr-old (0 = none, 1 ≤25%, 2 ≤50%, 3 ≤75%, 4 ≤100%) separate by age cohort: current yr, 1-, 2-, 3-, 4-yr old 1 Annual average; 2Total for all sources; 3Throughfall assumption for Base Cations; 4Bulk assumption for Base Cations; 5Bulk deposition; 6Dissolved Inorganic Nitrogen

6.3 Results and discussion The 25 stands varied in age from 55 to 90 years (Appendix 6.1). All were completely dominated by jack pine (one plot had one trembling aspen (Populus tremuloides) sub-canopy tree) with low canopy tree density and few sub-canopy trees. Canopy tree densities varied from an average of 175 (JP315) to 1325 stems per hectare while average sub-canopy tree densities varied from 0 to 425 stems per hectare. Tree diameters at 1.3 m (breast) height varied from an average of 13 to 28 cm (JP315) while average heights of the canopy trees varied from 10.0 to 18.6 m (Appendix 6.1). Since JP315 had a substantially lower density and higher dbh than other stands, the regressions for these variables were run with and without that site. This made no difference in the significance of the regressions. Average crown depth varied from 4.8 to 14.9 m for canopy trees and from 3.0 to 8.7 m for sub-canopy trees. This represented from about 32 to 80% of tree height for both crown classes. Trees were fairly open grown with the average crown closure rarely being above 2 (representing touching adjacent trees on two sides) (Appendix 6.1).

Annual branch growth varied from an average of 2 to 7.5 cm (Appendix 6.2). There was little loss of needles; for the most recent three years’ branch segments, average defoliation values were only once greater than 2 (representing up to 50% defoliation). The four-year-old branch segments had higher defoliation with averages varying from 1.6 to 3.6 (Appendix 6.2).

There were a few instances in which the tree and branch variables were significantly related to the deposition variables (Table 6.2). The relationships had low to moderate strength with R2 values varying from 0.199 to 0.332. Diameter at Breast Height of canopy trees was positively related to CALPUFF Base Cations (both throughfall and bulk) and negatively related to CALPUFF Potential acid Input (throughfall assumption for Base Cations) (Table 6.2). The strongest relationship was with CALPUFF Base Cations (throughfall) (Fig. 6.1). Height was significantly positively related to CALPUFF S and N concentration and deposition and all three IER deposition variables while being negatively related to distance from source (Table 6.2). The strongest relationship was with IER Base Cations (Fig. 6.1). Height of the sub-canopy trees was also significantly, positively related to CALPUFF N concentration and all three IER deposition variables (Table 6.2).

Branch internode length was significantly, negatively related to CALPUFF Potential Acid Input (bulk base cation assumption) (Table 6.2, Fig. 6.2) while branch defoliation was positively related to CALPUFF S concentration and CALPUFF Base Cations (throughfall) (Table 6.2, Fig. 6.2).

There were significant relationships between stand age and deposition variables (Table 6.2). Age was negatively related to CALPUFF Potential Acid Input (Throughfall Base Cation assumption) and positively related to CALPUFF Base Cations (bulk) (Table 6.2, Fig. 6.3). This raises the question of whether significant relationships of tree and branch variables with deposition variables could be an artefact of differences in stand age. The four stands with the lowest CALPUFF PAI (throughfall) values (to the left in Fig. 6.3) were JP104, JP303, JP307 and JP317. Only one of these (JP104) matches to the four points to the high (right) end of the graph of height versus IER Base Cations (Fig. 6.1); it had the highest average height but was actually no older than the overall average of

all stands (75 versus 74, Appendix 6.1). There were no particularly notable points in the graph of age versus CALPUFF Base Cations (bulk) (Fig. 6.3). The youngest (JP311) and oldest (JP317) stands had low and high values (respectively) for this deposition variable but they did not have the lowest/highest values (respectively) for height, diameter at breast height, or branch internode length (Appendix 6.1 and 6.2).

Table 6. 2 Results of regressions analyses of tree variables on the various deposition variables. Given is the significance (P), the R2 value, and the estimate for the slope of the regression. For the reasons explained in the methods any regression for which p < 0.05 was considered to be of potential interest CALPUFF variables are from Davies et al. (Chapter 4, this report); IER variables and Distance from source are from Fenn (Chapter 3, this report). See Table 6.1 for full names of variables. Grey shading indicates p > 0.05. For the following response variables none of the regressions had p < 0.05: Density, Crown Depth or Crown Closure (for all trees or for canopy or sub-canopy trees), Diameter at Breast Height for sub- canopy trees; Branch Internode Length or Branch Defoliation (overall for current year to 4 years old, separately for 1-, 2-, 3-, 4-years old); (see also Appendices 6.1 and 6.2). (A) Diameter at Breast Height Height Height (canopy trees) (canopy trees) (sub-canopy trees)

P R2 slope P R2 slope P R2 slope CALPUFF S conc 0.012 0.304 0.898 CALPUFF N conc 0.009 0.323 0.207 0.017 0.292 0.21 CALPUFF S deposition 0.019 0.269 13.19 CALPUFF N deposition 0.029 0.238 6.158 CALPUFF BC (TF) 0.024 0.253 2.038 CALPUFF PAI (TF BC) 0.033 0.229 -2.356 CALPUFF BC (bulk) 0.033 0.228 2.496 CALPUFF PAI (bulk BC) Distance from source 0.017 0.276 -0.025 IER S deposition (bulk) 0.008 0.331 1.409 0.022 0.272 0.95 IER DIN deposition (bulk) 0.008 0.332 0.975 0.023 0.270 1.38 IER Base Cations (bulk) 0.008 0.332 5.429 0.022 0.271 5.30 (B) Branch internode length Branch Defoliation Age P R2 slope P R2 slope P R2 slope CALPUFF S conc 0.041 0.202 0.226 CALPUFF N conc CALPUFF S deposition CALPUFF N deposition CALPUFF BC (TF) 0.043 0.199 0.430 CALPUFF PAI (TF BC) 0.043 0.1664 -8.647 CALPUFF BC (bulk) 0.014 0.2361 10.383 CALPUFF PAI (bulk BC) 0.024 0.242 -1.175 Distance from source IER S deposition (bulk) IER DIN deposition (bulk) IER Base Cations (bulk) 89

Figure 6. 1 Regressions of Diameter at 1.3m (Breast) Height and Tree Height for canopy trees (dominance classes 1 and 2) and versus deposition variables. Each point represents a site and the line shows the regression. CALPUFF Base Cations (Throughfall assumption) is from Davies et al. (Chapter 4, this report). IER Base Cations (Bulk) is from Fenn et al. (Chapter 3, this report). See also Table 6-2.

Figure 6. 2 Regressions of Branch Internode Length and Defoliation (both for current year) versus deposition variables. Each point represents a site and the line shows the regression. CALPUFF Potential Acid Input (PAI) with a bulk Base Cation assumption and CALPUFF SO2 concentration are from Davies et al. (Chapter 4, this report). See also Table 6-2.

Figure 6. 3 Regressions of Stand Age (in 2011) and versus two different deposition variables: CALPUFF Potential Acid Input (Throughfall Base Cation assumption) and CALPUFF Base Cations (bulk); (both from Davies et al. (Chapter 4, this report)). Each point represents a site and the line shows the regression. See also Table 6-2.

In agreement with these results, the 1998 TEEM monitoring program showed decreasing tree height, diameter at 1.3m (breast) height and crown length with increasing distance from the source of emissions. Following attempts to normalize these growth data to account for the influence of factors such as site, climate and stand age these relationships disappeared (Amec Earth & Environmental Ltd., 2000). However, the analysis of tree rings from the 10 plots sampled in that campaign suggested that annual diameter increment prior to the start of emissions was greater at sites closer to the source of emission but that this difference disappeared after the start of emissions (Amec Earth & Environmental Ltd., 2000). Results from that campaign also suggested that needle loss was greater in high emission areas (Amec Earth & Environmental, 2000) and this agrees with our findings.

Data from the 2004 TEEM monitoring program suggested that tree diameter and height were positively related to Potential Acid Input (PAI) (Jones & Associates, 2007). However, these relationships were driven by three sites with the highest predicted deposition – JP102, JP104 and JP 212. For this reason there was uncertainty as to whether a cause and effect relationship could be inferred between atmospheric depositions and tree growth (Jones & Associates, 2007). In the 2004 monitoring program, no significant relationship was found between Site Index (height at age 50) and PAI (Jones & Associates, 2007). It is not clear whether the influence of base cations on Potential Acid Input was accounted for in 2004. Notably, our results imply a negative relationship of tree diameter with Potential Acid Input (as calculated using the CALPUFF model, Davies et al., Chapter 4 this report). Further, our results were not driven by these three as they fell across the range in terms of tree diameter and while JP102 is now still predicted to have high PAI, JP212 is predicted to have moderate levels and levels at JP104 are predicted to be low (based on the CALPUFF model, Davies et al., Chapter 4, this report).

Previous studies have found variable effects of nitrogen (N) and sulphur (S) deposition on tree growth. Experimental additions of nitrogen and sulphur to nutrient-poor, coarse-textured soil Scots pine stands in Sweden resulted in increased tree growth; however associated leaching losses of macronutrients were a concerning side-effect (Derome et al., 2009). Experimental additions of sulphur and nitrogen to forests in Norway increased height growth of Scots pine saplings (Sogn and Abrahamsen, 1998) while atmospheric deposition of nitrogen was implicated in increased growth of pine and spruce forests in southern Norway (Solberg et al., 2004). In boreal Scots pine forest in Eurasia, evidence of long-term growth declines and increases were attributed to S and N deposition, respectively (Savva and Berninger, 2010).

Jung and Chang (2012) found that experimental additions of nitrogen to mesic sites of medium richness in the Athabasca Oil Sands Region increased growth of trembling aspen (Populus tremuloides), suggesting these forests are nitrogen limited. This is likely even more so for the dry, poor pine forests being monitored by TEEM. However, Jung et al. (2013) compared two sites at different distances from emission sources in the Athabasca Oil Sands Region and found no differences in basal area increment of jack pine.

The relationships between tree variables and deposition of nitrogen, sulphur, and base cations suggest, if anything, a positive effect of deposition on growth. However, tree diameter, height and age cannot be taken to directly reflect the influence of deposition since these are a reflection of establishment and growth which preceded the development of oil sands mining in the region. The branch variables, which would reflect only recent effects of deposition, suggested a negative effect of deposition; i.e., declining internode length with increasing Potential Acid Input and increasing defoliation with increasing S concentration. This contrasts with previous studies in the region which provided evidence for a positive correlation between atmospheric NO2 concentrations and sulphur deposition with jack pine foliar concentrations of nitrogen and sulphur, respectively (Laxton et al., 2010; Jones & Associates, 2007).

It is difficult to make firm conclusions regarding the effects of atmospheric deposition on growth of trees in these sites because of the potential confounding effects of stand age, climatic differences, or atmospheric enrichment of CO2 due to industrial sources (Jung et al., 2012). If there are positive effects of deposition on tree growth the role of base cations in buffering any potential acidification effects of N and S deposition (Watmough et al., 2014; Davies et al., Chapter 4) is likely important. With the elimination of negative effects of acidification, N and S fertilization effects might occur (Greaver et al., 2012) or there could be a fertilization effect of the base cations themselves. A detailed analysis of tree diameter growth – with proper cross-dating and de-trending for age-related trends in growth – would allow for standardized calculation of growth during the time period before emissions began and comparison of that with growth in recent years. This analysis, which is currently in progress, is critical to factor out potential pre-emissions differences among sites and to make conclusions about ecological effects of deposition.

6.4 References AMEC Earth & Environmental Ltd. 2000. Monitoring of the long-term effects of acid emissions on soil and vegetation in Jack Pine and Aspen forest of northeast Alberta – 1009 Annual Report. Wood Buffalo Environmental Association CE01742/7000.

Beckingham, J.D. and J.H. Archibald. 1996. Field guide to ecosites of Northern Alberta. Canadian Forest Service, Northwest Region, Northern Forestry Centre. Special Report 5.

Derome, J., A. Saarsalmi, and M. Kukkola. 2009. Effects of nitrogen and sulphur “stress” treatment on soil acidity and growth response of a Scots pine stand. Boreal Environment Research 14: 861-874.

Greaver, T., T. Sullivan, J. Herrick, M. Barber, J. Baron, B. Crosby, M. Deerhake, R. Dennis, J.-J. Dubois, C. Goodale, A. Herlihy, G. Larence, L. Liu, J. Lynch, K. Novak. 2012. Ecologcial effects of nitrogen and sulfur air pollution in the US: what do we know? Frontiers in Ecol. Env. 10: 365-372.

Gundale, M., F. From, L. Bach, and A. Nordin. 2013. Anthropogenic nitrogen deposition in boreal forest has a minor impact on the global carbon cycle. Global Change Biology 20: 276-286.

Jones, C.E. and Associates Ltd. 2007. Terrestrial Environmental Effects Monitoring: Acidification Monitoring Program: 2004 Sampling Event Report for Soils, Lichen, Understory Vegetation and Forest Health and Productivity. Prepared for: Wood Buffalo Environmental Association, Terrestrial Environmental Effects Monitoring Committee. Fort McMurray, Alberta: pp 858.

Jung, K. and S. Chang. 2012. Four years of simulated N and S depositions did not cause N saturation in a mixedwood boreal forest ecosystem in the oil sands region in northern Alberta, Canada.

Jung, K., Choi, W.-J., Chang, S., and Arshad, M. 2013. Soil and tree ring chemistry of Pinus banksiana and Populus tremuloides stands as indicators of changes in atmospheric environments in the oil sands region of Alberta, Canada. Ecol. Indicators 25: 256-265.

R Core Development Team. 2014. R: A Language and Environment for Statistical Computing R Foundation for Statistical Computing. Vienna, Austria. www.R-project.org.

Savva, Y., and Berninger, F. 2010. Sulphur deposition causes a large-scale growth decline in boreal forests in Eurasia. Global Biogeochemical Cycles 24: GB3002, doi:10.1029/2009GB003749.

Sogn, T., and Abrahamsen, G. 1998. Effects of N and S deposition on leaching from an acid forest soil and growth of Scots pine (Pinus sylvestris L.) after 5 years of treatment. Forest Ecology and Management 103: 177-190.

Solberg, S., Andreassen, K., Clarke, N., Tørseth, K., Tveito, O.-E., Strand, G.-H., and Tomter, S. 2004. The possible influence of nitrogen and acid deposition on forest growth in Norway. Forest ecology and Management 192: 241-249.

Watmough, S.A, Whitfield C.J., and Fenn, M.E. 2014. The importance of atmospheric base cation deposition for preventing soil acidification in the Athabasca Oil sands Region of Canada. Science of the Total Environment. 493. 1-11.

Chapter 7: Soil biology in the AOSR using phospholipid fatty acid analysis Jacynthe Masse, Carolyn Churchland and Sue J. Grayston, Department of Forest and Conservation Sciences, Faculty of Forestry, University of British Columbia, Vancouver, BC

7.1 Introduction Soil organisms are fundamental to a plethora of processes in soil, from decomposition of organic matter to nutrient cycling, together with the physical, structural development of the soils themselves (Paul, 2007). Soil microorganisms are sensitive to environmental conditions and are susceptible to disturbance (van Bruggen and Semenov, 2000; Berga et al., 2012). Due to their fast turnover rates, soil microorganisms react to changes in environmental conditions in soil long before plants or other biota (Horz et al., 2004; Berga et al., 2012). Therefore, studying soil microbes in the context of disturbances can be a useful tool to detect early signs of ecosystem perturbation.

Phospholipid fatty acid (PLFA) analysis assesses broad-scale microbial community structure and is thought to be the most powerful approach for demonstrating general changes in soil microbial community composition (Grayston et al., 2004; Ramsey et al., 2006). PLFA analysis is able to discriminate several major microbial groups: actinomycetes, arbuscular mycorrhizal fungi, fungi, Gram-positive and Gram-negative bacteria. PLFA analysis is also quantitative, and able to discern relative abundance of the above mentioned broad-scale microbial groups.

Actinomycetes are a type of Gram-positive bacteria commonly found in soils that play an important role in decomposition of lignin, cellulose and chitin (Paul and Clark, 1996). Gram- positive bacteria are typically considered k-strategist bacteria, with longer lifespans and longer turnover times than Gram-negative bacteria. Gram-negative bacteria (r-strategists) are the dominant bacteria associated with plant roots (rhizosphere) (Paterson et al., 2007) and many are beneficial to plant growth. There is one PLFA characteristic of all fungi, including saprotrophs (involved primarily in decomposition) and mycorrhizae (involved in tree-root symbiosis, and nutrient uptake). There is also one PLFA representative of arbuscular mycorrhizal fungi, which are symbiotic fungi associated with many plant roots; however, this PLFA has also been found in some bacteria. Although variations in these broadly categorized microbial groups cannot provide direct information on changes in functions or biogeochemical processes (i.e. nitrogen cycling), they are sensitive bio-indicators and can be used to recommend areas for further research.

7.2 Methods 7.2.1 Experimental design and laboratory analysis In August 2011, we collected four random sub-samples of soil within a 4m2 area from the top 0- 5-cm of mineral soil at each of four locations adjacent to each of four large replicate soil plots at each of the Forest Health monitoring plots. We used a 5cm diameter soil auger for collection of the mineral soil layers. For each sampling location we composited the four replicate soil samples in double plastic bags. In addition, at each site, we took eight random soil cores, thoroughly mixed them and then subdivided the soil into two equal portions. This replicate was used as quality control for the analyses.

All soil samples were stored on ice in cool boxes in the field and then at 4oC in the laboratory for a short time until shipping. The soils were sieved through a 2-mm mesh immediately on return to the laboratory. Samples were freeze-dried in preparation for phospholipid fatty acid (PLFA) analysis. Phospholipid fatty acids were extracted according to the Frostegård et al. (1991) method, based upon the Bligh and Dyer (1959) procedure, and further modified by White et al. (1979). Briefly, soil samples (1.2 g of freeze-dried soil) were vortex-extracted in a 0.8:1:2 (v/v/v) solution of citrate buffer, chloroform and methanol. The extracted lipids were then fractionated into neutral lipids, glycolipids and phospholipids on Accubond II Solid Phase Extraction silica columns (Agilent Technologies Inc., Santa Clara, CA) by elution with chloroform, acetone and methanol. A known amount of methyl nonadecanoate (19:0) was added to the fraction containing the phospholipids to act as an internal standard. Lipids were then transmethylated to their fatty-acid methyl esters using mild alkaline methanolysis. Following this, fatty-acid residues were flash-evaporated under N2 gas and stored at -20°C until analysis.

PLFA peaks were identified by means of a combination of mass spectra and retention times relative to the internal standard 19:0, and an external bacterial acid methyl-ester standard (BAME; Sigma-Aldric., 47080-U, Oakville, ON, Canada). PLFA nomenclature identifies the number of C atoms in the fatty-acid chain (e.g. 18 in 18:1ω7), the number of double bonds in the chain (e.g. 1) and the position of the first double-bonded C from the methyl end of the fatty-acid molecule (e.g. ω7). Abundance of identified fatty acids is expressed as nmols per gram of freeze- dried soil (nmols g-1). The following fatty acids were chosen to represent bacterial PLFAs: i15:0, a15:0, 15:0, i16:0, 16:1ω7c, i17:1ω8c, 10Me17:0, i17:0, a17:0, 17:O, 18:1ω7c, 18:1ω5c, 18:O, cy19:0 (Frostegård et al., 1993; Kroppenstedt, 1985; Zogg et al., 1997). The branched PLFAs i15:0, a15:0, i16:0, i17:0, and a17:0 were considered to be indicative of Gram-positive bacteria, and 16:1ω5c, 16:1ω7c, 16:1ω9c, 16:1ω5c, 17:1ω8c, 18:1ω5c, 18:1ω7c, cy17:0 and cy19:0 of Gram- negative bacteria. The branched PLFAs 10Me17:0, 10Me18:0 and cy10:0 were classified as actinomycetes. The remainder of the bacterial PLFA were classified as “other bacteria”. There was one fungal biomarker PLFA, 18:2ω6, 9; and one for arbuscular mycorrhizae, 16:1ω5c.

Data on atmospheric deposition, vegetation and soil characteristics in AOSR were provided by co-contributors and are presented in previous chapters of this report (Chapters 4, 5, this report)

7.2.2 Statistical analysis First, a set of univariate statistics were executed in order to examine if proximity to the industrial center affected the microbial community composition and abundance. Sites were split into three categories based on distance from the edge of applied in-situ projects and industry centers: within 40 km, between 40 and 80 km and more than 80 km. Those within 40 km include sites: JP102, JP103, JP104, JP109, JP212, JP303, JP304, JP307, JP310, JP311 and JP315. Sites between 40 and 80 km include: JP101, JP106, JP107, JP210, JP308, JP312, JP313 and JP316. The sites that were greater than 80 km away from applied in situ projects include: JP108, JP201, JP205 and JP213. Several sites were burned in recent forest fires: JP103, JP106, JP107, JP109, JP205, and JP213. All univariate statistical analysis was conducted using JMP system release 11 (SAS Institute Inc., Cary, NC). Data was analyzed using a 2-way ANOVA, followed by a post-hoc Tukey HSD test.

Secondly, multivariate spatial and canonical analyses were done in order to assess for the relationship among microbial communities, space, atmospheric deposition, vegetation, and soil characteristics. Multivariate spatial analysis was performed using a principal coordinates of neighbour matrices method (PCNM), whereas relationships among communities, deposition, vegetation, and soil characteristics were discerned with redundancy canonical analysis (RDA) (Borcard et al., 2011). Statistically significant differences were α<0.1. Multivariate statistics were performed using R (R core team, 2013).

7.3 Results 7.3.1 Proximity to the industrial centre and burned sites Fungi, Gram-negative bacteria and bacterial:fungal ratio are significantly affected by the distance to the industrial centre (Table 7.1). On the other hand, forest fire significantly influenced the abundance of arbuscular mycorrhizae, Gram-positive and total bacteria. Figure 7.1 shows that fungal abundance was greatest in sites that were furthest from the industrial center. Possible explanations for these results will be discussed in the next section.

Table 7. 1 Two-way analysis of variance of broad-scale PLFA groups. Only statistically significant results and strong trends (α<0.1) were included. Variable factor df F p Fungi Distance 2 3.4729 0.035 Arbuscular Fungi Forest Fire 1 3.4667 0.068 Bacteria:Fungi Ratio Distance 2 3.1409 0.047 Total Bacteria Forest Fire 1 3.1040 0.081 Gram-Positive Bacteria Forest Fire 2 5.6362 0.019 Gram-Negative Bacteria Distance 1 2.7237 0.07

Figure 7. 1 Mean fungal abundance (nmols/g dry weight of soil) over three distances: 1 = 0-40 km from the industrial center, 2 = 40-80 km from the industrial center, and 3= 80+ km from the industrial center. Vertical error bars represent standard error.

7.3.2 Soil microbial community spatial structure Principal Coordinates of Neighbour Matrices is a special case of Moran’s eigenvector maps and allows one to detect spatial structure in data. The basic idea behind this method is to first create a set of maps. These maps can be seen as waves connecting sampling sites. The first map (or PCNM) will generate waves where distance between two succeeding wave-crests will be at the maximum. These are used to describe broad-scale spatial structure. The last map will generate waves where distance between two succeeding wave-crests will be at the minimum and describe fine-scale spatial structure. All the PCNM between will be on that gradient. The second steps will be to do a canonical ordination with soil microbial communities as response variables and the PCNM as explanatory variables. PCNM techniques are able to discriminate if there is spatial structure in the data and, if so, at what scale this spatial structure is operating. It is an appropriate multivariate spatial analysis when faced with irregular sampling design spread over a large area (Brocard et al., 2011). Figure 7.2 shows the results of the PCNM run on the PLFA analysis.

Figure 7. 2 Principal Coordinates of Neighbour Matrices on PLFA analysis of microbial communities (p<0.01).

The PCNM analysis retained two significant spatial maps: PCNM 3 and 4. Figure 7.2 shows grouping of sites based on the position of each site on these two PCNM. This ordination highlights two main groups of elements. The first one (JP107, 109, 205, 106 and 102) is composed of fire- impacted sites. The second group (JP 103, 307, 104, 212 304, 213, 108, 313) comprises sites closer to the mining sites together with sites southeast of the mining centre. Other sites are grouped two-by-two (JP201 and 308; JP 101 and 311; JP210 and 316; JP303 and 312) according to their geographical positions on the map. Sites JP315 and JP310 are standing alone.

Figure 7.3 shows the same ordination in a biplot format. This RDA shows roughly the same two main groups of sites: the fire-impacted group of sites and the sites closer to the mining centre and Southeast of them. Some of the two-by-two sites are plotted together (JP101 and JP311; JP201 and JP308). However, JP303 and 312 as well as JP 210 and 316 are plotted with the mining- impacted sites. JP310 and 315 are still standing alone. This ordination allows us to see which sites are enriched or depleted in certain communities. Axis 1 shows a gradient from Gram-negative, Gram-positive, actinomycetes and arbuscular mycorrhizae communities. It is possible to see that sites JP101 and 311 have the highest abundance of Gram-negative and Gram-positive bacteria followed by site JP 310. Fire-impacted sites have higher abundance of actinomycetes and arbuscular mycorrhizae. Sites JP201 and 308 have the highest abundance of arbuscular mycorrhizae, fungi and other bacteria. Mining impacted sites have the lowest abundance of all communities.

Figure 7. 3 RDA biplot - scaling 1 showing relationships between sites, abundance of microbial communities and PCNM axis (p<0.01)

By regressing the spatial canonical axis on the environmental data, it was possible to determine which variables were related to each PCNM. Table 7.2 shows the significant variables for PCNM 3 as well as their coefficient. Analyses were made on standardized data. Soluble Mg, pH and PAI all have a negative correlation with PCNM 3. Soluble Fe has a positive coefficient with PCNM 3. This means that sites that are along this axis have less soluble Mg, low pH, low PAI and high concentration of soluble Fe. PCNM4 was related to longitudinal geographical coordinates.

Table 7. 2 Species presenting a significant relationship with PCNM 3 – *** P value<0.01 Species Coefficient F value P value Soluble Fe 1.33 -14.81 *** Soluble Mg -0.934 -11.08 *** pH -0.425 -30.73 *** PAI -0.209 5.317 ***

7.3.3 Soil microbial community relationships with soil chemical characteristics, atmospheric deposition and vegetation Redundancy canonical analysis (RDA) was used to identify relationships between soil communities’ abundances and soil chemical properties, atmospheric deposition and vegetation coverage at each site. RDA allows its user to assess and test relationships between two datasets: one comprising response variables and the other, the environmental variables. Figure 7.3 shows the RDA biplots in scaling 2 of the RDA done on microbial PLFA abundance, soil chemical characteristics and atmospheric deposition controlling for spatial coordinates. Scaling in the context of ordination refers to the way the ordination results are projected in the “reduced

100

space” of the biplot. Biplots are a projection in 2D of a multi-dimensional cloud of data. Just like in geographical mapping, there is no single way to optimally display objects and variables and keeping all the relationships among them together in the reduced space of the biplot. Two main type of projection are generally used in this kind of analysis: scaling 1 and scaling 2. Distances among objects (observations – sampling sites) in biplots drawn in scaling 1 are an approximation of their Euclidean distances in the multidimensional space while the angles among the descriptor vectors are meaningless. On the other hand, distances among objects in the biplots drawn in scaling 2 are not an approximation of their Euclidean distances in multidimensional space, however the angle between descriptors reflects their correlation (Legendre and Legendre, 1998).

The overall RDA presented in figure 7.4 and its canonical axes 1 to 3 are significant (p<0.01). The environmental variables explain up to 67.11% of the overall variability in microbial communities’ abundances. Total nitrogen and carbon in soil and PAI are significant variables in the ordination + (p<0.1). Axis 1 appears to be related to soil chemical characteristics (especially N, P, NH4 and cation exchange capacity (CEC)) explains 61.04% of this variation. Axis 2 which could be explained by deposition (nitrogen and sulphur compound deposition), accounts for 11.72% of the variation. Finally, axis 3 accounts for 7.55% of the variation in microbial communities’ abundances and is related to potential acid input (PAI).

This RDA highlights the negative correlation between nitrogen compound deposition and sulphur + compound deposition with soil characteristics such as N, NH4 and CEC. The N and S depositions are also opposed to Gram-positive, Gram-negative and arbuscular mycorrhizal communities meaning that sites with higher N and S depositions, have lower abundance of these microbial communities. Not surprisingly, NO3 seems to be perfectly correlated with nitrogen deposition. The use of the axis 3 shows that fungal communities are positively correlated with PAI.

101

Figure 7. 4 RDA biplots - scaling 2 showing relationships between abundance of microbial communities and soil chemical characteristics and atmospheric deposition. The top RDA shows axis 1 and 2, and the bottom RDA shows axis 1 and 3

102

Figure 7. 5 RDA triplot - scaling 1 showing relationships between abundance of microbial communities and soil chemical characteristics and atmospheric deposition.

Figure 7.5 shows the scaling 1 of the same RDA. This scaling can be used to study grouping of sampling sites when considering altogether the microbial communities abundances, soil chemical characteristics and atmospheric deposition. Axis 1 is discriminating between two groups of sites. The first group, on the right-hand site of the figure compromises sampling sites JP304, JP212, JP102, JP315, JP312 and JP201 which are almost exclusively sites closer to the industrial centre and are characterized, in this RDA, by high sulphur and nitrogen deposition and low abundances of Gram-positive, Gram-negative and arbuscular mycorrhizae. On the left hand side of the axis 1 is another group of sampling sites (JP101, 311 and 107) which are characterized by low deposition and high microbial community abundances.

Figure 7.6 illustrates the links between the same microbial communities (PLFA) and the soil characteristics, the atmospheric deposition and adding variables on the vegetation coverage at each site. Since we had fewer sites with information on all these variables, some variables not important in the ordination, were left out in order to have enough degrees of freedom to be able to test the relationships between the variables.

103

Figure 7. 6 RDA biplots - scaling 2 showing relationships between abundance of microbial communities and soil chemical characteristics and atmospheric deposition. The top RDA shows axis 1 and 2, and the bottom RDA shows axis 1 and 3

The RDA and canonical axes 1 to 3 are significant (p<0.01). This time, the environmental variables (including vegetation) included in ordination explain up to 84.6% of the variability in the microbial communities’ abundances. Total nitrogen and total carbon in soil are the only significant variables in the ordination (p<0.1). Axis 1, which is still related to soil chemical characteristics (especially N and pH), explains 62.72% of this variation. Axis 2 explains 15.07% of the variation and seems to

104

be driven by PAI. Finally, axis 3 explains 4.69% of the variation in microbial communities’ abundances with no clear drivers.

The second ordination confirms relationships seen in the first RDA. Nitrogen compound deposition is still negatively correlated to N content in the soil and to Gram-positive, Gram- negative bacteria and to arbuscular mycorrhizae and positively correlated to NO3 content of the soil. However, this ordination also illustrates a positive relationship between PAI and coverage of bryophytes on sampling sites, which are both positively correlated with abundance of fungi and other bacteria at the sampling sites. Actinomycetes abundance is also positively correlated to lichen coverage. Surprisingly, vascular plants coverage is negatively correlated to Gram-positive, Gram-negative, arbuscular mycorrhizae as well as with N and C content of the soil. However, axis 3 shows a positive correlation between nitrogen compound depositions and vascular plants coverage, suggesting a possible fertilization effect of nitrogen deposition on vascular plants.

Figure 7.7 shows this last RDA with scaling 1. Axis 1 seems to be driving the grouping of the sampling sites. On the right-hand side of the RDA we again find sites JP304, 212, 102, 315, 312 and 108. Axis 2 seems to discriminate site JP 308 from the rest of the sites, based on high abundance of fungi and other bacteria and high PAI deposition at this site.

Figure 7. 7 RDA triplot - scaling 1 showing relationships between abundance of microbial communities and soil chemical characteristics, atmospheric deposition and vegetation coverage.

105

7.4 Discussion 7.4.1 Spatial structure of microbial communities The PCNM analysis showed a significant spatial structure to the microbial communities found in soils from the different sites. Broadly, we were able to identify 2 main clusters. One group contains mainly sites that are closer to the industrial centre and on the path of regional dominating wind (southeast of the mining centre). They are also the sites with lower N and C content, highest deposition and lower abundances of soil microbial communities. Another group comprises sites that have been recently naturally disturbed by fire and the rest of the groups encompass sites with less PAI, more abundant microbial communities and the highest lichen coverage and they are grouped together according to their geographical position.

When the same analysis (the PCNM) was performed with the soil biological data collected by Visser in 2004 from the same region (Visser, 2006), no spatial structure was detected. There may be three possible reasons for this observation. First, it could be that the effect of deposition on soil microbial communities was not strong enough to be detected in 2004. However, seven years later effects were significant enough to be detected. Alternatively, it could also be that the biological variables used in 2004 (total microbial biomass using chloroform fumigation, soil respiration, ectomycorrhizal root tip characterization, soil faunal abundance and diversity) were not sensitive enough to detect differences in communities from site to site. Thirdly, in 2004 there were fewer sampling sites (13) and most were close to the industrial centre, so all sites at that time may have been under the influence of the industrial activity and, therefore, all biological communities equally affected. Visser (2006) concluded that in 2004, although the forest floor chemistry appeared to be changing as a result of acid deposition, the mineral soil was unaffected giving a strong indication that this region of the soil profile is in the early stages of impact from acid forming emissions and no measureable adverse effects on soil biological properties were apparent. Our results from the 2011 sampling suggest that we are now seeing the effects of acid forming emissions on microbial communities.

7.4.2 Impacts of deposition on soil microbial communities Our analysis showed a negative relationship between deposition (both nitrogen and sulphur) and soil microbial communities mediated through soil chemical characteristics (especially soil N and C content). This suggests that a site with high levels of deposition will have a lower N and C content and a lower abundance of microbial communities compared to sites with low deposition levels. The analysis also shows that there is a different response of bacteria and fungi to deposition. Although, like bacteria, fungi were also negatively correlated with nitrogen and sulphur deposition, fungal abundance seemed to be more related to PAI. Taken together these correlations suggest that there is a general decrease in all microbial communities combined with a shift from fungi towards bacteria dominating closer to the industrial centre. Bacteria are essential for nutrient cycling in soils. A decrease in the abundance of communities is concerning. However, further research is needed to know which specific communities were affected in order to make a conclusion on whether or not an ecological function was altered in these soils. Fungi are integral to plant health as they form symbiotic relationships with plant roots, are involved in

106

soil organic matter decomposition and are key regulators of soil nutrient cycling. Whether the decrease in fungi towards the industrial center is a decrease in all fungi uniformly, or specifically is a decrease in symbiotic or saprotrophic fungi is unclear, and requires further research. Changes in ectomycorrhizal abundance will impact tree growth, and seedling survival. Conversely, a decrease in saprotrophic fungi could be detrimental to soil health, as it would decrease decomposition, limit nutrient cycling and slow soil formation. This could hinder future growth of forest communities in these areas.

Vegetation did not seem to have a strong impact on soil microbial community structure, with two exceptions. Firstly, bryophyte cover seems to be highly related to fungal abundance in soil. Whether this relationship is mediated through PAI deposition or not, is impossible to determine in the analysis performed. Secondly, the presence of lichen was strongly associated with actinomycete abundance in soils; a relationship previously reported in the scientific literature (Gonzalez et al., 2005).

7.4.3 Impacts of natural forest fire on soil microbial communities Arbuscular mycorrhizal fungi and total bacteria both showed a strong trend (Table 7.1) towards greater abundance in sites that were burned. Similarly, Gram-positive bacterial abundance was significantly greater in burned sites than in unburned sites. This is a surprising result as lower microbial abundance is typically observed in burned sites (Bååth, 1995 – Scots pine; Campbell et al., 2008 – wet sclerophyll sites; Switzer et al., 2012 – Douglas fir). However, Gundale et al. (2005) found no difference between burned and unburned ponderosa pine sites after 2 years, and there is limited information on the impact of forest-fires on microbial community structure (Switzer et al., 2012). Both arbuscular mycorrhizal fungi and Gram-positive bacteria produce spores, which are highly resistant to environmental perturbations, which may explain their higher abundance in burned sites. Figure 7.3 showed a highest abundance of actinomycetes in these sites. Actinomycetes are a type of Gram-positive bacteria commonly found in soils that play an important role in decomposition of lignin, cellulose and chitin (Paul and Clark, 1996). It is not surprising to find them in high abundances in fire-impacted sites since there is plenty of wood to decompose. Microbial community abundances from the burned sites were also grouped together in the PCNM reinforcing this indication of a strong impact of forest fire on soil microbial communities.

7.5 Conclusions This indicator identified relationships between soil microbial community abundances, soil chemical characteristics, atmospheric deposition and vegetation coverage in AOSR. Our results demonstrate a negative relationship between deposition (both nitrogen and sulphur) and soil microbial communities mediated through soil chemical characteristics (especially soil N and C content). Although, both bacterial and fungal communities were negatively correlated with nitrogen and sulphur deposition, fungal abundance seemed to be more related to PAI and fungi were much less abundant near the industrial centre. In order to determine the consequences and effect of changes in the soil microbial community, further study is necessary.

107

At present there are few sampling sites that are further than 80 km from the industrial center. A greater sampling range (including sites to the south of the industrial center) would establish if this phenomenon occurs consistently throughout the landscape. Furthermore, it is important that new analyses examine soil fungal and bacterial community diversity. Specifically, future work could include molecular analysis to establish which groups of fungi (be it mycorrhizal or saprotrophic) are decreasing in those sites near the industrial center. Such research will give insight into the functional changes to soil biogeochemistry. This molecular analysis could be combined with further nutrient analysis or enzymatic analysis to establish if the sites closer to the industrial center are experiencing any loss in soil function or changes in biochemical cycling.

7.6 References Bââth, E., Frostegard, A., Pennanen, T., and Fritze, H. 1995. Microbial community structure and pH response in relation to soil organic-matter quality in wood-ash fertilized, clear-cut or burned coniferous forest soils. Soil Biology & Biochemistry. 27 (2): 229-240.

Berga, M., Székely, A.J., and Langenheder, S. 2012. Effects of Disturbance Intensity and Frequency on Bacterial Community Composition and Function. PLoS ONE 7(5): e36959. doi: 10.1371/journal.pone.0036959.

Bligh, E.G., Dyer, W.J. 1959. A rapid method of total lipid extraction and purification. Canadian Journal of Biochemistry and Physiology 37(8): 911-917.

Borcard, D., Gillet, F., and Legendre, P. 2011. Numerical Ecology with R. New York. 306p.

Campbell, C.D., Cameron, C.M., Bastias, B.A., Chen, C.R., and Cairney, J.W.G. 2008. Long term repeated burning in a wet sclerophyll forest reduces fungal and bacterial biomass and responses to carbon substrates. Soil Biology & Biochemistry. 40 (9): 2246-2252.

Frostegård, Å., Bååth, E., and Tunlid, A. 1993. Shifts in the structure of soil microbial communities in limed soils as revealed by phospholipid fatty acid analysis. Soil Biol. Biochem. 25:723-730.

Frostegård, Å., Tunlid, A., and Bååth, E. 1991. Microbial biomass measured as total lipid phosphate in soils of different organic content. J Microbiol Methods 14:151–163.

Gonzalez, I., Ayuso-Sacido, A., Anderson, A., and Genilloud, O. 2005. Actinomycetes isolated from lichens: Evaluation of their diversity and detection of biosynthetic gene sequences. FEMS Microbiology Ecology 54: 401-415.

Grayston, S.J., Campbell, C.D., Bardgett, R.D., Mawdsley, J.L., Clegg, C.D., Ritz, K., Griffiths, B.S., Rodwell, J.S., Edwards, S.J., Davies, W.J., Elston, D.J., and Millard, P. 2004. Assessing shifts in microbial community structure across a range of grasslands differing in management intensity using CLPP, PLFA and community DNA techniques. Applied Soil Ecology 25: 63-84.

108

Gundale, M.J., DeLuca, T.H., Fiedler, C.E., Ramsey, P.W., Harrington, M.G., and Gannon, J.E. 2005. Restoration treatments in Montana ponderosa pine forest: Effects on soil physical, chemical and biological properties. Forest Ecology and Management. 213(1-3): 25-38.

Hortz, H.P., Barbrook, A., Field, C.B., and Bohannan, B.J.M. 2004. Ammonia-oxidizing bacteria respond to multifactorial global change. PNAS. vol 101, no 42. 15136-15141.

Kroppenstedt, R.M. 1985. Fatty acid and menaquinon analysis of actinomycetes and related organisms In: Chemical methods in bacterial systematics (Eds.) M. Googfellow and D.E. Minnikin. Academic Press, London, pp 173-199.

Legendre, P., and Legendre, L. 1998. Numerical Ecology, 2nd English edition. Elsevier Science BV, Amsterdam.

Paterson, E., Gebbing, T., Abel, C., Sim, A., and Telfer, G. 2007. Rhizodeposition shapes rhizosphere microbial community structure in organic soil. New Phytologist 3:600-610.

Paul, E.A. 2007. Soil microbiology, ecology, and biochemistry in perspective in Paul E.A. Editor. Soil microbiology and biochemistry. Third Edition. Burlington. 535 p.

Paul, E.A., and Clark, F.E., 1996. Soil microbiology and biochemistry, 2nd edn. New York: Wiley.

R Core Team. 2013. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org/.

Ramsey, P.W., Rillig, M.C., Feris, K.P., Holben, W.E., and Gannon, J.E. 2006. Choice of methods for soil microbial community analysis: PLFA maximizes power compared to CLPP and PCR-based approaches. Pedobiologia 50:275–280.

Switzer, J. M., Hope, G.D., Grayston, S.J, and Prescott, C.E. 2012. Changes in soil chemical and biological properties after thinning and prescribed fire for ecosystem restoration in a Rocky Mountain Douglas-fir forest. Forest Ecology and Management. 275: 1-13.

Van Bruggena, A.H.C., and Semenov, A.M. 2000. In search of biological indicators for soil health and disease suppression. Applied Soil Ecology 15: 13-24.

Visser, S. 2006. Oil Sands Plant Emissions in Northeastern Alberta: Monitoring Soil Chemistry, Soil Biology and Ectomycorrhizae in Jack Pine Stands – 2004 Results. Report presented to Wood Buffalo Environmental Association Terrestrial Environmental Effects Monitoring Committee. 71 pages.

109

White, D.C., Bobbie, R.J., King, J.D., Nickles, J., and Amoe, P. 1979. Lipid analysis of sediments for microbial biomass and community structure. Methodology for biomass determinations and microbial activities in sediments, ASTM STP 673, C.D. Litchfield and P.L. Seyfried, Eds., American Society for Testing and Materials. pp 87-103.

Zogg, G.P., Zak, D.R., Ringelberg, D.B., MacDonald, N.W., Pregitzer, K.S., and White, D.C. 1997. Compositional and functional shifts in microbial communities due to soil warming. Soil Science Society of America Journal 61:475-481.

110

Chapter 8: Routine foliar and soil analysis of 15 years monitoring in the AOSR Doug G. Maynard, Natural Resources Canada, Canadian Forest Service, Victoria, BC

8.1 Introduction The objectives of the routine soil and foliar chemical analysis were to determine: i) if there were differences in the chemistry between sites and years; ii) if there were differences could they be attributed to oil sands development either directly or indirectly, and iii) were any differences observed in the soil and foliar chemistry which modified the ecosystem health of the jack pine forests in the Athabasca oil sands region (AOSR). These monitoring results are important because they are a regulatory and management requirement for some industry approvals to operate (Percy et al., 2012).

8.2 Materials and methods 8.2.1 Sites selected Monitoring of the soil and jack pine foliar chemistry for WBEA began in 1998 and was done at 10 sites. Additional sites were added in the 2004 resampling and one site (JP212) was added and soils sampled in 2001. Unfortunately, some of the original 10 sites were lost to mine expansion and to compensate, 12 new sites were added in 2004. A new sampling design (Percy et al., 2012; Jaques and Legge, 2013; see also Chapter 5 this report) was put in place for the 2011 resampling with the addition of 10 new sites along with the remaining sites from the two previous samplings. Two additional sites (JP317 and JP318) were established and sampled in 2012 to fill in a gap east of the main emission sources. A third site (JP108) that was excluded in the 2004 sampling was resampled in 2011. Thus, there were 25 sites established and available for detailed monitoring in the 2011 resampling as 6 sites were burned in 2011, 4 had soils sampled but no foliage and the other two (JP106 and JP213) had both soils and foliage sampled. Details on the sites and the sampling dates are given in Table 5.2, Chapter 5.

8.2.2 Jack pine foliage sampling Foliar (needle) collection for chemical analysis varied among sampling times. In 1998, 10 off-site trees were individually sampled for current needles from the middle to upper third of the crown. Samples were composited prior to analysis so there was only one sample per plot. In 2004, current as well as 1 and 2 year old needles were sampled from the middle to upper third of the crown and samples collected from off-site trees were mixed in to form one composite sample per needle age class per site. The sampling method was modified in 2011 to five trees per plot, though samples were not composited. Thus, there were five sample replicates in each of the three age classes for a total of 15 needle samples per site. In addition, a duplicate set for each age class was obtained per site.

111

8.2.3 Soil sampling One soil pit was dug at each site established in 2011 to classify the sites according to the Canadian System of Soil Classification (Soil Classification Working Group, 1998). Soils were sampled by pedogenic horizon (e.g., LFH, A, etc.) to provide chemical analyses needed to classify the soils. Pit locations, rationale and procedures are described in Foster et al., 2015. Routine soil sampling procedures also varied among sampling times and are described in Chapter 5, this report.

At all sampling times the surface organic horizon (LFH) was collected and the mineral soils were sampled by depth (0-5; 5-15; 15-30; and at new sites 30-50 cm). The main differences between sampling times was the number of subsamples per site and stratified sampling design (e.g., stemflow, throughfall and freefall zones). The sampling schemes for the routine soil sampling are given in Appendix 5.2.

8.2.4 Foliar chemical analysis The foliar chemical analyses for the 2011 sampling followed the protocols outlined in Foster et al. (2015, Table 8.1 outlines the various methods used). Samples were dried at 70°C for 24 hours and ground in a zirconium ball mill. Total carbon (C), nitrogen (N) and sulfur (S) were analyzed by dry combustion and other total elements were analyzed by inductively coupled atomic emission spectrometry following microwave digestion. Sulfate in the needles was analyzed by two different methods in 2011.

8.2.5 Soil chemical analysis Soil chemical analysis in 2011 followed the protocols (including quality controls) as outlined in Foster et al. (2015) and summarized in Table 8.1.

8.2.6 Statistical analysis Summary statistics were calculated for jack pine foliage and soil chemical characteristics for each site. Means were calculated for each foliar age class (current, 1-year old and 2-year old) and each soil layer (surface organic horizon (LFH), 0-5 cm, 5-15 cm, 15-30 cm and 30-50 cm). Means and standard deviations for all measurements are presented in Appendices 8.1 and 8.2.

Statistical analysis was performed using the SAS system (SAS Institute). All measurements were tested by analysis of variance by site. A nested model was used for the soil data with the 4 subsamples nested within a plot (4 plots per site) for each site using general linear models (GLM) with least square means separation in SAS (n=4). The GLM for foliar chemistry was performed on plot means with n=5.

112

Table 8. 1 Foliar and soil chemical measurements used in the 2011 sampling year. See Wood Buffalo Environmental Association Procedure Manual for complete details (Foster et al., 2015).

Chemical Analysis Methods Notes Foliage – all age classes Total carbon (C) Dry combustion analyzer Total nitrogen (N) Dry combustion analyzer Total sulfur (S) Dry combustion analyzer Total S also measured by ICP-OES following microwave digestion Total calcium, magnesium, Wet digestion with nitric acid; potassium, manganese, analyses by inductively iron, aluminium, copper, coupled plasma – optical boron, zinc emission spectrometer (ICP- OES). Inorganic Sulfur Two methods: Methods compared; HI- (a) hydriodic acid (HI) - reducible S higher than reducible S weak acid extraction (b) Weak acid extraction method. (0.01 M HCl); analysis by ion chromatography

Soil pH 0.01 M CaCl2 Total C, N and S Dry combustion analyzer Nitrate and Ammonium Extraction in KCl and colorimetric determination by segmented flow analyzer Inorganic Sulfate

Cation Exchange Capacity 1.0 M unbuffered NH4Cl (CEC) and exchangeable extraction; Exchangeable cations cations measured by ICP-OES; CEC by NH4 replacement Water-soluble ions Extraction with water and analysis by ICP-OES

Available phosphorus Extraction by NH4F/H2SO4; Bray’s P colorimetric analysis by segmented flow analyzer Base cation: aluminum ratio Calculated based on and % base saturation exchangeable cations and Al and CEC.

113

The effect of S, N, or base cation deposition on jack pine foliage and soil chemical characteristics was examined by simple linear regression analysis. Predicted deposition using the CALPUFF model (see Chapter 4, this report) was regressed against various soil properties or foliage chemistry. Regression lines are shown for those relationships that had α< 0.10 (actual P values are given for all regressions). Linear regressions were also determined using the ion exchange resin data (see Chapter 3, this report). Trends using this data were similar to those using the CALPUFF model and are not shown.

8.3 Results and discussion 8.3.1 2011 Foliar chemistry There were significant differences in the foliar chemistry among sites (Appendix 8.1). This included all foliar elements measured and was not limited to the main pollutants (i.e., N and S). This is not unexpected even though considerable effort was made to select sites that were as similar as possible (Jaques and Legge, 2013). Slight differences in microclimate or edaphic factors could result in differences in foliar chemistry. Overall the elemental concentrations measured were within the natural range of variability for jack pine foliage reported in the literature with a few exceptions (Table 8.2).

Table 8. 2 Range of nutrient concentrations of current jack pine foliage and comparison with other studies in Boreal Plains and Boreal Shield West P Ca Mg K Mn Al Fe Zn g kg-1 mg kg-1 WBEA 1.04 – 1.82 – 0.71 – 3.15 – 321 - 252 - 542 23- 209 41 – 73 sites1 1.28 4.01 1.05 4.64 1216 1.06 – 1.52 – 0.84 – 3.28 – 242 - ARNEWS2 195 - 493 28 - 59 na 1.55 3.32 1.29 6.84 660 Meadow 1.16 – 1.69 – 0.82 – 4.05 – 224 - Lake PP – 189 - 240 6 - 49 na 2.11 7.76 1.08 7.86 559 site 93 Kimmins 1.1 – 2.6 – 0.60 – 2.60 – et al. 60 na 153 44 2.0 3.1 1.30 5.87 1985 1Range of concentrations in current foliage collected in 2011. 2Acid Rain National Early Warning Sites in the Prairie Provinces (Maynard and Fairbarns 1994). 3Maynard, D.G. Unpublished data from current jack pine foliage collected yearly from 1992 to 1996.

Foliar elemental concentrations were regressed against predicted N, S, base cations and potential acid input (PAI) depositions estimated using the CALPUFF model (Chapter 4, this report). In addition, foliar concentrations were regressed against predicted N, S and base cation deposition predicted using the ion exchange resin data of Fenn (Chapter 3, this report). There were weak but significant correlations of foliar N and S concentrations with predicted N and S deposition

114

using the CALPUFF model (Fig 8.1a, b). The correlations were weakly significant (R2 in the 0.17 to 0.25 range; Figures 8.1). Similar correlations were observed with the predicted deposition based on the ion exchange resin data (data not shown).

12 Total Total (g/kg) N 10 R2 = 0.16; P= 0.069 y = 3.26x + 11.2 8 0 0.2 0.4 0.6 0.8 Modeled N deposition (keq/ha/a)

1.4

1.2

0.8

0.6

Total Total S(g/kg) 0.4 R2 = 0.24; P=0.026 0.2 y = 1.50x + 0.53 0 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 Modeled S deposition (keq/ha/a)

Figure 8. 1 Total nitrogen (top) and S in current jack pine foliage versus modeled N and S deposition from the CALPUFF model (Chapter 4, this report).

There were no correlations of foliar N and S concentrations with PAI when the base cation deposition was included as part of it (Figure 8.2). At most sites any acid input was more than neutralized by the base cation deposition (see Chapter 4), supporting the recent results of Watmough et al. (2014) who found base cation deposition exceeded the combined inputs of N and S in both bulk deposition and throughfall.

115

There are four sites (JP102, JP104, JP212 and JP304) within 20 km of the Suncor and Syncrude stacks; though only one site(JP104) had elevated N and S concentrations (14.0 and 1.22 g kg-1, respectively). These concentrations are still considered within the normal range for current jack pine foliage although at the higher end of the range. There were a few anomalies with foliar N and S concentrations at sites further away from the main oil sands development area. Sites 317 and 318 had higher N concentrations than the majority of other sites; however, these sites were established and sampled in 2012 to fill in a gap between the main oils sands development and the Saskatchewan border. Their analogue type is being reassessed in 2015. Site JP213 in Saskatchewan also had higher N values than expected but was affected by the forest fires that burned a large area north and east of the main oil sands development.

The foliar S concentrations at JP317 and JP318 were also at the high end of the range for current jack pine foliage. These sites are due east of the main development areas (41 and 77 km for JP317 and JP318) and the predicted N and S deposition at these sites is less than half the predicted deposition at the sites close in. Thus, these slightly higher than expected concentrations may be because of temporal variability. Foliar concentrations have been shown to vary among years primarily as a result of differences in climatic conditions in the year prior to sampling and the year of sampling (Leaf et al., 1970; Maynard and Fairbarns, 1994).

The strongest relationships (R2=0.43; Figure 8.3) were between current foliar Ca concentrations and CALPUFF predicted base cation throughfall deposition (Ca, Mg and Na; K was not included in the deposition model predictions). The highest cation deposition was predicted to occur at JP104 and decrease with distance from the main development areas. Thus, the highest cation deposition was predicted to occur at the sites with the highest N and S deposition. The concentrations of other elements potentially of concern as a result of oil sand development had variable responses to PAI. In particular, differences in response varied depending on the element and the age of the foliage.

There were 5 sites where 3 complete foliar samplings occurred and given that in the two previous samplings there was no replication any interpretation of temporal trends is limited. Sulfur concentrations tended to be higher in the current foliage sampled in 2011 compared to the two previous years at 3 of the five sites. Two of these sites (JP104 and JP212) were within 20 km of the main stacks. The S concentrations at JP102 (another close in site) did not change among sampling times (Table 8.3).

116

12 Total Total N (g/kg)

10 R2=0.04; P=0.40

8 -1.5 -1 -0.5 0 0.5 1 Modeled PAI (keq/ha/a)

1.6

1.2

0.8 Total Total S(g/kg)

0.4

R2=0.02; P=0.53

0 -1.5 -1 -0.5 0 0.5 1 Modeled PAI (keq/ha/a)

Figure 8. 2 Total nitrogen (top) and S in current jack pine foliage versus modeled potential acid input (PAI) deposition based on bulk deposition measurements from the CALPUFF model (Chapter 4, this report).

117

3 Total Total (g/kg) Ca 2 R2 = 0.43; P=0.001 y = 0.527x + 1.83 1 0 0.5 1 1.5 2 2.5 Modeled cation deposition (keq/ha/a)

Figure 8. 3 Total calcium in current jack pine foliage versus modeled base cation deposition (Ca, Mg, and Na) from the CALPUFF model (Chapter 4, this report). Table 8. 3 Sulfur concentrations in current foliage of jack pine (g kg-1) at monitoring sites established in 1998 with multi-year sampling

Site 19981 20041 20112

JP101 0.56 0.66 0.70 ± 0.07

JP102 0.78 0.83 0.82 ± 0.05 JP103 0.67 0.79 Burned JP104 0.75 0.88 1.05 ± 0.13 JP106 0.58 0.74 0.76 ± 0.10 JP107 0.61 0.80 Burned JP109 0.59 0.81 Burned JP212 0.833 0.77 0.95 ± 0.15 1Ten trees were sampled at each site and composited into one sample for analysis per site. 2Mean and standard deviation (n=5). 3JP212 was established and sampled in 2001

118

Nitrogen concentrations in current foliage were higher in 2011 compared to the other sampled years at all five sites (Table 8.4). It is not clear whether the increases observed in the current foliage at all sites were related to year to year variation, an artifact of the different sampling methodologies or related to increases in atmospheric N deposition. The PAI of N deposition at JP102, JP104 and JP212 was elevated relative to the other sites, thus, the higher N concentrations of the foliage at these sites may be an indication that they are related to the higher N deposition. However, the possible reason for the increases at the other two sites is not evident based on the modeled N deposition.

There were 5 sites with 2 sampling times (includes JP108 that was not sampled in 2004 and JP205 that was sampled for soils but not foliage). As noted above there can be temporal variability in foliar concentrations so to make a proper assessment of temporal change more than 2 sampling times are needed; however, a similar trend of higher foliar S and N concentrations in 2011 was observed.

Table 8. 4 Nitrogen concentrations in current foliage of jack pine (g kg-1) at monitoring sites established in 1998 with multi-year sampling.

Site 1998 2004 2011 JP101 9.2 8.0 11.2 ± 1.6 JP102 10.4 7.7 12.8 ± 1.0 JP103 9.6 8.9 Burned JP104 10.0 8.6 14.0 ± 2.2 JP106 9.0 8.0 10.9 ± 1.4 JP107 10.0 8.8 Burned JP109 10.2 8.7 Burned JP212 9.13 9.3 12.2 ± 2.0 1Ten trees were sampled at each site and composited into one sample for analysis per site. 2Mean and standard deviation (n=5). 3JP212 was established and sampled in 2001

119

Table 8. 5 Sulfur and nitrogen concentrations in current foliage of jack pine (g kg-1) at monitoring sites established in 2004 and resampled in 2011. Site1 20042 20114 Sulfur JP1083 0.56 0.80 ± 0.03 JP201 0.68 0.70 ± 0.08 JP210 0.77 0.73 ± 0.09 JP213 0.70 0.87 ± 0.12 Nitrogen JP1083 9.7 11.7 ± 1.3 JP201 8.0 10.2 ± 1.5 JP210 8.9 10.3 ± 1.2 JP213 8.1 13.2 ± 1.8 1JP205 was burned in 2011 and was sampled for soils but not for foliage. 2Ten trees were sampled at each site and composited into one sample for analysis per site. 3JP108 was established and sampled in 1998 but was not sampled in 2004. 4Mean and standard deviation (n=5).

8.3.2 2011 Soil chemistry There were significant differences among sites for all measured parameters in the surface organic horizon (LFH), 0-5, 5-15, 15-30 and 30-50 cm mineral soils; however, differences that could be attributed to atmospheric deposition were limited to the LFH and extractable S in the 0-5 and 5- 15 and 15-30 cm depths. There were several sites sampled following the spring fires in 2011. The burned sites where the trees were killed were removed from the regression analyses because many of the soil parameters measured were affected by the fire. Two sites JP106 and JP213 were partially burned but the trees survived and these sites were included in the regression analyses.

120

Figure 8. 4 Extractable sulfur (a) and total sulfur (b) in the surface organic horizon (LFH) versus predicted nitrogen and sulfur deposition from the CALPUFF model (Chapter 4, this report).

8.3.2.1 Surface organic horizon (LFH) There were significant correlations of extractable (Figure 8.4a) and total S (Figure 8.4b) in the LFH with the predicted N and S deposition from the CALPUFF model (see Chapter 4). There was no correlation of total N (Figure 8.5) or inorganic N (ammonium and nitrate, see appendix 8.2). Total N concentrations ranged between 8 and 12 g/kg among sites. Site JP102 had the highest total N concentration (12 g/kg) but the other three sites with the highest predicted N and S deposition were similar to sites predicted to be at or near background deposition levels. Similar observations + were found for NH4 concentrations with JP102 having the highest concentration (163 mg/kg).

121

Nitrate concentrations in the LFH were low (< 1.3 mg/kg at all sites) and the four sites with the - highest predicted deposition had NO3 concentrations in the LFH < 0.25 mg/kg.

The lack of correlation between LFH N and predicted deposition may be related to the N limitations of these jack pine ecosystems and the time since significant N deposition had occurred. At the current N deposition rate, any N being deposited may be taken up by the trees and other vegetation and the higher N concentrations in the jack pine needles at the four sites would support that conclusion. The lack of correlation between LFH N and predicted deposition would also indicate that at the current rate of N deposition the jack pine sites are not N saturated (at least not from the perspective of N losses, primarily leaching from the system).

There was no effect of predicted N, S or base cation deposition on the pH of the LFH with the pH of the LFH ranging from 3.46 to 5.03 (Fig. 8.6). The highest pH was at JP106 and was related to the wildfire in May 2011 (trees were still alive). Elevated pH often results following wildfires (Maynard et al., 2014) and was also observed at other sites burned in 2011. The increased pH is short lived and LFH pH usually returns to pre-fire levels within 2 or 3 years. In addition to elevated pH, the total S and N concentrations of the LFH generally decrease at the burned sites. Both S and N volatilize at relatively low temperatures, so following fire there will often be a decrease in total S and N concentrations. In boreal ecosystems the loss of N following fire has been suggested as one of the reasons for N limitations in that ecosystem. Under conditions of no anthropogenic atmospheric deposition, N is primarily returned to the soil through biological N fixation and it may take decades for the N levels to return to pre-wildfire concentrations.

6 R2 < 0.01; P = 0.99 Total Total N (g/kg) 4

0 0 0.2 0.4 0.6 0.8 1 Modeled N and S deposition (keq/ha/a)

Figure 8. 5 Total nitrogen in the surface organic horizon (LFH) versus predicted nitrogen and sulfur deposition from the CALPUFF model (Chapter 4, this report).

122

6.00

5.00

4.00

3.00 pH

2.00 R2 = 0.03; P=0.46 1.00

0.00 0 0.2 0.4 0.6 0.8 1 Modeled N+S deposition (keq/ha/a)

Figure 8. 6 pH in the surface organic horizon (LFH) versus predicted nitrogen and sulfur deposition from the CALPUFF model (Chapter 4, this report).

Table 8. 6 Total sulfur and nitrogen and pH of the surface organic horizon (LFH) of burned sites (burned in May 2011 prior to sampling) sampled in 2011. Mean (standard deviation; n=4).

JP sites 103 106 107 109 205

Total S 0.3 0.8 0.6 0.8 0.4 (g kg-1) (0.1) (0.2) (0.1) (0.0) (0.1)

Total N 5.9 9.2 7.1 9.2 8.0 (g kg-1) (1.7) (2.9) (1.1) (1.1) (2.1)

pH 5.90 5.03 5.34 5.48 4.97 (0.27) (0.37) (0.37) (0.31) (0.32)

123

8.3.2.2 Mineral soil 0-5 cm layer There was no significant relationship between total N or total S in the 0-5 cm mineral soil layer and predicted N and S deposition (CALPUFF model). Extractable sulphate was the only measured parameter that showed a positive correlation to the predicted N and S deposition (Figure 8.7). Most of the S deposited would be as inorganic sulphate and in these sandy soils any excess sulfate not taken up by the vegetation would be susceptible to leaching and movement down through the profile.

2.5

1.5

Sulfate (mg/kg) 1 R2 = 0.12; P=0.084 y = 1.48x + 0.213 0.5

0 0 0.2 0.4 0.6 0.8 1 Modeled N and S deposition (keq/ha/a)

Figure 8. 7 Extractable sulphate in the 0-5cm mineral soil layer versus modeled N and S deposition from the CALPUFF model (Chapter 4, this report)

124

6.00

5.00

4.00

3.00 pH R2 = 0.002; P=0.84 2.00

1.00

0.00 0 0.2 0.4 0.6 0.8 1 Modeled N and S deposition (keq/ha/a)

Figure 8. 8 Soil pH in the 0-5 cm mineral layer versus modeled nitrogen and sulfur deposition from the CALPUFF model (Chapter 4).

4.00

3.50 R2 = 0.19; P=0.029 3.00 y = -1.65x + 2.25 2.50

2.00

1.50

1.00 Ammonium (mg/kg) 0.50

0.00 0 0.2 0.4 0.6 0.8 1 Modeled N and S deposition (keq/ha/a)

Figure 8. 9 Ammonium concentration in the 0-5 cm mineral layer versus modeled nitrogen and sulfur deposition from the CALPUFF model (Chapter 4).

The other parameters were not correlated to the predicted N and S deposition (e.g., pH; Figure 8.8) with the exception of ammonium concentrations which were negatively correlated with N + + and S deposition. Nevertheless, NH4 concentrations were low at all sites (< 4 mg/kg NH4 -N).

8.3.2.3 Mineral soil 5-15 and 15-30 cm layers Extractable sulphate was significantly correlated with predicted N and S deposition (CALPUFF model) at both soil layer depths (Figures 8.10 and 8.12). No other measured soil parameter (e.g., pH; Figures 8.11 and 8.13) was correlated with the predicted N and S deposition. In addition, ion

125

exchange membranes (PRSTM probes) were deployed at JP104 and JP201 to determine soluble ion concentrations (Percy et al., 2013). Soluble sulphate levels were higher at JP104 (0.74 keq/ha/a N and S deposition) than at the background site JP201 (0.23 keq/ha/a N and S + - deposition); however, soluble NH4 and NO3 levels were similar at both sites. This is further evidence that at the current deposition levels, nitrogen is being taken up by the vegetation and there is minimal leaching of N. Sulfur in the soil at the sites within 20 km appears to be leached down to at least 30 cm.

9 8 R2 = 0.28; P=0.020 7 y = 5.06x + 0.365 6 5 4 3

Sulfate (mg/kg) 2 1 0 0 0.2 0.4 0.6 0.8 1 Modeled N and S deposition (keq/ha/a)

Figure 8. 10 Extractable sulfate in the 5-15cm mineral soil layer versus modeled N and S deposition from the CALPUFF model (Chapter 4, this report)

6.00

5.00

4.00

3.00 pH R2 = 0.07; P = 0.23 2.00

1.00

0.00 0 0.2 0.4 0.6 0.8 1 Modeled N and S deposition (keq/ha/a)

Figure 8. 11 Soil pH in the 5-15 cm mineral layer versus modeled nitrogen and sulfur deposition from the CALPUFF model (Chapter 4).

126

9 8 7 6 5 4

3 Sulfate (mg/kg) 2 R2 = 0.21; P = 0.046 1 y = 4.39x + 1.93 0 0 0.2 0.4 0.6 0.8 1 Modeled N and S deposition (keq/ha/a)

Figure 8. 12 Extractable sulfate in the 15-30 cm mineral soil layer versus modeled N and S deposition from the CALPUFF model (Chapter 4, this report)

5.20 5.10 5.00 4.90 4.80

pH 4.70 4.60 4.50 R2 = 0.08; P = 0.22 4.40 4.30 0 0.2 0.4 0.6 0.8 1 Modeled N and S deposition (keq)/ha/a

Figure 8. 13 Soil pH in the 15-30 cm mineral layer versus modeled nitrogen and sulfur deposition from the CALPUFF model (Chapter 4).

8.3.2.4 Percent Base Saturation (% BS) and Base Cation to Aluminum (BC:Al) Ratio Changes in the % BS and the BC:Al ratio of the soil are the parameters in the regional effects- based emissions management framework for the oil sands that could trigger a change in emission levels (CEMA, 2004). For example, emissions reduction are required if model simulations show the chemical thresholds for %BS and BC:Al ratio are reached within 15 years. Chemical thresholds on a site-specific basis are calculated as half the change between the estimated historical condition and a fixed-effect end point (%BS = 10%; BC:Al = 2), calculated through model simulations (see Whitfield et al., 2009). Eight of the current monitoring sites were sampled in

127

1998, 2004 and 2011 (exception was JP108 that was sampled in 1998 and 2011 but not 2004). Of these 4 were burned in 2011. Thus, at these sites we were able to compare %BS after 14 years for each soil depth (Table 8.7). It was not possible at this time to compare BC:Al ratios among years because of differences in the analytical methodology and calculations.

The %BS varied considerably among sites. However, there was no correlation of %BS with modeled N and S deposition in the mineral soil layer (Figure 8.14b only 0-5 cm mineral soil layer shown). There was a weakly significant positive correlation of %BS with modeled N and S deposition (R2= 0.19; P=0.048; Figure 8.14a) in the LFH. The %BS increased was probably related to higher base cation deposition at the sites nearer to the main mine developments.

As noted, the trigger for a management response to acid deposition is based on changes to %BS with time. Whitfield et al. (2009) modeled changes in %BS at 15 and 35 years after S deposition based on the current deposition and at double the current deposition rate. The changes were estimated using a dynamic model with the exchangeable cations and CEC calculated based on a weighted average by soil horizon to a depth of 30 cm. The model predicted that %BS did not decrease below the threshold at any site based on the current and two times the current S deposition. Similar to the modeled results of Whitfield et al. (2009) the threshold level for %BS was not reached for any site or any depth. The %BS of the LFH increased at the four burned sites and this is consistent with increased pH at these sites (Table 8.6). There are a few anomalies but overall the results support the modeled values for %BS of Whitfield et al. (2009).

The comparison among years is limited given the uncertainty associated with the data and the few sites with multiple sampling years. The %BS of the mineral soils tended to be lower than in the previous two samplings; however, these results are difficult to interpret. There are only 5 sites; 3 in areas of higher deposition and 2 sites considered in areas of low deposition. However, decreases were observed at all sites thus, acid deposition and leaching losses cannot be the sole explanation. There were no error estimates for the first two samplings but if we assume they were similar to the variation shown in 2011 then, in many cases the mean differences may not be significant.

128

Table 8. 7 The % base saturation (%BS) in soil layers at monitoring sites established in 1998 with multi- year sampling. Site Sampling year Threshold1 1998 2004 2011 LFH JP101 80 76 71 ± 5 45 JP102 82 88 85 ± 2 46 JP104 72 74 68 ± 4 41 JP108 72 --3 60± 10 41 JP2122 82 88 116 ± 10 46

0 – 5 cm JP101 72 69 59 ± 7 41 JP102 67 64 59 ± 12 44 JP104 52 48 41 ± 3 33 JP108 37 --3 37 ± 10 28 JP212 69 55 46 ± 10 40

5-15 cm JP101 80 69 59 ± 6 45 JP102 67 41 59 ± 11 39 JP104 62 21 36 ± 5 36 JP108 26 --3 18 ± 5 18 JP212 60 37 35 ± 11 35

15-30 cm JP101 87 99 68 ± 4 48 JP102 69 94 64 ± 14 40 JP104 84 91 52 ± 12 47 JP108 41 --3 24 ± 10 26 JP212 72 80 48 ± 11 41 1Threshold was calculated as half the change between the 1998 value and a fixed-effect end point of %BS = 10%. For the purposes of this exercise we considered the 1998 value as the historical base. 2JP212 was established and sampled in 2001 3JP108 was established and sampled in 1998 but was not sampled in 2004.

There was even greater variability with the BC:Al ratio than %BS as small changes in Al concentration or one of the base cations can cause considerable changes in the ratio. Similar to %BS there was no correlation of BC:Al ratio with modeled N and S deposition in the LFH (Figure 8.15a or any mineral soil layer (Figure 8.15b; only 0-5 cm mineral layer shown). Whitfield et al. (2009) modeled the changes in BC:Al using the same scenarios as outlined above for %BS. The BC:Al ratio was calculated using exchangeable cations and Al on a molar charge basis. In the earlier reports, the BC:Al ratio was based on the soluble cations and Al; however for the 129

modeling, Whitfield et al. (2009) used the ratios based on the exchangeable concentrations. In the model of Whitfield et al. (2009), for the majority of sites under the two deposition scenarios,

140.00 a 120.00

100.00

80.00

60.00

40.00 R2 = 0.19; P=0.048 % base saturation % saturation base (%BS) 20.00 y = 28.24x + 71.97

0.00 0 0.2 0.4 0.6 0.8 1 Modeled N and S deposition (keq/ha/a)

80.00 b 70.00

60.00

50.00

40.00

30.00

20.00 R2 <0.01; P=0.87 % base saturation % saturation base (%BS) 10.00

0.00 0 0.2 0.4 0.6 0.8 1 Modeled N and S deposition (keq/ha/a)

Figure 8. 14 % Base Saturation (BS) in the (a) LFH and (b) 0-5cm mineral soil layer versus modeled N and S deposition form the CALPUFF model (Chapter 4, this report).

130

50.00 45.00 a 40.00 35.00 30.00 25.00 20.00 15.00 Base Base cation:Al ratio 10.00 R2 = 0.05; P=0.31 5.00 0.00 0 0.2 0.4 0.6 0.8 1 Modeled N and S deposition (keq/ha/a)

10 b 9 8 R2 <0.001; P= 0.94 7 6 5 4 3

Base Base cation: Al ratio 2 1 0 0 0.2 0.4 0.6 0.8 1 Modeled N and S deposition (keq/ha/a)

Figure 8. 15 Base cation to aluminum (BC:Al) ratio in the (a) LFH and (b) 0-5cm mineral soil layer versus modeled N and S deposition form the CALPUFF model (Chapter 4, this report). the BC:Al ratio threshold was not exceeded. In the modeling exercise presented in Chapter 4, any exceedances of the critical load for acid deposition were offset by base cation deposition. A comparison among sampling times was not relevant as the reported ratios in the earlier reports was calculated on the soluble cations and Al.

8.4 Summary The 2011 soil and foliar sampling included 25 sites (including two added in 2012). Current, one- year-old and two-year-old jack pine needles were sampled and soils were sampled at four depths; surface organic horizon (LFH), 0-5, 5-15, 15-30 and 30-50 cm. The latter depth was only sampled at the sites established in 2011 for baseline conditions.

131

Measurements of N, S, Ca and some metals associated with oil sands development show elevated concentrations in the needles of jack pine within ~ 20 km from the source complex similar to several of the studies reported in the recent book on Alberta oil sands (Percy and Krupa, 2012). A similar pattern was measured for extractable sulfur in the soil to a depth of 30 cm but not for nitrogen. The lack of correlation with N and predicted deposition suggests that the N deposited as a result of oil sands development was being taken up by the vegetation and was not building up in the soil. It is unknown at what N deposition rate or how long it would take (if at all) under the current deposition for N leaching to occur. While increased soluble sulphate has been measured at depth (at least 30 cm) there was no correlation of pH, base cations or aluminum concentrations with N and S modeled deposition in 2011. It may be that any effects of acid inputs have been offset by relatively high base cation deposition (Watmough et al., 2014; Watmough and Whitfield, Chapter 12); however, it is not known at what rates of S (and N) deposition or how long before the indirect effect of acid deposition would be a concern and the potential for effects if base cation deposition were reduced.

Currently, measurable changes in foliar or soil parameters that could be related to oil sands development occurred at the routine monitoring sites within ~20 km. There were no apparent indirect effects on soil or foliar elemental concentrations at these sites; however, continued monitoring is warranted given the planned increase in development and potential changes in atmospheric deposition in the AOSR.

8.5 References CEMA. 2004. Recommendations for the Acid Deposition Management Framework for the Oil Sands Region of North-Eastern Alberta. Prepared by the Cumulative Environmental Management Association, NOx/SOx Management Working Group. 39 pp.

Foster, K. R., Baines, D., Percy, K.E., Legge, A.H., Maynard, D.G., and Chisholm, V. 2015. Wood Buffalo Environmental Association Terrestrial Effects Monitoring: Forest Health Monitoring Program 2011 Procedures Manual. WBEA, Fort McMurray, Alberta.

Jacques, D.R., and Legge, A. H. 2012. Ecological analogues for biomonitoring industrial sulfur emissions in the Athabasca Oil Sands Region, Alberta, Canada. Chapter 10. pp 219-241. IN K.E. Percy (Ed.) Alberta Oil Sands:Energy, Industry and the Environment. Elsevier, Oxford, UK.

Kimmins, J.P., Binkley, D., Chatarpaul, L., and de Catanzaro, J. 1985. Biogeochemistry of temperate forest ecosystems: literature on inventories and dynamics of biomass and nutrients. Can. For. Serv., Petawawa Natl. For. Inst., Chalk River, Ontario. Inf. Rep. PI-X-47E/F.

Leaf, A.L., Berglund, J.V., and Leonard, R.E. 1970. Annual variation in foliage of fertilized and /or irrigated red pine plantations. Soil Sci. Soc. Amer. Proc. 34: 677-682.

132

Maynard, D.G., and Fairbarns, M.D. 1994. Boreal ecosystem dynamics of ARNEWS plots: base line studies in the Prairie Provinces. Nat. Resour. Can., Can., For. Serv. Northwest Reg., North. For. Cent., Edmonton, Alberta. Inf. Rep. NOR-X-327.

Maynard, D.G., Paré, D., Thiffault, E., Lafleur, B., Hogg, K.E., and Kishchuk, B. 2014. How do natural disturbances and human activities affect soils and tree nutrition and growth in the Canadian boreal forest? Environmental Reviews 22: 161-178. Doi: 10.1139/er-2013-0057.

Percy, K.E., Maynard, D.G., and Legge, A. H. 2012. Applying the forest health approach to monitoring boreal ecosystems in the Athabasca oil sands region. pp 193-217. IN K.E. Percy (Ed.) Alberta Oil Sands:Energy, Industry and the Environment. Elsevier, Oxford, UK.

Percy, K.E., and Krupa, S. 2012. Concluding remarks. pp 469-483. IN K.E. Percy (Ed.) Alberta Oil Sands:Energy, Industry and the Environment. Elsevier, Oxford, UK.

Soil Classification Working Group. 1998. The Canadian System of Soil Classification. Agriculture and Agri-Food Canada Publication 1646 (Revised). NRC Research Press, Ottawa, ON.

Watmough, S. A., Whitfield, C.J., and Fenn, M.E. 2014. The importance of atmospheric base cation deposition for preventing soil acidification in the Athabasca Oil Sands Region of Canada. Science Total Environment 493: 1-11.

Whitfield, C.J., Aherne, J., and Watmough, S.A. 2009. Modeling soil acidification in the Athabasca oil sands region, Alberta, Canada. Environ. Sci. Technol. 43: 5844-4850.

133

Chapter 9: Responses of understory vegetation to deposition from oil sand processing operations Ellen Macdonald, Department of Renewable Resources, University of Alberta, Edmonton, AB

9.1 Introduction Atmospheric nitrogen and sulphur deposition can result in a fertilization effect - causing increased tree growth and understory plant abundance – or can cause soil acidification and nutrient imbalances – reducing productivity and biodiversity and changing community composition (Greaver et al., 2012). Effects on forest understory plant communities can be direct - through damage due to exposure, or indirect - through effects on soil pH and nutrient availability. These, in turn, influence competitive interactions among understory species. Responses can also be mediated indirectly through impacts of deposition and soil chemistry on canopy trees, which then affects the understory through below- or above-ground competition (Bobbink et al., 2010). Nitrogen addition has also been implicated in increased disease incidence in forest understory shrubs (Nordin et al., 2005).

Nitrogen deposition to boreal forests has been associated with increased abundance of nitrophilous species – including grasses, forbs (broadleaf non-woody plants) and shrubs but decreased abundance of ericaceous species, bryophytes (mosses, liverworts and hornworts) and lichens (Nordin et al., 2005; Bobbink et al., 2010). Bobbink et al. (2010) reported that there was little evidence for effects of nitrogen deposition on species richness (number of species per given area) in boreal forests but experimental acidification in Swedish boreal forests has been associated with declines in species richness (Hallbäcken and Zhang, 1998; Nordin et al., 2005) and dramatically reduced abundance of mosses (Mäkipää, 1998). Nitrogen fertilization in Swedish boreal forests has also been associated with reduced abundance of ground “reindeer” lichens such as Cladina rangiferina (Olsson and Kellner, 2006). In temperate forests, which have a very abundant and rich understory community, nitrogen additions have been associated with major shifts in community composition and declining species richness (Gilliam and Roberts, 2003; Gilliam, 2006; Clark et al., 2013).

The dry, nutrient poor jack pine sites chosen for the TEEM monitoring program were classified as the a1.1 Pj/bearberry/lichen ecosite phase (Beckingham and Archibald, 1996). This ecosite type was expected to be the most sensitive to atmospheric deposition resulting from industrial emissions associated with oil sands operations in northeastern Alberta. The sites were chosen to be in locations anticipated to represent a gradient of deposition and studies have verified that the chosen sites are, indeed, located along a gradient of deposition of nitrogen, sulphur and base cations (Davies et al., Fenn, Chapters 3 and 4 in this report). Previous studies in the region showed a positive correlation of atmospheric NO2 concentrations with nitrogen in lichens and in jack pine (Pinus banksiana) needles (Laxton et al., 2010). The 2004 monitoring of the TEEM showed that

134

jack pine foliar sulphur, as well as nitrogen and sulphur concentrations in lichen tissue, were positively associated with atmospheric deposition (Jones and Associates Ltd., 2007). There were no observed effects on forest productivity.

Herein variation in abundance, richness and composition of the understory plant community at the TEEM monitoring sites in relation to gradients of atmospheric deposition is examined.

9.2 Methods For the analyses we conducted, data from 19 sites were included; i.e. all sites sampled in 2011/12 (see Foster this report Table 5.2) excluding those which had burned in the 2011 wildfire (JP103, JP106, JP107, JP109, JP205, JP213). Sampling of understory vegetation occurred in the 10m x 40m vegetation plot (see Chapter 5, Foster this report). Visual estimates of percent cover to species were made in the 10 small (0.4 m x 1m) subplots, the two medium (1m x 1m) subplots and in the one large (2m x 20m) subplot. Actual cover values were recorded as well as cover class values (see Foster this report). For the analyses conducted herein the actual cover values were used. All vascular plants were identified to species, including tree species that occurred as seedlings or saplings. The dominant mosses and lichens were also identified to species (see Appendix 9.3). Cladonia sp. other than Cladonia uncialis were present in low abundance and these were grouped at the genus level. The data should not be taken to represent a complete species list for bryophytes or lichens at the sites since there was no intensive search of specific microhabitats (e.g., downed dead wood, live and dead standing trees) important for these taxa.

Cover estimates are likely more precise in the small and medium subplots and these also serve as small, focused locations for repeated measures over time. However, the large subplot provides a more complete picture of the species present, and thus richness. For the purposes of the analyses conducted, average cover value per species was calculated by taking the average of: the average of the ten small subplots, the average of the two medium subplots, and of the one large subplot. Cover (% summed across taxa) and richness (number of species) values were calculated from these data as follows: total (all vascular plants, mosses and lichens); vascular plants; mosses; lichens; forbs (here including broadleaf non-woody species plus prostrate or trailing woody species such as bearberry (Arctostaphylos uva-ursi), bog cranberry (Vaccinium vitis-idaea), or bunchberry (Cornus canadensis)); and shrubs (all other woody species, including tree seedlings or saplings).

To examine relationships between understory vegetation and atmospheric deposition, regression and constrained ordination analyses were utilized. For these, distance from the centre of industrial development (see Fenn, Chapter3, this report) was used as an independent variable along with two different sets of atmospheric deposition variables: 1) predictions based on source emissions as derived from the CALMET/CALPUFF dispersion model (See Davies et al., Chapter 4); and 2) predictions modelled on the basis of ion exchange resin data collected at a network of sites in the region (see Fenn, Chapter 3).

135

From the CALPUFF model, deposition values were predicted based on emission sources within the Lower Athabasca Region (LAR) and for LAR sources plus sources outside this region. Also, for CALPUFF two different assumptions were used for base cations, and thus potential acid input: bulk deposition and throughfall through the forest canopy. Thus, a total of 11 CALPUFF variables were used: predicted SO2 and NO2 concentration, sulphur deposition (LAR), sulphur deposition (Total), nitrogen deposition (LAR), nitrogen deposition (total), base cations (bulk), base cations (throughfall), potential acid input (PAI) (bulk BC assumption) and PAI (throughfall BC assumption) (Appendix 9.1). The model based on the ion exchange resin data predicted deposition of: ammonium (NH4-N), nitrate (NO3-N), dissolved inorganic nitrogen (DIN), sulfate (SO4-S) and base cations (BC). For each of these, separate predictions were made for bulk deposition and for throughfall resulting in a total of 10 IER deposition variables (see Appendix 9.2).

Linear regressions were utilized to examine the potential influence of atmospheric deposition on understory vegetation cover and richness. As a preliminary step, individual linear regressions were conducted for each of the 12 understory vegetation response variables (cover and richness for: total, mosses, lichens, vascular plants, forbs, shrubs) versus each of the 22 independent variables (11 CALPUFF deposition variables, distance from source, 10 IER deposition variables). These allowed determination of which variables were significantly related, and thus should be further considered for inclusion in the subsequent multiple regression model; also, this allowed determination of non-linear relationships between response and predictor variables. For each of these regressions, the residuals were examined for normality, homogeneity of variance and to look for evidence of non-linearity. No transformations were required to meet these assumptions. Those independent variables that were significant were retained for subsequent model selection procedures for each of the dependent variables, as follows. For a given dependent variable, a stepwise multiple regression was conducted including all the significant CALPUFF independent variables. Likewise a separate stepwise multiple regression was conducted including all the significant IER independent variables. The different CALPUFF variables and IER variables were often highly correlated and these stepwise regressions produced a reduced set of CALPUFF variables and of IER variables, which were internally independent and had the greatest explanatory power. A final step-wise multiple regression was conducted including the reduced set of CALPUFF variables, distance from source (when significant), and the reduced set of IER variables. Through this analysis the most parsimonious model was chosen by comparison of Akaike Information Criterion (AIC) values. All regression analyses were conducted in R. version 3.1.1 using the “lm” “drop1” and “step” procedures (R Core Development Team, 2014).

To examine the potential influence of atmospheric deposition on understory composition, I used constrained ordination (Redundancy Analysis). The response dataset consisted of Hellinger- transformed cover values for each vascular species along with summed cover for mosses and lichens. The decision was made to sum the latter two taxonomic groups because the vast majority of cover for each was represented by only one species. The decision followed preliminary analysis based on cover values for all taxa. The environmental dataset consisted of the 22 independent variables, as described above. A model selection procedure similar to that described above for the linear regressions was followed. Two separate analyses using backwards selection and

136

examination of variance inflation factors were used to choose a reduced set of CALPUFF deposition variables and a reduced set of IER deposition variables. Backwards selection was then run using an environmental dataset that included distance from source and the two reduced sets of deposition variables. This analysis produced a final reduced set of independent variables. The significance of the overall ordination and of each axis was tested using permutation procedures. The ordinations were conducted in R version 3.1.1 (R Core Development Team, 2014) using the “rda”, “vif.cca”, “ordistep” and “anova.cca” procedures in the “vegan” package (Oksanen et al., 2013).

9.3 Results and discussion Seven of the 12 understory vegetation cover and richness variables were significantly, positively related to the deposition variables and negatively related to distance from source (see Appendix 9.1 and 9.2). Total cover, shrub cover, bryophyte cover, lichen cover, and lichen richness were not significantly related to any of the deposition variables. The only significant negative relationship with a deposition variable was moss richness with CALPUFF Potential Acid Input with a Bulk Base Cation assumption. In all but one instance the vegetation variables were more strongly related to the deposition variables than they were to distance from source. Total richness, vascular cover, forb cover and shrub richness were significantly related to many more deposition variables than were the other vegetation response variables.

The best model, as chosen using the model selection procedure, always included just one independent variable. Indeed, only two independent variables were chosen for the seven ‘best’ models: CALPUFF Potential Acid Deposition (PAI) with a Bulk Base Cation (BC) assumption (from Davies et al., Chapter4) for vascular plant cover and richness, forb cover and richness, and moss richness; IER Base Cations – Throughfall (from Fenn, Chapter 3) was the independent variable chosen for the model of total richness and shrub richness (Table 10.1). The models had moderate to low explanatory power with R2 values ranging from 0.22 to 0.35.

Although the regressions were not particularly strong the relationships indicated are striking. Predicted total richness varied from 26 species to 33 over a gradient in IER Base Cations (throughfall) of 0.9 to 1.8 keq/ha/yr (Fig. 9.1). Predicted vascular plant cover varied from 12 to 22% and vascular richness varied from 12 to 18 species over a CALPUFF PAI (bulk BC) gradient of -0.8 to 0.3 keq/ha/yr (Fig. 9.1). The increase in total richness was almost exclusively due to increased vascular plant richness, which in turn was due to increased richness of forbs and shrubs. There were very few graminoids (grasses or sedges) in the community and they had extremely low cover. Likewise the increase in vascular plant cover was due to increased forb cover – bearing in mind that this included prostrate trailing woody species such as bearberry, bog cranberry and twin flower (Linnaea borealis).

137

Table 9. 1 Results of the best regression model for each of the understory vegetation response variables. Given is the independent variable that was chosen during the model selection procedure, the intercept, the estimate of the slope, the R2 value, and significance (P) of the regression. CALPUFF PAI (Potential Acid Deposition) – Bulk BC (Base Cation) assumption is from Davies et al. (Chapter 4, this report); IER Base Cations – Throughfall is from Fenn (Chapter 3, this report). The following understory vegetation response variables were not significantly related to any of the predictor variables: Total cover, shrub cover, bryophyte cover, lichen cover, lichen richness (see also Appendix 9.1 and 9.2). Response Variable Independent Variable1 Intercept Estimate R2 P Vascular Cover CALPUFF PAI (Bulk BC) 19.37 9.25 0.2264 0.0395 Forb Cover CALPUFF PAI (Bulk BC) 13.79 9.71 0.3497 0.0077 Total Richness IER Base Cations - Throughfall 19.04 0.0074 0.3038 0.0144 Vascular Richness CALPUFF PAI (Bulk BC) 16.93 6.05 0.2932 0.0166 Shrub Richness IER Base Cations - Throughfall 1.77 0.0013 0.2952 0.0162 Forb Richness CALPUFF PAI (Bulk BC) 11.00 5.07 0.3113 0.0131 Bryophyte Richness CALPUFF PAI (Bulk BC) 4.62 -1.63 0.2890 0.0176 1 units for both variables are keq/ha/yr

35 20

30 15

Total Richness 10

20 Vascular RichnessPlant 0.8 1.0 1.2 1.4 1.6 1.8 -1.0 -0.6 -0.2 0.0 0.2 0.4 IER Base Cations - Throughfall (keq/ha/yr) CALPUFF PAI - Bulk BC (keq/ha/yr)

Figure 9. 1 Total Richness, Cover of Vascular Plants, and Vascular Plant Richness versus deposition, as chosen during the model selection procedure. Each point represents a site and the line shows the regression. Regression equations were as follows: Total Richness = 20.119 + 7.45*IER-Base Cations (Throughfall)1 ; P = 0.014, R2= 0.304, Vascular cover = 19.37 + 9.245* CALPUFF PAI (Bulk BC)2; P = 0.040, R2 = 0.463, Vascular Richness = 16.93 + 6.049* CALPUFF PAI (Bulk BC)2 ; P = 0.017, R2 = 0.017, 1from Fenn (Chapter 3); 2from Davies et al. (Chapter 4). (See also Table 10.1).

While the regression for moss richness showed a negative effect of atmospheric deposition, the magnitude of the effect was small: moss richness varied from 6 to 4 species over a gradient in CALPUFF PAI (bulk BC) gradient of -0.8 to 0.3 keq/ha/yr. Given that all moss species except big red-stem (Pleurozium schreberi) had extremely low cover, and the fact that a thorough bryophyte survey of the sites was not conducted, this negative response needs to be interpreted with caution. Nevertheless, bryophytes have been identified as being particularly susceptible to the effects of nitrogen additions with responses varying among species depending on their affinity for rich habitats (Mäkipää, 1995; Mäkipää, 1998; Nordin et al., 2005; Bobbink et al., 2010). A

138

more thorough examination of bryophyte assemblages across a gradient of atmospheric deposition in the Athabasca Oil Sands Region could be revealing.

In the constrained ordination, four CALPUFF deposition variables and three IER deposition variables were selected for inclusion in the Redundancy Analysis (Table 10.2). The overall ordination was significant and explained 75.2 % of the variation in species composition. CALPUFF Potential Acid Input with a bulk base cation assumption loaded to the low end of axis 1 while CALPUFF base cation with either a bulk or throughfall assumption loaded to the positive end on axis 1; these three did not separate on axis 2 (Fig. 9.2). The remaining variables: CALPUFF S deposition, IER sulfate, IER dissolved inorganic nitrogen, and IER base cations (all bulk deposition) all loaded to the low end of axis 2 and did not separate on axis 1. Cover of mosses was positively associated with CALPUFF Potential Acid Input (bulk base cation assumption) while cover of lichens was positively associated with CALPUFF base cation (bulk assumption). Cover values for the trailing shrub bearberry and (to a lesser extent) common blueberry (the shrub Vaccinium myrtilloides) and were positively associated with all the other deposition variables (Fig. 9.2). Average cover of bearberry at the five sites with the highest predicted IER base cation deposition values (JP102, JP104, JP212, JP304, JP315; see Fenn, Chapter 3) was 12.6%, compared to an average of 2.9% at the other 14 sites.

Previous studies have shown that atmospheric nitrogen deposition is associated with increased abundance of grasses and forbs but decreases abundance of dwarf shrubs while not affecting species richness (Nordin et al., 2005; Bobbink et al., 2010). In contrast, we found no such increases in grasses or forbs – except for the trailing shrub bearberry – and we did find increased total and vascular plant species richness. The latter is opposite to findings from pine and spruce forests in Sweden, where atmospheric depositions were associated with reduced vascular plant richness (Hallbäcken and Zhang, 1998). They also found reduced abundance of mosses (Mäkipää, 1998; Nordin et al., 2005) and ground lichens (Olsson and Kellner, 2006) whereas we saw no relationship between atmospheric deposition and abundance of these groups. Given that the sites we sampled were dry with low soil nutrient content and were species-poor, it is perhaps not surprising that the increase in vascular plant cover we saw were owing to increases in the dominant vascular plant, which happened to be the trailing ericaceous shrub bearberry. The increases in vascular plant cover were minor, however, and overall vascular plant cover was low. Thus, the vascular plant community probably has little competitive effect on the ground lichen community and this may partly explain why we saw no significant effects for the latter.

One can only speculate as to the mechanisms underlying the observed associations between atmospheric deposition and forest understory vegetation. The fact that regressions were stronger with the deposition variables than with distance from source reassures us that the observed trends are not due simply to some confounding effects of geographic location. Given the evidence that any potential acidification effects of nitrogen and sulphur deposition are buffered by base cation deposition (Watmough et al., 2014; Davies et al., Chapter 4), it is possible that there is a nitrogen and sulphur fertilization effect occurring (Greaver et al., 2012). Likewise, the base cation inputs could themselves have a fertilization effect. Jung and Chang (2012) found

139

that four years of simulated nitrogen and sulphur deposition in the Athabasca Oil Sands Region decreased soil exchangeable calcium and magnesium; if base cation deposition balances nitrogen and sulphur deposition in this region then the concerns Jung and Chang (2012) raised about nutrient imbalances may be allayed.

Table 9. 2 Results of constrained ordination of the understory vegetation composition (percent cover for each vascular species and for all mosses combined and all lichens combined).Given is the F value and significance of the overall model and for the two significant axes. Also given is the percent variation explained by the overall model and by the first two axes. The adjusted R2 value of the overall model is also given. In the lower portion of the table are listed the independent variables1 that were chosen by means of backwards selection and their scores on the first two ordination axes. See also Figure 10.2. % Variation F P explained Overall 4.762 0.001 75.2 Axis1 26.577 0.001 59.9 Axis2 3.802 0.009 8.6 adjusted R2 = 0.593982 RDA1 RDA2 CALPUFF PAI (bulk BC)1 -0.382 -0.275 CALPUFF BC (bulk)1 0.252 -0.097 CALPUFF BC (throughfall)1 0.209 -0.350 CALPUFF S deposition1 -0.199 -0.703 2 IER SO4 (bulk) -0.220 -0.785 IER DIN (bulk)2 -0.218 -0.784 IER BC (bulk)2 -0.219 -0.781 1 Deposition variables based on emission sources and modelled using CALPUFF (from Davies et al. Chapter 4 in this report): CALPUFF PAI (bulk) = Potential Acid Input (PAI) – bulk base cation assumption; CALPUFF BC (bulk) = Base Cation – bulk assumption; CALPUFF BC (throughfall) = Base Cation (throughfall assumption); CALPUFF S deposition = Total S deposition (all sources) 2 Deposition variables modelled based on data from ion exchange resins (IER) (from Fenn, Chapter 3 this report): IER SO4 bulk = SO4-S deposition (bulk), IER DIN (bulk) = Dissolved Inorganic Nitrogen (bulk), IER BC(bulk) = Base Cations (bulk).

140

Figure 9. 2 Results of constrained ordination (Redundancy Analysis) of understory vegetation composition on atmospheric deposition variables, as chosen by backward selection. Symbols represent sites (with labels in the bottom panel). In the top panel the deposition variables are shown as vectors. Deposition variables based on emission sources and modelled using CALPUFF (from Davies et al. Chapter 4): PAI.bulk = CALPUFF Total Potential Acid Input – bulk base cation (BC) assumption , BC.bulk = CALPUFF Base Cation (bulk assumption), BC.TF = CALPUFF Base Cation (throughfall assumption), Total S = CALPUFF Total S deposition (See Appendix 9.1); Deposition variables modelled based on data from ion exchange resins (IER) (from Fenn Chapter 3): SO4 bulk = IER SO4-S deposition (bulk), DIN.bulk = IER Dissolved Inorganic N (bulk), BC.bulk2 = IER Base Cations (bulk) (see Appendix 9.2). In the bottom panel the most strongly discriminating species are shown as four letter codes (genus and species) enclosed in a box. See also Appendix 9.3.

141

Further, their experiment demonstrated that the forests in this region are nitrogen-limited, and this is likely even more true for the poor, dry jack pine forests we sampled. We must consider the possibility that the observed relationships could be caused by other factors that are correlated with the deposition variables. For example, it is possible that atmospheric CO2 enrichment derived from industrial sources creates a gradient similar to that for atmospheric depositions and that this could be causing a CO2 fertilization effect on forest vegetation (Jung et al., 2012).

It is important to note that the 19 sites included in these analyses were not perfect ecological analogues of one another. Seventeen fell into two ecological analogue types (EAT 2 or 3) while JP101 and JP315 were in EAT 1 and 7, respectively (Jaques and Legge, 2012). This could have influenced the results. JP101 had the lowest vascular cover and second lowest forb cover while JP315 had the fourth highest vascular and forb cover. However, both sites were in the mid-range for all the richness variables that showed a significant relationship with the deposition variables and they were fairly centrally located in the ordination as well.

The observed relationships apply only to these dry, poor jack pine forests. Other forest types in the region could well respond differently, although studies in mixed wood (aspen – white spruce) forests in the region suggest they also have the capacity to respond positively to nitrogen and sulphur deposition, at least in terms of tree growth (Jung et al., 2012). The effects of higher levels of longer-term exposure, however, are unknown. It is important to realize that even after atmospheric deposition levels are decreased, subsequent vegetation recovery can be very slow (Nordin et al., 2005).

9.4 Conclusions On these dry, poor jack pine sites increasing atmospheric deposition of nitrogen, sulphur and base cations was associated with increased cover and richness of vascular plants and a shift in species composition. In particular there was a notable increase in cover of bearberry, a common trailing shrub. Increasing deposition was not, however, associated with lower cover of ground lichens; all sites were still heavily dominated by the common reindeer lichen Cladina mitis. Overall, the results are suggestive of a fertilization effect of atmospheric deposition. Although vegetation responses associated with deposition were minor, further increases in deposition or accumulation over longer time periods could result in more dramatic changes in understory vegetation – such as increased grass cover, more dramatic increases in vascular plant cover and richness, and reduced abundance and richness of mosses and lichens. This would change the fundamental nature of these forests.

Both the CALPUFF and IER deposition variables gave similar results. Among the CALPUFF deposition variables Potential Acid Input with a base cation bulk assumption seemed to have the strongest predictive power. For the IER deposition variables it was IER Base Cations (throughfall) that was the strongest. There was no evidence of negative effects of acidification on the understory plant community and this is likely due to the buffering effect of base cation depositions.

142

9.5 References Beckingham, J.D. and Archibald, J.H. 1996. Field guide to ecosites of Northern Alberta. Canadian Forest Service, Northwest Region, Northern Forestry Centre. Special Report 5.

Bobbink, R., Hicks, K., Galloway, J., Spranger, T., Alkemade, R., Ashmore, M., Bustamante, M., Cinderby, S., Davidson, E., Dentener, F., Emmett, B., Erisman, J.-W., Fenn, M., Gilliam, F., Nordin, A., Pardo, L., and De Vries, W. 2010. Global assessment of nitrogen deposition effects on terrestrial plant diversity: a synthesis. Ecological Applications 20: 30-59.

Clark, C., Morefield, P., Gilliam, F., and Pardo, L. 2013. Estimated losses of plant biodiversity in the United States form historical N deposition (1985 - 2010). Ecology 94: 1441-1448.

Gilliam, F., and Roberts, M. 2003. The herbaceous layer in forests of eastern North America. Oxford University Press, New York, New York, USA.

Gilliam, F. 2006. Responses of the herbaceous layer of forest ecosystems to excess nitrogen deposition. J. Ecol. 94: 1176-1191.

Greaver, T., Sullivan, T., Herrick, J., Barber, M., Baron, J., Crosby, B., Deerhake, M., Dennis, R., Dubois, J.-J., Goodale, C., Herlihy, A., Larence, G., Liu, L., Lynch, J., and Novak, K. 2012. Ecologciale ffects of nitrogen and sulfur air pollution in the US: what do we know? Frontiers in Ecol. Env. 10: 365-372.

Hallbäcken, L., and Zhang, L.-Q. 1998. Effects of experimental acidification, nitrogen addition and liming on ground vegetation in a mature stand of Norway spruce (Picea abies (l.) Karst.) in SE Sweden. Forest Ecol. Mgmt. 108: 201-213.

Jaques, D.R., and Legge, A.H. 2012. Ecological analogues for biomonitoring industrial sulfur emission in the Athabasca oil sands region, Alberta, Canada. pp 219-241. IN K.E. Percy (Ed.) Alberta Oil Sands: Energy, Industry and the Environment. Elsevier, Oxford, UK.

Jung, K., Choi, W.-J., Chang, S., and Arshad, M. 2012. Soil and tree ring chemistry of Pinus banksiana and Populus tremuloides stands as indicators of changes in atmospheric environments in the oil sands region of Alberta, Canada. Ecol. Indicators 25: 256-265.

Jung, K., Chang, S.X., Ok, Y.S., and Arshad, M.A. 2013. Critical loads and H+ budgets of forest soils affected by air pollution from oil sands mining in Alberta, Canada. Atmos. Environ. 69: 56-64.

143

Mäkipää, R. 1995. Sensitivity of forest-floor mosses in boreal forests to nitrogen and sulphur deposition. Water Air Soil Poll. 85: 1239-1244.

Mäkipää, R. 1998. Sensitivity of understorey vegetation to nitrogen and sulphur deposition in a spruce stand. Ecol. Engineer. 10: 87-95.

Nordin, A., Strengbom, J., Witzell, J., Nasholm, T., and Ericson, L. 2005. Nitrogen deposition and the biodiversity of boreal forests: Implications for the nitrogen critical load. Ambio, 34: 20-24.

R Core Development Team. 2014. R: A Language and Environment for Statistical Computing R Foundation for Statistical Computing. Vienna, Austria. www.R-project.org

Oksanen, J., Blanchet, F. G., Kindt, R., Legendre, P., O’Hara, P. R. M. R. B., Simpson, G. L., Solymos, P., Stevens, M. H. H., and Wagner, H. 2013. Vegan: community ecology package. R package version 2.0-7.

Olsson, B.A., and Kellner, O. 2006. Long-term effects of nitgen fertilization on ground vegetatin in coniferous forests. Forest Ecol. Mgmt. 237: 458-470.

144

Chapter 10: Chemical composition and lichen community change in the AOSR Keith Puckett, ECOFIN, Waldemar, ON

10.1 Introduction The purpose of the Terrestrial Environmental Effects Monitoring [TEEM] Program is to detect, characterize and quantify the impact that air emissions have had or may have in the longer term, on terrestrial ecosystems and on traditional land resources. One of the approaches identified in the TEEM Strategic plan is the measurement of environment conditions (see Chapter 1 of this report) using various indicators and endpoints to answer the questions: what has changed, where has it changed, how has it changed, is the change biologically significant? In the AOSR case, lichen can be used as: 1) bioindicators of dispersion and deposition of air pollutants - involving the sampling and chemical analysis of the lichens Hypogymnia physodes and Evernia mesomorpha, as well as 2) indicators of the impacts of air quality- involving describing changes in the epiphytic macro-lichen community. The two target epiphytic lichen species are abundant throughout the region, are pollution tolerant and have been used extensively, both in the AOSR and elsewhere to map the distribution of pollutants.

Earlier work using lichens to describe the deposition and impact of air pollutants in the AOSR has been reported by Douglas and Skorepa (1976), Wylie (1978), Addison and Puckett (1980) and by Addison (1984), Dobbs (1985), and Pauls et al. (1996). These studies as well as those referenced below all involve measuring the level of pollutants in the lichens as an indirect means of describing air pollutant dispersion and deposition patterns. The same two lichens were included in the initial phase of the TEEM Forest Health Monitoring Program with lichen collection and chemical analysis in 2004 (C. E. Jones and Associates, 2007). In addition a pilot study in 2004 (Berryman et al., 2004) and a comprehensive survey focused on source identification in 2008 (Berryman et al., 2010) using the same two lichens has added considerable information about lichen element levels in the AOSR. Another lichen [Usnea laponica], was included in earlier Forest Health Monitoring Program activities (AGRA, 1999; AMEC, 2001).

This chapter describes the results of the chemical analyses of Hypogymnia physodes in 2011. The status of the lichen communities was not assessed in 2011. Nevertheless, the methodologies and results from previous assessments of lichen vigour, growth and community structure are included for information.

10.2 Methods 10.2.1 Lichen element levels Lichen element levels were determined by following the methodology outlined in the WBEA TEEM Forest Health Monitoring Program Procedure Manual (Foster et al., 2015) with modifications as described in Puckett et al., 2015 [In preparation]. In essence, the lichen

145

Hypogymia physodes was collected from branches of Pinus banksiana at 22 sites (Fig. 10.1). Samples were cleaned and analysed for 46 elements (Appendix 10.1) [For details, see Puckett et al., 2015. In preparation]. Only the data for 25 elements are discussed in this report.

10.2.2 Epiphytic lichen community composition Changes in lichen community composition may occur in response to exposure to air contaminants, as abundance of sensitive lichen species decreases and tolerant species abundances increase. Epiphytic lichens, reliant on the air for nutrient and water supply, are likely the most susceptible to changes in the quality of the air. Examination of the lichen community in previous years (2004) was carried out following the methodology outlined in the WBEA TEEM Forest Health Monitoring Program Procedure Manual. In essence all epiphytic macro-lichens (excluding crustose species) within the vegetation plot are identified by common and scientific (genus, species) names. The abundance of each species within the vegetation plot was estimated using the R6 abundance scoring scale (Table 10.1). In some instances, lichen vigour was assessed by measurements of the length and width of 20 of the largest individuals at each site. In addition, the absence or presence of three indicators of poor health (morphological responses to pollution stress), were also determined. The three indicators were: hyper abundance of asexual reproductive structures, presence of parasitic fungi and necrosis and/or discolouration of the thallus.

Figure 10. 1 Study site locations in the AOSR

146

Table 10. 1 Lichen Abundance Scoring Scale Code Abundance 1 Rare (<3 individuals in the plot) 2 Uncommon (4 to 10 individuals in the plot) 3 Common (11 to 40 individuals in the plot) Very common (>40 individuals in the plot, fewer than ½ of the branches covered), sub- 4-X scored as: 4-1 Individuals few (40 to 80) and widely scattered in the plot 4-2 More than 80 individuals, concentrated on a few trees 4-3 Many trees have up to 20 individuals 4-4 Many trees have more than 20 individuals 4-5 More than ½ of the trees have up to 20 individuals 4-6 More than ½ of the trees have more than 20 individuals 5 Abundant (more than ½ of the branches covered)

10.2.3 Elemental concentrations in lichen Element levels in H. physodes were not evenly distributed across the study area. Most elements show a marked relationship with distance from the mining/upgrading activities. For the most part, lichen element concentrations increase with increasing proximity to the industrial sources and the strength of the relationship varies from one element to another. For some elements, concentrations decreased with increasing proximity and for other elements, there was no clear relationship with distance.

Element concentrations were regressed against distance with the point of origin being the mid- point between the largest upgrading operations. The elements that showed the highest power law fit (R2 > 0.65) for the relationship between concentrations and distance are aluminium, arsenic, chromium, cobalt, iron, molybdenum, nickel, silicon, sodium, titanium and vanadium (Figure 10.2) suggesting that source proximity is the major determinant of the lichen element composition.

147

148

Figure 10. 2 Al, As, Cr, Co, Fe, Mo, Ni, Si, Na, Ti, and V in H. physodes as a function of distance from industrial area.

Weaker power law relationships (R2 > 0. 35-0.51) were seen for the elements calcium, copper, magnesium, potassium, strontium and sulphur. (Figure 10.3).

149

Figure 10. 3 Ca, Cu, Mg, K, Sr and S concentrations in H. physodes as a function of distance from industrial area.

The power law fit was poor (R2< 0.15) with distance for barium, lead, nitrogen and phosphorus Figure 10.4.

150

Figure 10. 4 Ba, Pb, N, and P concentrations in H. physodes as a function of distance from industrial area.

In contrast, concentrations of cadmium, manganese, mercury, and zinc showed a tendency to decrease with increasing proximity to the industrial complex (Figure 10.5).

Comparison of the relationships with distance for the various elements in this study with those reported by Berryman et al. (2004) and Graney et al. (2012) shows marked similarities which are further examined in the Discussion.

For those elements that showed higher concentrations closer to the sources, a comparison between concentrations at sites closer to the sources (<25 km) and concentrations to sites further removed (> 100 km) showed differences ranging from 1.2 to 5.4 fold increases. Comparison with the results of Edgerton et al. (2012) who also examined concentration differences between sites closest and furthest away from the sources showed that for the most part the elements showed very similar distal/proximal ratios in both collection years (2008, 2011) with the exception of aluminium, chromium, cobalt, iron, molybdenum, silicon, titanium and vanadium which showed higher proximal/distal ratios in 2011 indicating that deposition and/or uptake of these elements have increased over the intervening three years.

151

Figure 10. 5 Cd, Mn, Hg, and Z concentrations in H. physodes as a function of distance from industrial area.

10.2.4 Lichen community structure at sampling sites While lichen vigour and community structure were not assessed in 2011, reductions in lichen vigour have been recorded in all previous surveys of lichen health (Table 11.2).

Table 10. 2 List of lichen impact studies in the Athabasca Oil Sands Region. Year [Sites] Lichen Author Parameter Response Douglas and Not reported Re- 1975 [12] H.physodes Skorepa (1976) Cover, frequency analysis changes in 1976 [56] Parmelia sulcata Peterson and vigor Douglas (1977) Epiphytic lichen Cover, frequency, 1977 [11] Addison (1979a,b) No change assemblages vigor Lulman et al., No change 1979 [12/56] Parmelia sulcate Vigor, growth (1980) No change Addison and 1978 [69] E. mesomorpha Vigor Changes in vigor Puckett (1980) H.physodes 1981 [11] E. mesomorpha Case (1982) Vigor Changes in vigor C. mitis Changes in vigor 1983 [56/38] P.sulcata Dabbs et al., (1985) Vigor, growth Reduced growth H. physodes Berryman at al., 2002 [44] Vigor Changes in vigor E. mesomorpha (2004) E. mesomorpha C.E. Jones and 2004 [13] Vigor Changes in vigor Bryoria furcellata Associates (2007) Changes in vigor, E. mesomorpha Wieder et al., Vigor, Species no significant 2010 Lichen community (2013) number differences between sites

152

The earliest study of the impacts on lichens was by Douglas and Skorepa (1976). In their feasibility study, cover and frequency estimates were made for all species but there was no assessment of vigour or community structure. In this context, vigour is a qualitative estimate of the wellbeing of the lichens and poor vigour is seen as change in the lichen morphology (matting, increased production of asexual propagules, change in colour, dwarfism etc.). Following the feasibility study, a network to establish baseline lichen conditions of 56 permanent plots symmetrically radiating out from the periphery of the Syncrude lease, was established in the summer of 1976 (Peterson and Douglas, 1977) .

In 1983, 38 of the original 56 sites identified by Peterson and Douglas were resurveyed (Case et al., 1985; Dabbs, 1985). Lichen performance was judged in the terms of changes in both vigour and growth. A subjective assessment of the lichen vitality was made based on field observations and quadrat photographs. A vigour rating was assigned to each site on a scale of 0-3 where 0 indicated absence, 1 indicated morbidity, 2 indicated visible injury and 3 indicated the lack of injury. General vitality improved as the distance from the oil sands plants increased. Also, re- examination of the 1976 observations by Dabbs (1985) showed that a similar relationship also existed for the lichens examined in 1976 and that general vitality within a 15 km radius had decreased even further from 1976 to 1983.

Growth was measured as the difference in the radius of individual lichen thalli in 1976 and 1983 as determined by a comparison of aerial photographs of the same thallus in both years. Positive growth increments were restricted to the west, north and south and a zone of negative growth was associated with the locations of the oil sands extraction plants. There were also zones of reduced growth to the east of the plants which seem to be associated with topographic heights of land downwind of the emission sources.

Addison (1980a) established lichen biomonitoring plots in 13 jack pine stands in 1976-77. Sites were situated 2.8 to 80km from the SUNCOR lease with 3 sites within 5 km. In the vicinity of each site, 10 spruce branches with a lichen community of mainly fruticose and foliose types were selected. Examining lichen community structure and composition in 1976-77 provided no evidence of changes in vigour. Biomonitoring plots established in 1976-77 were revisited in 1979 (Addison, 1980b). No significant differences in cover between 1976-77 and 1979 were consistently found for any species group at any site.

A survey of the vigour of Evernia mesomorpha collected from 69 sites (black spruce) in the vicinity of GCOS [SUNCOR] operations (Addison and Baker, 1979; Addison and Puckett, 1980) involved assessing thalli on a scale from 1 (degraded) to 5 (healthy) to describe sample condition. Vitality was poorer along a north south axis running along the Athabasca River valley and in the easterly and westerly directions, poorer vitality was restricted to within a few km. of the river valley. Indeed, there was a marked distribution of thallus condition with the least luxuriant thalli being found closer (< 10 km) to the industrial complex and more luxuriant material being found further away.

153

Forest monitoring efforts focusing on jack pine stands in 2004 included information on lichen communities and vitality (C.E. Jones and Associates, 2007). The cover of the ground was assessed at each of 13 jack pine stands. The lichen community was dominated by Cladina species and the cover estimates varied from 17-72%. However, the differences in cover very likely reflected differences in site factors (C.E. Jones and Associates, 2007) rather than differences in cumulative emissions deposition as there was no clear pattern that could be linked with distance from the mining/upgrading activities.

This study also addressed the vigour of certain epiphytic lichen species and assessed arboreal lichen communities. Vigour was addressed by the measurement of width and length of individual thalli of E. mesomorpha and Bryoria furcellata. In addition to length and width measurements, the presence or absence of three indicators of poor health [morphological responses] to pollution stress was determined. Observations on vigour suggested that lichens at certain sites showed signs of stress [elongated lichens with evidence of parasitism and hyper- asexual reproduction] while lichens from other sites appeared healthy [short, broader thalli that were not hyper- reproductive]. The remaining sites showed intermediate characteristics (C.E. Jones and Associates, 2007). In addition, there was no clear relationship between lichen vigour and distance from the oil sands processing plants. Lichen community structure showed remarkable similarity across all sites; the same 11 macrolichen species were present at every site and the variation in abundance between sites was low allowing the conclusion that local industrial emissions did not appear to have influenced the distribution of epiphytic lichens at these sites.

In 2002, as part of a depositional gradient pilot study, Berryman et al. (2004) surveyed lichen communities at 44 sites located along 4 transects radiating in each cardinal direction from a point equidistant between the Suncor and Syncrude operations. Patterns of lichen community were weakly correlated with distance from the oil sands facilities (Berryman et al., 2004). It appears that air emissions from the mining operations have little impact on epiphytic macrolichen community species diversity and richness in the AOSR. Some lichens were less abundant but not absent at sites close to oil sands mining and extraction facilities although the less abundant lichens were often dwarfed and showed signs of stress at sites close to the industrial operations. In terms of vigour, lichens found closer to the industrial operations showed signs of morphological change including dwarfed growth, hyper-growth of asexual structures, discolouration of the lichen thallus, and increased parasitism by fungi. However, it was not possible to establish a clear distance cut–off after which lichens showed no morphological change.

In 2010, Wieder et al. (2013) examined the vigour and community structure of lichens from bog communities at 5 sites located between 12 and 251 km from the oil sands processing plants. Qualitative observations on the condition of E. mesomorpha at sites closer to the oil sand processing plants showed that the lichen was showing clear signs of stress expressed as dwarfed growth and hypersorediate thalli (excessive production of asexual structures). However, the chlorophyll content, another measure of vigour, was found to be higher in sites closer to the oil sands processing complex and this increase was attributed to the higher nitrogen levels in the

154

lichen. The number of species varied among sites, but did not differ significantly amongst sites. However, total lichen cover was lowest at the Mildred Lake site, the site with the highest N and S deposition and where lichen condition was observed as poor based on morphology.

10.3 Discussion 10.3.1 Spatial variation in lichen element concentrations In 2011, higher concentrations for some elements were found at sites closer to the industrial complex than at sites far removed from the mining and upgrading activities. This has been a consistent feature of previous studies in the AOSR over the past 30 years (Addison and Puckett, 1980; Pauls et al., 1996; Berryman et al., 2004, 2010; and, Edgerton et al., 2012).

The elements aluminium, arsenic, chromium, cobalt, iron, molybdenum, nickel, silicon, sodium, titanium and vanadium in H.physodes showed a strong relationship with distance. This group of elements was very highly inter-correlated indicative of common sources and/or delivery mechanisms. Berryman et al. (2004) in analyzing the element content of H.physodes and E. mesomorpha describe an ordination approach which gave a two dimensional solution with one axis representing the distance gradient from the oil sands operations. The elements most correlated with this first axis were aluminium, chromium, iron, molybdenum, nickel, silicon and vanadium. Earlier Pauls et al. (1996) noted that the aluminium, iron, nickel, titanium and vanadium in E. mesomorpha and Usnea spp. showed a strong trend of decreasing concentration with increasing distance from the oil sand plant developments. More recently, Jaques (2011) showed that aluminum, arsenic, calcium, cobalt, iron, lead, magnesium, nickel, sodium, strontium, and titanium in Cladina mitis increased with proximity to the oil sand operation. The strong influence of proximity to the oil sands development on the concentrations of these element concentrations indicates that the oil sands industrial activities are a significant source.

A weaker relationship (> R2 0.35) with distance was noted for the elements calcium, copper, magnesium, potassium, and strontium which as a group were highly inter correlated (r2> 0.85). This grouping is very similar to the second of the axes identified by Berryman et al. (2004) and represented by magnesium, calcium, strontium and barium. This second axis explained little variation and was not strongly related to distance from the mines. Also, Pauls et al. (1996) found a weak but statistically significant trend towards lower concentrations with increasing distance from the mines/upgraders for barium, copper, and strontium in Evernia mesomorpha.

It should be noted that some elements, notably, barium, calcium, potassium, magnesium, lead, strontium and zinc are correlated with both axes identified by Berryman et al. (2004). This is in keeping with the observation in this study that while calcium, magnesium, potassium and strontium are highly inter correlated (r2 > 0.85) these elements also showed strong correlations (r2> 0.75) with aluminium, iron, silicon and titanium etc., indicating that while proximity is a determinant, other factors play a role in influencing the variation in these elements.

155

A weaker power law relationship (R2 0.44) was also seen for sulphur which showed weak correlations with both axes identified by Berryman et al. (2004). Decreases in sulphur concentrations in lichens with distance have been described by Addison (1980 a, b); Pauls et al. (1996); AMEC (2001) and, Davies (2012) with the highest concentrations being recorded within a nominal 20 km radius of the mining/upgrading complex (Proemse and Mayer, 2012; Davies, 2012). Sulphur is only weakly correlated with aluminium, iron, nickel, titanium and vanadium consistent with the view that sulphur is largely emitted in a gaseous form (Landis et al., 2012) whereas aluminium, iron and other elements are accumulated as particulates [see below]. Sulphur being emitted as gaseous sulphur dioxide is thought to be generally transported longer distances than particulates (Landis et al., 2012) but high sulphur concentrations in this study seem to be limited to within 20 km of the mining/upgrading complex and there does not appear to be any evidence of a more gradual decline in sulphur concentration consistent with the suggested longer range dispersion of sulphur dioxide.

Of the elements, only nitrogen showed a strong correlation with sulphur and the power law fit with respect to distance was also poor (R2 = 0.14). Nitrogen concentrations showed little response to increased proximity to the source region except for collection sites close (<20 km) to the mining/upgrading complex. In previous studies, nitrogen levels in H. physodes were at their highest within 10 km of the mines (Berryman et al., 2004; Proemse and Mayer, 2012; Davies, 2012) and concentrations did not change with increased distance.

Two other elements, lead and phosphorus, also showed very weak power relationships [≈0.10] implying that proximity to mining/upgrading activities is not the only factor or even a factor influencing these element concentrations in H.physodes. Berryman et al. (2004) and C.E Jones and Associates (2007) also found that lead concentrations were not strongly linked to distance.

However, Graney et al. (2012) points out that the high power law function fit for the lead isotope ratio versus distance plot suggested that proximity is a major determinant of isotopic lead in AOSR lichens, indicating that the oil sands extraction and processing activities do contribute to the lichen lead content but that other source contributions mask the relationship with distance.

For phosphorus, as for lead, the power law relationship was weak (R2. 0.10). Phosphorus showed no relationship with distance for H.physodes and E. mesomorpha in 2002 (Berryman et al., 2004) and 2004 (C.E. Jones and Associates, 2007) and even earlier (E. mesomorpha; Pauls et al., 1996). Hence, distance is not a key determinant of the phosphorus concentration. Also, phosphorus, like lead, was not highly correlated with any other element in 2011.

The metals cadmium, manganese, mercury and zinc show a different response in 2011 with reduced concentration levels with increased proximity to the source region. This accumulation pattern is not consistent from year to year, sometimes being very apparent and other times difficult to detect.

156

With respect to cadmium and manganese, this is not a new observation in that Berryman et al. (2004), and Edgerton et al. (2012) have pointed out this feature for Cd and Mn as have Pauls et al. (1996) for Mn. However, a similar pattern was not seen by Jaques (2011) using the ground lichen Cladina mitis. For cadmium, the power law fit is not very good but nevertheless the pattern seems to be consistent over time. Cadmium was also consistently correlated with manganese and zinc in 2004 and 2011.

For manganese the power law fit is reasonable and consistent over time and manganese concentrations are correlated with cadmium and zinc. Lichens found on the forest floor, such as Cladina mitis, also showed lower manganese concentrations closer to the industrial activity (Berryman et al., 2004).

This response for cadmium and manganese has now been described at different times, shown by different lichen species and is not a function of the analytical technique. The response appears to be unique to the AOSR and this pattern may not be significant in terms of describing impacts on forest health in general, nevertheless it may be useful to understand the reasons for this response in order to better understand the utility of using H.physodes as a biomonitor.

Similarly for mercury, concentrations in H. physodes in 2011 are lowest at sites closest to the Oil Sands industrial activity and increase with increasing distance from the bitumen upgrading sites. This response was observed earlier by Blum et al. (2012). However, surveys of the mercury content on H.physodes in 2002 (Berryman et al., 2004) and 2004 (Jones and Associates, 2007) showed no obvious relationship with distance in contrast to the observations in 2008 and the current study. In addition, mercury levels in E. mesomorpha increased with increasing distance from the mines in 2002 but the report cautions care in the interpretation given the high analytical variation (Berryman et al., 2004) and there was no obvious relationship with distance in 2004 for the same lichen (C.E. Jones and Associates, 2007). In 2011, mercury concentrations were not significantly correlated with any other element.

The relationship with distance for zinc in different lichen species and at different times is variable. This study shows a weak inverse relationship for H. physodes with distance. Berryman et al. (2004) refer to suggestive evidence that zinc levels in H.physodes in 2002 also increased with distance from the mines. There was no relationship with distance for either E. mesomorpha or H.physodes in 2004 (C.E. Jones and Associates, 2007). An inconsistent response was observed for E. mesomorpha and Usnea spp. with the inverse pattern only apparent in some sampling years (Pauls et al., 1996). In general, zinc concentrations are correlated with cadmium and manganese. Given the variable response it seems reasonable to conclude that distance is not a significant factor in determining zinc concentrations in H. physodes.

The observed pattern shown by cadmium, manganese, mercury and zinc may be reflection of both how the elements are deposited to the lichen thallus and the retention and loss characteristics of H. physodes. In terms of deposition, manganese is found primarily in the coarse fraction of ambient particulate matter whereas cadmium is found in the fine fraction. Both are

157

found in the regional snowpack, in both particulate and water-soluble form. In terms of lichen specificity, it is well documented that some lichens accumulate more elements than others at a given site (Chiarenzelli et al., 2001). A comparison of the element content of H. physodes and E. mesomorpha shows distinct differences [see below] with H. physodes accumulating much higher concentrations of cadmium, manganese and mercury which suggest that lichen morphology/ chemistry may explain in part the reduced concentrations close to the industrial activity.

10.3.2 Causality - Linkage between changes in vigour, growth, community structure and air quality in the AOSR In general, concentrations of some elements [e.g. aluminium] in lichens in the AOSR are approaching what are judged to be unusual levels (Bennett, 2000) with the implication that these levels may be impacting lichen well-being.

Unlike higher plants, lichens do not possess a cuticle or roots and lichens accumulate elements directly from the atmosphere via wet and dry deposition to levels far greater than their expected physiological needs and indeed lichens growing directly on mineral deposits accumulate and tolerate very high metal concentrations. Purvis and Pawlik-Skowronska (2008) review the tolerance mechanisms to both naturally and anthropogenically derived metals which include the production of lichen substances [lichen-specific carbon based secondary products], metal oxalates, melanin pigments and organic phosphates as well as potential detoxification mechanisms acting at the cellular level. Localised calcium deposits were identified as a possible sink for metals in H. physodes growing in polluted areas (Budka et al., 2002)

Morphological abnormalities in H. physodes and other lichens have been linked to elevated metal levels (Scott and Hutchinson, 1990; Otnyukova, 2007). However, there is only limited evidence that the current levels of elements in H. physodes are impacting lichen vigour, growth and community structure in the AOSR (Puckett 2015 in preparation). Indeed, comparisons of element concentrations in lichens and the associated lichen vigour/growth are few in number.

A relationship between total element levels and poor vigour was described by Case (1982) but no specific elements were identified as being more important than others in explaining the variation in vigour. Addison and Puckett (1980) compared estimates of lichen vigour [E. mesomorpha] and element concentrations and concluded that 22% of the variation in the lichen condition was a function of the lichen sulphur concentration and that vanadium was responsible for a further 6%. An additional 18% was linked to the presence of varying aluminium and titanium levels. C.E. Jones and Associates (2007) introduced physical measurements of thallus breadth and length into the assessment of vigour offering a more quantitative assessment. However there was no relationship between lichen condition and element concentration other than a link to lead which is unexpected given that the lead concentrations were more or less uniform across the sampling domain in 2004.

Another approach to examine causality has been to transplant lichens from remote areas to areas close to the industrial activity and monitor changes in the lichen response to the poorer air quality

158

and greater deposition of elements of concern. Addison (1984) transplanted epiphytic lichens from distant jack pine sites to areas adjacent to the industrial activity and the lichen community was monitored over a three year period. Significant cover reductions of some lichen groups occurred on transplanted branches within 3-8 km of the pollution source. Evernia, Cetraria, and Bryoria groups were reduced whereas the Hypogymnia grouping showed no response over a 3 year measurement period. While changes in cover were observed there were no corresponding measurements of lichen sulphur or metal concentrations in the lichen community.

10.3.3 Causality – Air quality levels known to impair AOSR lichens An additional approach to discussing causality is to consider whether the measured levels of air pollutants and deposition in the AOSR are sufficient to induce changes based on existing relationships between pollutants and their effects in other regions.

The United States Forest Service National Lichens & Air Quality Database and Clearinghouse summarizes literature ratings based on field observations on whether lichens are sensitive, intermediate or tolerant to a range of gaseous pollutants and deposition. (http://gis.nac se.org/lichenair/?page=sensitivity#Hypphy) and includes information on the response of E. mesomorpha and H. physodes which is summarized below.

Hypogymnia physodes

Based on the summary of literature ratings on field observations (see above), the overall rating to sulphur dioxide was either as intermediate sensitivity or tolerant as H. physodes was present as levels greater than 40ppb. In terms of acid deposition (S and N), H.physodes was rated as intermediate to tolerant: (Farmer et al., 1992).

For nitrogen, there was no information on the response to ambient NOX concentrations. For nitrogen deposition, the evidence suggested that H. physodes was a eutroph, a grouping which tolerates and is usually enhanced by N deposition loads above 4.5 kg N/ha/yr and having a broad N requirement (McCune and Geiser, 2009).

Information on critical levels or loads of particulate matter and specific elements other than nitrogen and sulphur metal is not available.

Evernia mesomorpha

Based on the summary of literature ratings on field observations (see above), the overall rating was sensitive as the lichen was present at 10-23 ppb annual average and absent at 29 ppb (LeBlanc, 1969; LeBlanc et al., 1972a).

159

There is no information on the response of E. mesomorpha to ambient levels of NOx, sulphate/or nitrate deposition, particulate matter or particular elements, such as cadmium, copper or mercury.

10.3.4 Causality - Comparison of measured and predicted air pollutant concentration in the AOSR and lichen response Air quality has been monitored by WBEA in the AOSR since 1997 (Percy, 2013) and there has been little change over that period with the exception of recent increases in ambient NOX concentrations (Kindzierski et al., 2009). The current air quality ambient monitoring data for the region show minimal impacts of oil sands development on regional air quality (Royal Society of Canada, 2010). For example, in 2011, the Alberta ambient air quality objective for sulphur dioxide (annual average 8.0 ppb) was not exceeded (Percy et al., 2012), neither was the annual objective for nitrogen dioxide (annual average 24ppb). A comparison of the current and historical annual sulphur dioxide concentrations with known injury threshold levels show that the annual ambient air concentrations are considerably lower than those known to influence the distribution of E. mesomorpha and H. physodes.

There are no injury threshold levels for nitrogen dioxide for the two lichens so at this time it is not possible to conclude whether or not historical and current ambient NOx concentrations are influencing the lichen wellbeing and distribution. There has been considerable discussion of the response of vegetation and lichens in particular to nitrogen deposition in the context of eutrophication (Pardo et al., 2011; Geiser and Nadelhoffer; 2011). For the AOSR, there is evidence that nitrogen deposition is approaching levels that may impact lichen communities (Berryman and Straker, 2008). However, specific critical loads for AOSR lichens have not been developed.

In conclusion, at the present, comparisons of critical loads and levels for sulphur and nitrogen indicate that current exposure to measured ambient or deposited sulphur is well below those levels expected to impair vigour/growth of the two lichens (E. mesomorpha and H. physodes). For nitrogen, the deposition might be approaching levels that could induce shifts in the lichen community. However, this conclusion is based on a limited number of studies and on the absence over time of systematic and continuous epiphytic community based, comprehensive studies to examine changes in lichen community structure in AOSR and relate observed differences in space and time to changing air quality and deposition.

10.4 Summary A longer term view shows that over the past 40 years, the lichen community has been exposed to varying emissions. In that period, SO2/NOX and particulate matter emissions have changed, influenced both by increasing production and changing emission control technology. Increased element concentrations and decreased lichen vigour were seen in the lichen community within 10 years of the onset (1967) of mining/upgrading activities in the AOSR (Addison, 1984; Case et al., 1985) and these two observations have been a constant feature in subsequent studies until

160

the present day (Wieder et al., 2013). In addition, and in the absence of more systematic studies, there is no evidence that the impacts have increased over the intervening years.

Impacts of air quality on vegetation are sometimes described in terms of a continuum or cascade of events (National Research Council, 1989). The first stage is this cascade is an increase in concentration of pollutants in the plant. For lichens this stage is well documented (Addison and Puckett, 1980; Pauls et al., 1996; Berryman et al., 2004). The second stage involves change in the wellbeing of the individual plants. Again, this is well documented for lichens in the AOSR with observations of reduced vigour (e.g. C.E. Jones and Associates, 2007). The third stage involved impacts on the community. At this time the impacts continuum does not seem to have developed to this stage as measured by changes in species number in the jack pine sites examined in the studies to date.

Several factors preclude more definitive statements on the lichen response to changing air quality in the AOSR. The lack of a consistent methodology over the past 4 decades makes it difficult to compare results at different times during this period. This discontinuity is due in part to the inability to maintain appropriate monitoring activities. Several networks have been initiated with subsequent sampling in some instances but unfortunately these networks were abandoned as attention and resources were needed elsewhere.

However, NOX emissions are projected to increase significantly and the issue of nitrogen fertilization may become more significant. In this instance the more appropriate measure of change in the lichen community is not species number or disappearance, but a shift in community composition to reflect lichens which are more tolerant of higher nitrogen deposition (Johansson et al., 2012). However, the absence of ongoing comprehensive, consistent and continuous monitoring programs will preclude further commentary on the impact of changing air quality in the AOSR.

The lichen community description methodology used by C. E. Jones and Associates (2007), and Berryman et al. (2004) has created a baseline for further lichen community surveys. This approach has been used extensively though out the United States but has not been applied to a large extent in the boreal forest. Further consideration of the applicability/sensitivity of this approach in terms of detecting change in the AOSR epiphytic lichen community is warranted. In addition, the issue of quantification and extent of changing lichen vigour remains to be fully addressed, given that it seems to be a widespread and consistent response to changing air quality.

10.5 References Addison, P.A. 1980a. Baseline condition of jack pine biomonitoring plots in the Athabasca Oil Sands Region 1976-1977. For the Alberta Oil Sands Environmental Research Program. Project LS 3.4.2. 42pp.

161

Addison, P.A. 1980b. Ecological bench- marking and biomonitoring for detection of airborne pollutant effects on vegetation and soils. For the Alberta Oil Sands Environmental Research Program. Project LS 3.4. 48pp.

Addison, P.A .1984. Quantification of branch dwelling lichens for the detection of air pollution impact. Lichenologist: 16: 297-304.

Addison, P.A., and J. Baker. 1979. Interim report on ecological bench-marking and biomonitoring for detection of air-borne pollutant effects on vegetation and soils 1975 to 1978. Alberta Oil Sands Environmental Research Project Report. No. 46, 1-41.

Addison, P.A and K.J. Puckett. 1980. Deposition of atmospheric pollutants as measured by lichen element content in the Athabasca Oil Sands Area. Can. J. Bot. 58: 2323-2334.

AMEC Earth and Environmental Limited. 2000. Monitoring the long term effects of acid emissions on soil and vegetation in the jack pine and aspen forests of north eastern Alberta. CE 01742/7000. Submitted to the Wood Buffalo Environmental Association, Fort McMurray, Alberta. 806 pp.

AMEC Earth and Environmental Limited. 2001. Jack pine acid deposition monitoring network site selection. CE 02078. Submitted to the Wood Buffalo Environmental Association, Fort McMurray, Alberta. 51 pp.

AMEC Earth and Environmental Limited. 2002. Establishment of site JP212 as a TEEM Jack Pine site acid deposition monitoring site. CE 02301. Submitted to the Wood Buffalo Environmental Association, Fort McMurray, Alberta. 108 pp.

Bennett, J.P. 2000. Statistical baseline values for chemical elements in the lichen Hypogymnia physodes. In: Environmental Pollution and Plant Responses. Edited by S.H. Agrawal and M. Agrawal. Lewis Publishers. pp 343-353.

Berryman, S., Geiser, L., and Brenner, G. 2004. Depositional gradients of atmospheric pollutants in the Athabasca Oil Sands Region, Alberta, Canada: an analysis of lichen tissue and lichen communities. Lichen Indicator Pilot Program 2002-2003. Report prepared for the Wood Buffalo Environmental Association, Fort McMurray, AB .171pp.

Berryman, S., Straker, J. 2008. Nitrogen loading and terrestrial vegetation—Assessment of existing regional vegetation data and recommendations for future monitoring. Report to the Cumulative Environmental Management Association (CEMA) NOx-SO2 Management Working Group (NSMWG) and Eutrophication Task Group. Sydney, BC, Canada: CE Jones and Associates.

162

Berryman, S., Straker, J., Krupa, S., Davies, M., Ver Hoef, J., and G. Brenner. 2010. Mapping the characteristics of air pollutant deposition patterns in the Athabasca Oil Sands Region using epiphytic lichens as bioindicators. Interim Report Submitted to the Terrestrial Environmental Effects Monitoring (TEEM) Science Sub-committee of the Wood Buffalo Environmental Association (WBEA). Fort McMurray, Alberta, Canada.

Blum, J.D., Johnson, M.W, Gleason, J.D, Demers, J.D., Landis, M.S., and Krupa, S. 2012. Mercury concentration and isotopic composition of epiphytic tree lichens in the Athabasca Oil Sands Region. pp 373-390. IN K.E. Percy (Ed.) Alberta Oil Sands: Energy, Industry and the Environment. Elsevier, Oxford, UK.

Budka, D., Przybylowicz, W.J., Mesjasz-Przybylowicz, J., Sawicka-Kapusta, K. 2002. Nuclear Instruments and Methods. 189: 499.

Chiarenzelli, J., Aspler, L., Dunn, C., Cousens, B., Ozarko, D., Powis, K. 2001. Multi-element and rare earth element composition of lichens, mosses, and vascular plants from the Central Barrenlands, Nunavut. Canada. Applied Geochemistry 16: 245-270.

Case, J.W. 1982. Report on the condition of lichen vegetation in the vicinity of the Syncrude lease. Report to Syncrude Canada, Calgary, AB. 16pp.

Case, J.W., Dabbs, D.L., Nasi, M.E., and Bryant, S.E. 1985. The Syncrude Biomonitoring Network. A Research Report prepared for Syncrude Canada Ltd. Fort McMurray, AB 57pp. and figures. Draft report.

C. E. Jones and Associates Ltd. 2007. Terrestrial Environmental Effects Monitoring Acidification Program. 2004 Sampling event for soils, lichens, understory vegetation and forest health and productivity. Report prepared for the Wood Buffalo Environment Association, Fort McMurray, AB. 729pp.

Dabbs, D.L. 1985. Atmospheric emissions monitoring and vegetation effects in the Athabasca Oil Sands Region. Environmental Research Monograph1985-5. Syncrude Canada Ltd, Fort McMurray, AB.127pp.

Davies, M.J.E. 2012. Air Quality Modelling in the Athabasca Oil Sands Region. pp 267 - 309. IN K.E. Percy (Ed.) Alberta Oil Sands:Energy, Industry and the Environment. Elsevier, Oxford, UK.

Douglas, G.W and A.C. Skorepa. 1976. Monitoring Air Quality with Lichen: A Feasibility Study. Environmental Research Monograph 1976-2. Syncrude Canada Ltd., Fort McMurray, AB 69pp.

Dobbs, D.L. 1985. Atmospheric emissions monitoring and vegetation effects in the Athabasca Oil Sands Region. Ecological Research Monograph 1985-5. Syncrude Canada Ltd. 127pp.

163

Edgerton, E.S., Fort, J. M., Baumann, K., Graney, J.R., Landis, M.S., Berryman, S., and Krupa, S. 2012. Method for extraction and multi-element analysis of the lichen Hypogymnia physodes from the Athabasca Oil Sands Region. pp 315-342. IN K.E. Percy (Ed.) Alberta Oil Sands:Energy, Industry and the Environment. Elsevier, Oxford, UK.

Farmer, A.M., Bates, J.W., Bell, J.N.B. 1992. Ecophysiological effects of acid rain on bryophytes and lichens. pp 284-313. IN Bates, JW/Farmer, AM (eds.): Bryophytes and Lichens in a Changing Environment. Clarendon Press, Oxford.

Foster, K.R., Baines, D., Percy, K., Legge, A., Maynard, D., and Chisholm, V. 2015. WBEA TEEM Forest Health Monitoring Program. Draft Procedures Manual. Version 1. Wood Buffalo Environmental Association. Fort McMurray, Alberta.201pp.

Geiser, L.H., and Nadelhoffer, K. 2011. Taiga, Chapter 6. In: Pardo, L.H.; Robin-Abbott, M.J.; Driscoll, C.T., (Eds). 2011. Assessments of Nitrogen deposition effects and empirical critical loads of nitrogen for ecoregions of the United States. Gen. Tech. Rep. NRS-80. Newtown Square, PA: U.S. Department of Agriculture, Forest Services, Northern Research Station: 49-60.

Graney, J.R., Landis, M.S., and Krupa, S. 2012. Coupling lead isotopic and element concentrations in epiphytic lichens to track sources of air emissions in the Athabasca Oil Sands Region. pp 343- 372. IN K.E. Percy (Ed.) Alberta Oil Sands: Energy, Industry and the Environment. Elsevier, Oxford, UK.

Jaques, D.R. 2011. TEEM. Ecological Analogue Site Selection Program – 2010 Report. Submitted to the Wood Buffalo Environmental Association, Fort McMurray, AB. 67pp.

Johansson, O., Palmqvist, K., Olofsson, J. 2012. Nitrogen deposition drives lichen community changes through differential species responses. Global Change Biology. 8: 2626-2635.

Kindzierski, W. B., Chelme-Ayala, P., and Gamal El-din, M. 2009. Wood Buffalo Environmental Association Air Quality Data Summary and Trend Analysis. Department of Public Health Sciences, University of Alberta, Edmonton, AB.

Landis, M.S., Pancras, J.P., Graney, J.R., Stevens, R.K., Percy K.E., and Krupa, S. 2012. Receptor modeling of epiphytic lichens to elucidate the sources and spatial distribution of inorganic air pollution in the Athabasca Oil Sands Region. pp 427-467. IN K.E. Percy (Ed.) Alberta Oil Sands: Energy, Industry and the Environment. Elsevier, Oxford, UK.

LeBlanc, F. 1969. Epiphytes and air pollution, In: Air Pollution, Proceedings of the First European Congress on the Effects of Air Pollution on Plants and Animals. Wageningen, pp 211-220.

LeBlanc, F., Rao, D.N., Comeau, G. 1972a. The epiphytic vegetation of Populus balsamifera and its significance as an air pollution indicator in Sudbury, Ontario. Can. J. Bot. 50: 519-528.

164

Lulman, P.D., Fessenden, R.J., and McKinnon, S.A. 1980. Lichens as air quality monitors. In: Symposium on Effects of Air Pollutants on Mediterranean and Temperate Forest Ecosystems, June 22-27, 1980, Riverside, California, USA. P.241.

McCune, B., Geiser, L. 2009. Macrolichens of the Pacific Northwest.Oregon State University Press, Corvallis, Oregon.

Pardo L.H., and 22 others. 2011. Effects of nitrogen deposition and empirical nitrogen critical loads for ecoregions of the United States. Ecological Applications 21: 3049-3082.

Pauls, R.W, Abboud, S.A., and Turchenek, L.W. 1996. Pollution deposition impacts on lichens, mosses, wood and soil in the Athabasca Oil Sands area. Syncrude Canada. 222pp.

Percy, K.E. 2013. Geoscience of climate and energy 11. Ambient air quality and linkage to ecosystems in the Athabasca oil sands, Alberta. Geoscience Canada. 40: 182-201.

Percy, K.E., Hansen, M.C., and Dann, T. 2012. Air Quality in the Athabasca Oil Sands Region 2011. pp 47-91. IN K.E. Percy (Ed.) Alberta Oil Sands:Energy, Industry and the Environment. Elsevier, Oxford, UK.

Peterson, W.L., and Douglas, G.W. 1977. Air quality monitoring with a lichen network: Baseline data. Environmental Research Monograph1977-5. Syncrude Canada Ltd., Fort McMurray, AB. 79pp.

Proemse, B.C., and Mayer, B. 2012. Tracing industrial nitrogen and sulfur emissions in the Athabasca oil sands region using stable isotopes. pp 243-266. IN K.E. Percy (Ed.) Alberta Oil Sands: Energy, Industry and the Environment. Elsevier, Oxford, UK.

Puckett, K.J. 2014. Retrospective on the impacts of changing air quality on the lichens of the Athabasca Oil Sands Region. Prepared for the Wood Buffalo Environmental Association, Fort McMurray. 21pp. Draft.

Puckett, K.J., Blum, J.D., Edgerton, E., Graney, J., Landis, M., and Studabaker, W. 2014. TEEM Lichen project report. Spatial and temporal variation in the element and PAH concentrations in the lichen Hypogymnia physodes in the Athabasca Oil Sands Region. Prepared for the Wood Buffalo Environmental Association, Fort McMurray. Draft.

Purvis, O.W., and Pawlik-Skowronska, B. 2008. Lichens and metals. In: Avery S, Stratford, M. can West P (eds) Stress in yeasts and filamentous fungi. Elsevier, Amsterdam, pp 175-200.

Royal Society of Canada. 2010. Environmental and Health Impacts of Canada’s Oil Sands Industry. Report of Expert Panel.

165

Scott, M.G., and Hutchinson. 1990. The use of lichen growth abnormalities as an early warning indicator of forest dieback. Environmental Monitoring and Assessment. 15: 213-218.

Wieder, R.K., Vile, M.A., Vitt, D.H., and Berryman, S. 2013. Development of monitoring protocols for nitrogen- sensitive bog ecosystems including further development of lichen monitoring tools 2009-2012. Report prepared for the Wood Buffalo Environmental Association, Fort McMurray, AB.

Wylie, E. 1978. Diversity and vitality of 10 epiphytic lichens measured over standardized sections of jack pine white spruce and black spruce within an 11 kilometer radius from the Great Canadian Oil Sands Limited Complex in July 1977. Prepared for the Great Canadian Soil Sands Ltd by Loman and Associates. Calgary. 79pp.

166

Chapter 11: The application of critical loads and estimates of exceedance for sulphur and nitrogen deposition to forests in the AOSR Shaun Watmough, Trent University, Peterborough, ON and Colin Whitfield, Centre for Hydrology, University of Saskatchewan, Saskatoon, SK

11.1 Critical loads and their application to forest ecosystems The critical load (CL) of acidity (sulphur (S) + nitrogen (N)) is the maximum load of acidic deposition that an ecosystem can receive during the long-term without damage to specified sensitive biological components (Nilsson and Grennfelt, 1988). For terrestrial ecosystems, critical loads are usually estimated using the Simple Mass Balance (SMB) model (Sverdrup and De Vries, 1994). Critical loads for nutrient nitrogen (CLnut(N)) can also be estimated by a simple mass balance model or by an empirical approach that are determined from observations of detrimental responses of an ecosystem or ecosystem component (e.g. lichen vigour; tree health) published in the peer-reviewed literature (e.g. Bobbink et al., 2010; Pardo et al., 2011; Blett et al., 2014). Critical loads depend upon the ecological receptor and in the SMB the critical limit for acidity is usually based on a molar base cation:aluminum (Al) ratio in soil solution that will not adversely impact tree health (damage roots; see Sverdrup and Warfvinge, 1993; Cronan and Grigal, 1995). For nutrient N the critical limit is usually based on an acceptable level of nitrate leaching below the rooting zone. The SMB has been widely applied to estimate the ecosystem risk from acid deposition in Europe (e.g., De Vries and Reinds, 1994; Hettelingh et al., 2007; Lorenz et al., 2008), the United States (U.S.; e.g., McNulty et al., 2007; McDonnel et al., 2010) and in Canada (e.g., Arp et al., 1996; Foster et al., 2001; Ouimet et al., 2006; Whitfield et al., 2006, 2010) to assess the regional impact of acidic deposition, and serve as a valuable tool in the development of emissions management policies (Environment Canada, 1998).

11.2 Critical load calculations 11.2.1 The simple mass balance model for acidity The SMB model represents sources and sinks of acidity for forest soils (Sverdrup and De Vries, -2 -1 1994) (all units expressed as mmolc m yr ):

CL (S + N) = BCdep – Cldep + BCw – Bcu +Ni + Nu + Nde – ANCle(crit) where: BC refers to base cations (BC = Na+ (sodium) + Ca2+ (calcium) + Mg2+ (magnesium) + K+ (potassium)); the subscripts dep, w, u, i, de and le refer to deposition, weathering, uptake, immobilisation, denitrification and leaching, respectively; Cl = chloride; N = Nitrogen; and ANCle(crit) is the critical (or acceptable) leaching of ANC (Acid Neutralizing Capacity), defined as (UBA, 2004):

167

1 2  Bc Bc Bc  3 Bc Bc Bc  ANC  Q 3 1.5 w dep u  1.5 w dep u  le(crit)   (Bc: Al) K   (Bc: Al)    crit gibb   crit  where Q is the annual water flux through the soil at the bottom of the rooting zone (soil percolation), (Bc:Al)crit is the chemical criterion associated with ecosystem damage. The gibbsite 3+ + equilibrium constant (Kgibb) is used to describe the relationship between Al and H ions in soil solution.

11.2.2 The simple mass balance model for nutrient N. The SMB model takes into account N sinks in forest soils.

CLnut(N ) = CLmin(N) + Nle(acc) / (1 – fde)

Where: CLmin(N) refers to Ni + Nu; Nle(acc) refers to the amount of acceptable nitrate leaching and fde the fraction of atmospherically deposited nitrogen that is denitrified.

Empirical critical loads are determined from observations of detrimental responses of an ecosystem or ecosystem component to an observed nitrogen deposition input that have been published in the peer-reviewed literature. This level of deposition is set as the critical load and extrapolated to other similar ecosystems. For boreal forests a critical load for CLnut(N ) of 5 – 10 -1 -1 -2 -1 kg N ha y (35 – 70 mmolc m yr ) has been recommended (Bobbink et al., 2010; Pardo et al., 2011).

11.3 Previous estimates of critical loads and exceedance for forests in the AOSR The Jack Pine forest in the AOSR is considered particularly sensitive to the impacts of acid and N deposition (Whitfield et al., 2010). Emissions of S and N from industrial activities are high (see Chapters 1-4 for details)and the region has large areas of acid-sensitive sandy soils, which are considered to be N-limited owing to historically low N deposition and a frequent fire-return interval (Laxton et al., 2011). Previous work has shown that upland soils in the region have base cation weathering rates that are extremely low making them potentially vulnerable to acidification (Whitfield et al., 2010). Whitfield et al. (2010) estimated critical loads for acid deposition for 333 sites in the region and reported that under conditions of complete N retention, 34% of the sites receive acid deposition (early 21st century) in excess of their critical load; if all N deposition is leached, 62% of the sites are exceeded. Modelling studies have also suggested that soil solution chemistry will be more responsive to changes from acid deposition than base saturation and that some sites may exceed the critical threshold prescribed under the acid deposition management framework for the region within 30 years (Whitfield et al., 2009). There are very few long-term monitoring data that can be used to assess model predictions, but the few studies that exist indicate that soils are not currently acidifying, and in fact soils may be becoming less acidic (Jung et al., 2013) suggesting model predictions are wrong or other factors

168

such as base cation deposition (Watmough et al., 2014) are not being adequately considered. Critical loads for nutrient N for forests in the region have not been previously estimated.

11.4 Importance of weathering rates, N dynamics, appropriate chemical limits and deposition When determining critical loads and estimating exceedance for forest soils where biomass removals through harvesting do not occur, the key parameters included in the model are 1) base cation weathering, 2) appropriate critical chemical limits, 3) N dynamics and 4) atmospheric deposition of S, N and BC.

Traditionally, estimates of base cation weathering are considered the most uncertain parameter in estimates of critical loads (Hodson and Langan, 1999; Klaminder et al., 2011; Li and McNulty, 2007). Weathering rates have been estimated using a variety of approaches in the AOSR and have shown that despite the approach used, acid-sensitive soils in the region exhibit weathering rates that are among the lowest in Canada. Based on the application of PROFILE to more than 60 sites located on acid-sensitive soils, Watmough et al. (2014) reported that the median base -2 -1 cation weathering rate was just 17 mmolc m yr and more than 80% (52 of 63) of the sites had -2 -1 a base cation weathering rate <40 mmolc m yr . Previous work in the study region (AOSR) used the Zr-depletion method and the Pedological Mass Balance method to estimate base cation weathering rates for 33 soil profiles and reported weathering estimates ranges of 0 – 51 mmolc -2 -1 -2 -1 m yr and 0 – 58 mmolc m yr , respectively (Whitfield et. al., 2011). Similarly, the Soil Texture Approximation method was applied to 290 sites in the region and values of between 0 and 74 -2 -1 -2 -1 mmolc m yr were reported, with a median value of 10 mmolc m yr (Whitfield et al., 2010). Overall, previous studies indicate that weathering rates of the acid-sensitive soils in the region are generally very low and are thus highly sensitive to elevated acid deposition levels.

The choice of an appropriate critical chemical criterion for potential adverse effects is much debated (Skeffington, 2006). Cronan and Grigal (1995) reported that, based on a survey of the literature, that there is a 50% chance of tree death if the Ca:Al molar ratio in soil solution falls below 1.0. While a clearly demonstrable link between tree health and soil solution chemistry has not been established, the ratio of base cations or Ca to Al in soil solution has been the most widely used criteria (UNECE, 2004). The most common criteria are a BC:Al of 1.0 (sum of base cations, BC = Ca2+ + Mg2+ + K+ as Na+ does not provide protection against Al toxicity) or a Ca:Al ratio of 1.0. In Canada, a more conservative approach has been used whereby a critical BC:Al ratio of 10 has been used to maintain soil base saturation over 20% (Ouimet et al., 2006). Consequently, to be consistent with other applications in Canada a BC:Al ratio of 10 is used in the AOSR, which is considered to be conservative.

The estimation of critical loads requires an estimate of the long-term immobilisation of N in soils and the amount of N lost through denitrification. Both terms are uncertain but recommended -2 -1 values for Ni can range from 0–35 mmolc m yr or even higher depending on soil type, soil age, disturbance level and vegetative species present. In well-drained upland soils Nde is minimal. In

169

some cases, a highly conservative approach is used whereby long-term N immobilization is set to zero (Ouimet et al., 2006).

Accurate estimates of atmospheric deposition of S, N and BC are required for the calculation of the critical load (BCdep) and the level of exceedance (Sdep and Ndep). Atmospheric deposition of S, N and BC is very high close (< 20km) to the industrial center in the AOSR, although recent work has shown that base cation deposition is mostly greater than S + N deposition (Watmough et al., 2014). These high base cation deposition levels were not included in previous critical load assessments as the data were not available (Whitfield et al., 2009, 2010). Furthermore it is assumed that most of the base cations that are deposited onto forest ecosystems in the AOSR are derived from industrial activities and thus are subject to temporal change as industrial activities evolve. Because background base cation deposition is very low compared to that originating from industrial sources it is recommended that BCdep be included in the exceedance term as BCdep will likely change over time. Hence the exceedance of the critical load for acidity will increase proportionally to any decrease in atmospheric base cation deposition in the absence of changes in S and N deposition.

Therefore:

CL (S + N) = BCw – Bcu – Cldep +Ni + Nu + Nde – ANCle(crit)

And exceedance is given as:

Ex = CL (S + N) – Sdep – Ndep + BCdep

11.5 Modified application of critical loads to TEEM plots Critical loads and exceedance were estimated for 10 TEEM sites that had estimates of base cation weathering obtained using PROFILE (Whitfield et al., 2009). Deposition data for S, N, Ca2+, Mg2+ and Na+ for each site were obtained from Davies et al. (Chapter 4, this report). In this application K+ and Cl- deposition are excluded as they are unavailable and a relatively conservative approach -1 -1 -2 -1 was used whereby Ni value of 0.2 kg N ha y (1.4 mmolc m yr ) was selected based on Rosen et al. (1992) and which is recommended in the European critical load mapping manual (UNECE, 2004) and Nde was set to zero as denitrification is not expected in well drained upland soils. It is assumed that the sites are not harvested so Nu and BCu are set to zero and a critical BC:Al ratio -1 of 10 was used. For Clnut(N) acceptable nitrate leaching was set to 0.2 mg N L (UNECE, 2004) and -2 -1 a conservative empirical CLemp(N) for boreal forests of 35 mmolc m yr was selected (Bobbink et al., 2010).

Consistent with previous work, base cation weathering rates are very low in the sandy soils that are dominated by quartz with low amounts of weatherable minerals (Table 1). Base cation -2 -1 weathering rates are <10 mmolc m yr at all the sites and critical loads for S + N range from a

170

-2 -1 -2 -1 low of 4.9 mmolc m yr (JP107) to a high of only 25.6 mmolc m yr (JP104) and at all 10 sites the combined N + S deposition exceeds the critical load for acidity (Table 11.1). Base cation -2 -1 -2 -1 deposition, however, is between 8 mmolc m yr (JP107) and 171 mmolc m yr (JP104). Consequently, the critical load for acidity is exceeded at three sites, with the greatest exceedance -2 -1 (31.3 mmolc m yr ) occurring at JP102. At the 3 exceeded sites BC deposition exceeds S deposition and the exceedance is brought about by the additional N deposition and the assumption that only 0.2 kg N ha-1 y-1 can be immobilized in the long-term and there is negligible denitrification.

All 10 sites receive N deposition in excess of the critical load for nutrient N, whereas 4 of the 10 sites receive N deposition in excess of the empirical acid critical load for N (Table 11.1).

Table 11. 1 Base cation weathering rates, critical loads for acidity and N, S, N and BC deposition, and estimated exceedance for 10 TEEM plots. (Sites which were exceeded are indicated by negative -2 -1 numbers and are italicized): All units are mmolc m yr Site BCw CL CLnut(N) CLemp(N) Sdep Ndep BCdep ExCL(S+N) ExCLNnut ExCLNemp ID (S+N) AH8- 3.2 11.7 2.8 35 12 20 50 29.7 -17.2 15 R 5 JP205 1.5 6.1 2.8 35 10 16 22 2.1 -13.2 19 9 JP210 5.0 24.2 2.7 35 13 18 130 132.2 -15.3 17 8 JP212 4.3 18.0 2.8 35 32 60 89 15.0 -57.2 -25 5 JP213 9.5 14.0 2.9 35 11 17 12 -2.0 -14.1 18 5 JP102 5.9 12.7 2.8 35 29 49 34 -31.3 -46.2 -14 7 JP104 1.7 25.6 2.8 35 27 47 171 122.6 -44.2 -12 3 JP103 4.2 18.0 2.8 35 11 38 90 59.0 -35.2 -3 6 JP101 9.5 17.2 2.7 35 12 16 37 26.2 -13.3 19 2 JP107 2.1 4.9 2.8 35 7 15 8 -9.1 -12.2 20 1

11.6 Discussion: Implications for forest health/monitoring 1. This work suggests that the risk of soil acidification and adverse biological effects from acid deposition at the 10 TEEM sites we studied is low as long as base cation deposition remains high. In the event of a decrease in BC deposition (dust control) without a

171

concomitant decrease in S and N deposition the risk of acidification-related impacts would increase. 2. Sulphur dynamics in soils are fairly well understood, whereas there is much more uncertainty with respect to the long-term fate of N. Presently, even though data are scarce, there is no evidence of elevated N leaching from forest soils in the region (even at sites with high N deposition) (Laxton et al., 2010; Watmough et al., 2013) and increasing the immobilisation rate to 5 kg N ha-1 y-1 would result in zero sites receiving acid deposition in excess of the CL and 4 of 10 sites receiving N deposition in excess of the CL for nutrient N. These same 4 sites also receive N deposition in excess of the empirical critical load for N, which indicates that at the TEEM sites, the risk of adverse effects due to eutrophication is currently greater than potential acidification effects. 3. The estimated risk of N-related impacts on forests is highly dependent upon the choice of the empirical CL and the assumptions over N behavior in soils, which require further study.

11.7 References Arp, P. A., Oja, T., and Marsh, M. 1996. Calculating critical S and N loads and current exceedances for upland forests in southern Ontario, Canada. Canadian Journal of Forest Research.26: 696– 709.

Blett, T.F., Lynch, J.A., Pardo, L.H., Huber, C., Hauber, R., and Pouyat, R. 2014. FOCUS: A pilot study for national-scale critical loads development in the United States. Environmental Science and Policy, 38: 225-236.

Bobbink, and 16 others. 2010. Global assessment of nitrogen deposition effects on terrestrial plant diversity: a synthesis. Ecological Applications, 20: 30-59.

Bobbink, R., and J.P. Hettelingh. (eds.). 2010. Review and revision of empirical critical loads and dose-response relationships: proceedings of an expert workshop, Noordwijkerhour, 23-25 June 2010. National Institute for Public Health and the Environment, The Netherlands.

Cronan, C.S., and Grigal, D.F. 1995. Use of calcium/aluminum ratios as indicators of stress in forest ecosystems. Journal of Environmental Quality 24: 209–226.

De Vries, W., and Reinds, G.J. 1994. Assessment of critical loads and their exceedance on European forests using a one-layer steady-state model. Water Air and Soil Pollution 72: 357-394.

Environment Canada. 1998. The Canadian-wide acid rain strategy for post-2000: strategy and supporting document. Halifax, Nova Scotia.

Foster, K.R., McDonald, K., and Eastlick, K. 2001. Development and application of critical, target and monitoring loads for the management of acid deposition in Alberta, Canada. Water Air and Soil Pollution 1: 135-151. 172

Hettelingh, J.P., Posch, M., Slootweg, J., Reinds, G.J., Spranger, T., and Tarrason, L. 2007. Critical loads and dynamic modelling to assess European areas at risk of acidification and eutrophication. Water Air and Soil Pollution 7: 379–384.

Jönsson, C., Warfvinge, P., and Sverdrup, H. 1995. Uncertainty in predicting weathering rate and environmental stress factors with the profile model. Water Air and Soil Pollution. 81: 1-23.

Jung, K., Chang, S.X., Ok, Y.S., and Arshad, M.A. 2013. Critical loads and H+ budgets of forest soils affected by air pollution from oil sands mining in Alberta, Canada. Atmospheric Environment 69: 56–64.

Klaminder, J., Lucas, R.W., Futter, M.N., Bishop, K.H., Köhler, K.S.J., Egnell, G., and Laudon, H. 2011. Silicate mineral weathering rate estimates: are they precise enough to be useful when predicting the recovery of nutrient pools after harvesting? Forest Ecology and Management 261: 1–9.

Laxton, D.L., Watmough, S.A., and Aherne, J. 2011. Nitrogen cycling in Pinus banksiana and Populus tremuloides stands in the Athabasca Oil Sands region, Alberta, Canada. Water Air Soil Pollution 223: 1-13.

Li, H., and McNulty, S.G. 2007. Uncertainty analysis on simple mass balance model to calculate critical loads for soil acidity. Environmental Pollution 149: 315–326.

Lorenz, M., Nagel, H.D., Granke, O., and Kraft, P. 2008. Critical loads and their exceedances at intensive forest monitoring in Europe. Environmental Pollution 155: 426–435.

McNulty, S.G., Cohen, E.C., Moore Myers, J.A., Sullivan, T.J., and Li, H. 2007. Estimates of critical acid loads and exceedances for forest soils across the conterminous United States. Environmental Pollution 149: 281–292.

McDonnell, T.C., Cosby, B.J., Sullivan, T.J., McNulty, S.G., and Cohen, E.C. 2010. Comparison among model estimates of critical loads of acidic deposition using different sources and scales of input data. Environmental Pollution 158: 2934-2939

Nilsson, J., and Grennfelt, P. (Editors). 1988. Critical loads for sulphur and nitrogen. Miljörapport 1988:15. Nordic Council of Ministers, Copenhagen.

Ouimet, R., Arp, P.A., Watmough, S.A., Aherne, J., and DeMerchant, I. 2006. Determination and mapping critical loads of acidity and exceedances for upland forest soils in eastern Canada. Water Air and Soil Pollution 172: 57–66.

Pardo L.H., and 22 others. 2011. Effects of nitrogen deposition and empirical nitrogen critical loads for ecoregions of the United States. Ecological Applications 21: 3049-3082.

173

Rosen, J.G., Gundersen, P., Tegnhammar, L., Johansson, M., and Frogner, T. 1992. Nitrogen enrichment of Nordic forest ecosystems – The concept of critical loads. Ambio 21: 364-368.

Skeffington, R.A. 2006. Quantifying uncertainty in critical loads: (A) literature review. Water Air and Soil Pollution 169: 3–24.

Sverdrup, H., and De Vries, W. 1994. Calculating critical loads for acidity with the Simple Mass Balance Method. Water Air and Soil Pollution 72: 143–162.

Sverdrup, H., and Warfvinge, P. 1993. The effects of soil acidification on the growth of trees, grass and herbs as expressed by the (Ca+Mg+K)/Al ratio. In: Reports in Ecology and Environmental 1993: Department of Chemical Engineering II, Lund University, Lund, Sweden.

UBA. Manual on methodologies and criteria for Modelling and Mapping Critical Loads & Levels and Air Pollution Effects, Risks and Trends. Berlin: Umweltbundesamt; 2004.

UNECE. 2004. Manual on methodologies and criteria for modelling and mapping critical loads and levels and air pollution effects, risks and trends. Federal Environment Agency, Texte 52/04, Berlin, Germany, URL: http://www.icpmapping.org/Mapping_Manual

Warfvinge, P., and Sverdrup, H. 1992. Calculating critical loads of acid deposition with PROFILE - A steady-state soil chemistry model. Water Air and Soil Pollution 63: 119–143.

Watmough, S.A., Koseva, I., and Landre. 2013. A comparison of tension and zero-tension lysimeters and PRS-TM probes for measuring soil water chemistry in sandy Boreal soils in the Athabasca oil sands region, Alberta, Canada. Water, Air and Soil Pollution 224: 1663

Watmough, S.A., Whitfield, C.J., and Fenn, M.E. 2014. The importance of atmospheric base cation deposition for preventing soil acidification in the Athabasca Oil Sands region of Canada. The Science of the Total Environment 493:1-11.

Whitfield, C.J., Aherne, J., and Watmough, S.A. 2009. Modeling soil acidification in the Athabasca Oil Sands Region, Alberta, Canada. Environmental Science and Technology 43:5844-5850.

Whitfield, C.J., Aherne, J., Watmough, S.A., and McDonald, M. 2010. Estimating the sensitivity of forest soils to acid deposition in the Athabasca Oil Sands Region, Alberta. Journal of Limnology 69(Suppl. 1): 201–208.

Whitfield, C.J., Watmough, S.A., and Aherne, J. 2011. Evaluation of elemental depletion weathering rate estimation methods on acid-sensitive soils of north-eastern Alberta, Canada. Geoderma 166(1):189-197.

174

Chapter 12: State of the Jack pine forest Thomas A. Clair and Kevin E. Percy, Wood Buffalo Environmental Association, Fort McMurray, AB

12.1 Program evolution The WBEA TEEM Acid Deposition Monitoring Program (AMP) was initiated in 1998 with the first sampling was completed on 10 plots in 1998. Between 1998 and the second sampling in 2004, five plots were added and two lost to oil sands development. In 2004, 13 plots were sampled in the second cycle of the 6-year re-measurement program. The resulting report (Jones and Associates, 2007) revealed a number of issues including variability in ecological type between the 13 plot network. TEEM completed a science review based on the report findings that lead to a series of changes in network design. Among them were adoption of the forest health concept, stratification of plots across regional deposition zones, expansion of the spatial coverage employing the ecological analogue concept, and installation of key plot infrastructure to aid in determination of cause-effect should there be changes in biological or chemical indicators (Percy et al., 2012).

Along with needed changes in design to reduce intra- and inter-site variability and increase the capacity to detect change, indicator workshops were held to validate existing indicators and consider new ones. Important changes were made, including the decision to not composite soil samples in the future. Between 2008-2011, the forest health network (FHN) expanded to 25 plots following three years of field and aerial work to locate and ground-truth through analysis additional type 3 analogue plots (Jaques and Legge, 2012). Unfortunately, in May-June 2010, six of the plots were completely or partially burned in the 700,000 ha Richardson Fire.

In 2011, the third cycle of sampling was completed on 21 of the 25 plots that were not completely burned in 2010. In all, eight plots were common across the 1998, 2004 and 2011 samplings, five of which were not burned in 2010. This report provides soil and foliar data from these sites to evaluate change over time.

Maynard (Chap. 8) was able to compare Jack pine foliar chemistry of five sites sampled in 1998, 2004 and 2011 and found slight increases in total and extractable sulphur at 3 of the sites (Table 8.3), as well as N increases in all five sites with most of the changes occurring between 2004 and 2011 (Table 8.4). When he compared a further four sites sampled in 2004 and 2011 (Table 8.5), he found the same pattern with half the sites showing S increases, and all of the sites showing N increases. Atmospheric deposition is changing the chemistry of the region’s soils.

A comparison of soil microbiota from sites visited in 2004 and 2011 (Chap. 7) also suggested that communities had changed. The same statistical approaches done on 2011 data were also applied to the 2004 collected by S. Visser. Where the more recent data showed structural patterns in community composition, this was not detected using the 2004 values. The lack of change could

175

be due to the fact that: 1) the effect of deposition on soil microbial communities was not strong enough to be detected in 2004; 2) the indicators used in 2004 (total microbial biomass using chloroform fumigation, soil respiration, ectomycorrhizal root tip characterization, soil faunal abundance and diversity) were not sensitive enough to detect differences in communities; 3) that there were fewer sampling sites (13) in 2004 and most were close to the oil sands operations which did not provide the variability needed to statistically detect change; or 4) that variation (i.e. depth to water table, moss cover)from type 3 analogue in several older sites may have introduced a confounding influence. Nevertheless, there were differences in microbial communities detected in 2011 that were not present in 2004.

Puckett (Chap. 10) provided a review of the historical literature from the AOSR which showed that tree lichens were showing the effects of pollution stress, particularly nearer to the operations. The review suggested that historical changes in air quality have affected lichen species composition, but that because of a consistent approach to producing baseline conditions, it is quite difficult to quantify changes in this plant community in comparing the 2011/12 data to the past work.

The newest suite of sampling sites and data have allowed TEEM-sponsored scientists to provide state-of-the-art analyses of current conditions, providing excellent analyses of the spatial extent of current day air quality and deposition parameters (Chapters 2, 3, 4, this report), as well as vegetation distribution and chemistry patterns (Chapters 8, 9 and 10). The current sampling regime has been assessed and has been found suitable to allow temporal trends to be assessed with some of the 2004 results (Chapters 7 and 8), and also is thought to provide a robust base for future studies looking at trends in ecosystem variables in the AOSR.

12.2 Air chemistry and atmospheric deposition implications The three studies which estimated air quality and deposition of atmospheric contaminants (Hsu and Bytnerowicz, passive air quality samplers, Chap. 2; Fenn, wet and dry deposition captured on resins in the open and in forest canopy throughfall, Chap. 3; and Davies et al., using an air modeling approach, Chap. 4) all showed the same patterns of concentration and deposition distribution for S and N. The general pattern of air quality and deposition unsurprisingly follows air flow with two main patterns visible. The first is the general west to east air flow pattern which causes S and N concentrations and deposition to be higher east of the mining/upgrading operations than to the west. There is also a north-south spatial pattern in concentrations, showing the effect of the Athabasca River valley on wind and thus dispersion. The other significant finding from these studies is that the operations effects on N, and S air concentrations and deposition return to background levels ~50 km of mines and upgraders.

A significant difference between atmospheric deposition of S and N, as well as metals and other contaminants in the AOSR with the situation in eastern North America, is that over 70% of deposition is in the form of gases, and fine particulates (commonly referred to as dry deposition), as opposed to rain and snow (wet deposition). This difference is mainly due to the generally low

176

precipitation in this region that makes the accurate estimation of deposition amounts more difficult than in wetter regions. Because of this difficulty, estimations of deposition, which is important in the determination of critical loads and biological soil processes, is difficult and dependent on complex atmospheric micro-models which don’t always mirror reality.

As an example, total dissolved inorganic nitrogen (DIN) air concentrations were measured using passive samplers (Fig. 2.9) and were compared to results from the modeling approach (Fig. 4.1). The comparison of these figures suggests that the passive approach produces air concentrations 25-30% lower than the mass balance model. However, there is also a significant discrepancy between the reported dominance of NOx emissions and deposition data as the IER results (Chapter 3) show that NH4-N deposition is double that of NO3-N apparently due to unreported elevated emissions of particulate NH4 and likely also because of underestimates in reported NH3 emissions. Emissions of particulate matter mass is reported in the emissions inventory (NPRI, 2010, 2011), but chemical characterization of particulate matter emissions is not.

Total dissolved inorganic N deposition (as opposed to air concentration) was estimated using the three approaches described above. Both the modeling (Fig. 4.2) and ion exchange resin (IER) approaches produce deposition estimates similar to each other, while ionic N estimates from the passive sampler approach (Figs. 2.18 for NH3, 2.19 for NO3) are approximately 1/3 of the other two methods. Both the N concentration and deposition estimates suggest that the passive approach probably underestimates N in the region, though this discrepancy might be resolved by modifying the models used in the conversion of passive collector resin concentrations to air concentration estimates. The other crucial factor in this comparison is that the passive samplers only provide estimates of dry deposition and not total, so only account for ~70% of deposition. Nevertheless, all three methods provided the same patterns of air chemistry and deposition in the region, and suggest that oil exploitation activities no longer impact air quality past 50 km from the sources.

An important result from the N passive sampling work shows that the southern portion of the AOSR receives high NH3 concentrations during the summer (Fig. 2.2) which originate in the agricultural or industrial regions of central Alberta. These concentrations are greater than any measured in the vicinity of oil sands operations.

Comparison of estimated S dry deposition (Fig. 2.11) and modelled air concentrations (Fig 4.1) on the other hand, though shown for different time periods, that total deposition seem more similar to each other than the DIN situation. The high SO2 distribution is localized and limited to a small area close to the major emission during summer months. SO2 concentrations in winter were generally higher than those in summer. SO2 spatial variations in winter 2010 and winter 2011 followed a similar pattern with the highest concentrations observed near the major emission sources. The lower mixing height and stable atmosphere, wind speed and wind direction are the major factors for reducing the SO2 dilution effect in winter months, leading to the locally greater concentrations (Chap. 2).

177

When S deposition estimates from the passive, IER and modeling approaches are compared however, the passive sampler approach (Fig. 2.20) estimate produces values approximately 1/3 of the model (Fig. 4.2) and considerably lower than the IER (Fig. 3.4, 3.21). If we assume that the IER and modeled values are more accurate, these results suggest that the use of passive air sampling for the study of air pollutant impacts on ecosystems needs to be assessed or modified to improve the quality of its estimates. While values of total wet and dry deposition calculated with the modeled CALPUFF are usually higher at most stations, the CALPUFF results were surprisingly lower than the passive results at a few sites (BM10, BM7, and SM7) so that there is a need for explaining those unexpected results.

The IER and modeling approaches also estimated base cation deposition which generally followed the same spatial patterns as S and N. When the base cation (BC) deposition was compared to the total acid anion values, the potential acidification index (PAI) deposition patterns indicated that acid forming emission contributions are completely neutralized by BC emissions, especially near oil sands mining operations (Chap. 4). This suggests that overall, there is currently no regional acidification effect of oil sands operations in the AOSR, but that any changes in BC, N, or S emissions may change this balance depending on what inputs are modified.

12.3 Soil and plant interactions with atmospheric deposition and critical loads The main question asked when assessing the potential impact of oil sand extraction and processing on the regional forest ecosystem is whether or not soil and plant communities are showing changes which could be related to changes in air quality/deposition. The study of soil chemistry and biology both show that soils were affected by atmospheric deposition. Sulphur in soils was correlated with modeled S+N deposition at the LFH, 0-5 and 5-15 and 15-30 cm depths at the plots though neither nitrogen nor pH showed any correlation with deposition. The reason for S being correlated with deposition is due to its conservative nature. S is not taken up in any significant amounts by plants and it is not volatile, and thus will remain where it is deposited, unless washed out by rain or snowmelt.

Deposited N is a nutrient which is in scarce supply in Boreal forests and thus disappears into plant material or under anaerobic conditions is denitrified, so is much more mobile than S. Results of the EIR study (Chapter 3) also support the hypothesis that eutrophication effects to sensitive organisms such as epiphytic lichens may be of greater concern than acidification because acidic deposition is matched by equivalent amounts of buffering base cation deposition (Watmough et al., 2014).

Soil microflora fungi, Gram-negative bacteria and bacterial:fungal ratio are significantly affected by the distance to the industrial centre as fungal abundance was greatest in sites that were furthest from the industrial center (Chap. 7). Total nitrogen, carbon in soil and PAI (i.e. base cations) were significant indicators correlating with differences in microbial communities, though as could be expected, forest fire also had a major effect.

178

The vegetation indicators were strongly related to the deposition. Total richness, vascular cover, forb cover and shrub richness were significantly related to many more deposition variables than were the other vegetation response variables (Table 9.1). However, the most significant driver modifying plant communities was base cation deposition, and not eutrophication or acidification. To confirm this, previous studies elsewhere showed that atmospheric nitrogen deposition was associated with increased abundance of grasses and forbs but decreases abundance of dwarf shrubs while not affecting species richness (Nordin et al., 2005; Bobbink et al., 2010). In contrast, no such increases in grasses or forbs were found except for the trailing shrub bearberry. There was however, increased total and vascular plant species richness in the plots. This may have been due to the fact that not all plots were pure ecological analogues.

Though the understory plant community was affected by base cation deposition, metals and major ions were correlated with distance from mining sites to the lichen community. Element concentrations were regressed against distance with the point of origin being the mid-point between the Suncor and Syncrude operations. The elements that showed the highest power law fit [R2 > 0.65] for the relationship between concentrations and distance are aluminium, arsenic, chromium, cobalt, iron, molybdenum, nickel, silicon, sodium, titanium and vanadium (Figure 10.2) suggesting that source proximity is the major determinant of the lichen element composition. Concentrations of most of these metals above background levels seemed to persist up to 80 km past their points of origin, extending the range of influence of operations further than the 50 km to background for most of the base cations (Fig. 10.3). The Puckett study (Chap. 10) however, was not able to identify patterns in lichen community structure which could be linked to distance from the mining/upgrading activities.

There were weak but significant correlations between foliar N and S concentrations and modeled N and S deposition amounts (Fig 8.1a, b), with sites closer to the industrial activity having higher N and S concentrations. The best relationships between deposition and needle concentrations (R2=0.43; Figure 8.3) were between current foliar Ca concentrations and CALPUFF predicted base cation throughfall deposition.

The vascular plant community, as well as the soil and plant chemistry results were confirmed by the critical load analysis done in Chapter 11. This work suggests that the risk of soil acidification and adverse biological effects from acid deposition at the 10 TEEM sites studied is low as long as base cation deposition remains high. In the event of a decrease in BC deposition (dust control) without a concomitant decrease in S and N deposition the risk of acidification-related impacts would increase. The estimated risk of N-related impacts on forests however, is dependent on the choice of the empirical critical load and the assumptions concerning N behavior in soils, which require further study.

The main message coming from the soil and plant work is that base cation deposition is very significant in determining indicator status, and that nitrogen may also be increasing in importance as a eutrophying element. The lichen work also suggests that metals are being taken up by a portion of the plant community. Sulphur, which was the initial pollutant of concern at

179

the onset of the industrial developments has been shown to be of lesser concern. The main reasons for that is that significant S emission controls have been effective in reducing the amounts going to the atmosphere, and that the base cations which come from the operations and are deposited in the landscape are neutralizing sulfuric acids as well as nitric acids. Any changes in emissions of base cations, S and N will change this acid-base balance occurring and could change the situation as it now exists.

The trigger for a management response to acid deposition is based on changes to %Base Saturation with time. Our results show that %Base Saturation threshold level was not reached for any site or any depth. Similar to %Base Saturation there was no correlation of Base Cation: Al ratio with modeled N and S deposition in the LFH (Figure 8.15a or any mineral soil layer (Figure 8.15b; only 0-5 cm mineral layer shown).

12.4 Future directions As oil sands development increases into the future, it is likely that more and more attention will be placed on its environmental management in the AOSR. Because of this, it will be important to maintain if not enhance, the FHN to ensure that an essential long-term data set is gathered to examine trends for cumulative effects management. Results from the studies described here show that there are changes in soil chemical composition with distance from the operations. The work also shows that acidification does not pose a great risk.

In order to better understand the effects of future oil sands activities on the AOSR Boreal forest, it will be important to enhance and optimize monitoring activities. To that effect, the WBEA TEEM program has deployed 23 early warning edge forest plots in the region, at the interface between the forest and clearings or wetlands. Forest edges are more exposed to air pollutants than trees within a stand some distance from the edge, and thus indicators will change more rapidly.

The FHN has been mostly focused in the region north of Fort McMurray, while much new development has been occurring south of the city. The approach taken by the FHN should be applied in the southern region to ensure continuity in monitoring approaches for the whole oil sands region. Another initiative recommended is harmonization of design and approaches between WBEA and The Lakeland Industry Community Association (LICA) to the south, both located within the Lower Athabasca Regional Plan (LARP) which sets triggers for cumulative effects management on a land use framework scale.

Oil sands development is a large-scale endeavor which receives a great deal of attention from media, citizens, industry and environmental groups. Because of this, it is important that the work done to monitor the AOSR ecosystem be scientifically valid and accepted by regulators and the scientific community. The continuing evolution and long-term stability of the FHN is an important step in ensuring that the public and scientific community receives information that they can trust, and can use used for informed decision making.

180

APPENDICES

181

Chapter 1 Appendices Appendix 1.1 Location of current continuous air monitoring sites used for this study ID Name UTM Z12 UTM Z12 Latitude Longitude Easting Northing AMS 1 Fort McKay 461292 6338657 57.189450 -111.640450 AMS 2 Mildred Lake 465795 6323069 57.049783 -111.563817 AMS 3 Lower Camp 469264 6321073 57.032100 -111.506400 Met Tower AMS 4 Buffalo 463980 6317153 56.996500 -111.592883 Viewpoint AMS 5 Mannix 470697 6313920 56.967933 -111.481967 AMS 6 Patricia 470869 6289809 56.751350 -111.476367 McInnes AMS 7 Athabasca 476127 6287698 56.732683 -111.390200 Valley AMS 8 Fort 489780 6507596 58.708400 -111.176400 Chipewyan AMS 9 Barge 463769 6339612 57.198233 -111.599617 Landing AMS 10 Albian Mine 468304 6348789 57.281000 -111.525733 Site AMS 11 Lower Camp 469597 6320483 57.026817 -111.500833 AMS 12 Millenium 475641 6314036 56.969267 -111.400650 Mine AMS 13 Syncrude UE- 461119 6334165 57.149083 -111.642617 1 AMS 14 Anzac 497706 6256086 56.449283 -111.037217 AMS 15 CNRL 455437 6351437 57.303717 -111.739617 Horizon AMS 16 Albian 469313 6345230 57.249100 -111.508567 Muskeg River AMS 17 Wapasu 505865 6343925 57.238371 -110.902831 AMS 18 Conklin 489115 6163945 55.621290 -111.172843

182

Appendix 1.2 Location of Forest Health plots and meteorological measurement equipment Site UTM Z12 UTM Z12 FHS Passive Met_Tower Latitude Longitude Easting Northing JP101 421558 6266856 1 1 0 56.539503 -112.275600 JP102 467073 6307455 1 1 0 56.909616 -111.540720 JP103 465614 6365119 1 0 0 57.427494 -111.572630 JP104 474239 6330715 1 1 1 57.119019 -111.425425 JP106 489928 6391161 1 0 0 57.662610 -111.168812 JP107 474195 6416576 1 1 1 57.890264 -111.435236 JP108 565696 6285080 1 1 0 56.705154 -109.926993 JP109 464415 6377615 1 0 0 57.539645 -111.594417 JP201 333992 6324286 1 0 1 57.032109 -113.735517 JP205 532716 6410782 1 1 0 57.837770 -110.449002 JP210 534138 6236579 1 1 0 56.272802 -110.448742 JP212 475180 6323437 1 1 0 57.053691 -111.409165 JP213 575855 6323351 1 1 1 57.047348 -109.749686 JP303 493348 6298912 1 0 0 56.833990 -111.109017 JP304 477251 6323373 1 0 0 57.053223 -111.375019 JP307 498609 6357289 1 0 0 57.358456 -111.023121 JP308 387669 6328374 1 0 0 57.085007 -112.853495 JP310 451300 6283450 1 0 0 56.692592 -111.795140 JP311 441755 6269315 1 0 1 56.564529 -111.947774 JP312 534615 6298597 1 0 0 56.829919 -110.432767 JP313 557254 6280086 1 0 0 56.661405 -110.065964 JP315 435571 6329789 1 0 0 57.106911 -112.063674 JP316 554474 6245747 1 1 1 56.353265 -110.118498 JP318 539544 6320996 1 0 0 57.030737 -110.348504 JP317 503328 6325897 1 0 0 57.076447 -110.945104 AH3 492511 6283600 0 1 0 56.696417 -111.122283 AH7 453144 6298705 0 1 0 56.829817 -111.767833 AH8-R 434851 6329241 0 1 0 57.101883 -112.075417 BM7 424381 6435907 0 1 0 58.058200 -112.281433 BM10 415871 6353877 0 1 0 57.320050 -112.396967 BM11 445778 6394708 0 1 0 57.691317 -111.909517 NE7 508089 6333722 0 1 0 57.146683 -110.866317 NE10 549559 6274101 0 1 0 56.608517 -110.192633 NE11 486916 6349453 0 1 0 57.287883 -111.217067 R2 474018 6330233 0 1 0 57.114677 -111.429024 SM7 448736 6171395 0 1 0 55.685650 -111.815367 SM8 489125 6228533 0 1 0 56.201617 -111.175283 WF4 440461 6334275 0 1 0 57.147867 -111.984033 183

Chapter 2 Appendices Appendix 2.1 Annual NH3 N Dry Deposition (kg ha-1 a-1) at the forest monitoring sites site 2006 2007 2008 2009 2010 2011 2012 AH3 1.26 0.54 0.80 0.48 0.67 0.76 2.09 AH7 0.65 0.64 0.64 0.49 0.59 AH8 0.36 0.67 0.48 0.48 0.72 1.45 AMS1 1.21 1.75 1.00 0.69 1.00 1.06 1.44 AMS6 1.91 1.53 2.08 1.41 1.42 1.32 1.96 AMS14 1.11 1.26 0.73 3.12 BM10 0.53 0.39 1.05 1.58 BM11 0.39 0.71 1.03 1.44 BM7 0.66 1.05 0.61 1.30 JP101 0.81 0.57 0.59 0.40 0.62 0.74 1.29 JP102 0.98 0.61 0.74 0.60 0.77 1.11 1.42 JP104 1.08 0.51 0.75 0.53 0.47 0.66 1.70 JP107 0.78 0.54 0.58 0.75 0.54 0.90 JP205 0.68 0.42 0.47 0.52 0.56 0.38 0.81 JP210 0.84 0.53 0.67 0.42 0.64 0.74 1.34 JP212 0.58 0.53 0.78 1.47 1.32 JP213 1.30 0.53 0.81 0.78 0.75 1.09 NE10 0.48 0.65 1.05 1.84 NE11 0.55 0.52 1.26 NE7 0.80 0.71 0.63 1.55 SM7 0.65 0.61 2.35 2.09 SM8 0.57 0.47 0.60 1.39 WF4 0.41 0.43 1.56 1.74

184

Appendix 2.2 Annual HNO3 as N Dry Deposition (kg ha-1 a-1) at the forest monitoring sites. site 2006 2007 2008 2009 2010 2011 2012 AH3 0.96 0.30 0.34 0.38 0.37 1.81 1.09 AH7 1.20 0.29 0.29 0.37 0.27 AH8 0.16 0.20 0.57 0.30 0.57 0.95 AMS1 1.30 0.37 0.55 0.70 0.36 1.30 1.27 AMS 6 1.17 0.49 0.71 0.71 0.46 1.57 1.61 AMS14 0.32 0.55 1.92 1.58 BM10 0.29 0.24 1.43 1.52 BM11 0.30 0.23 1.24 2.10 BM7 0.27 0.27 1.25 2.70 JP101 1.15 0.25 0.40 0.41 0.28 0.77 1.12 JP102 2.09 0.36 0.40 0.78 0.50 0.95 0.80 JP104 1.35 0.43 0.41 0.48 0.35 1.06 0.55 JP107 0.91 0.29 0.32 0.62 0.42 0.97 JP205 0.98 0.36 0.25 0.33 0.41 1.48 1.20 JP210 1.32 0.34 0.31 0.68 0.29 1.33 0.87 JP212 0.25 0.52 0.25 2.04 0.54 JP213 1.68 0.30 0.44 1.10 0.37 1.21 NE10 0.18 0.23 1.04 0.76 NE11 0.16 0.53 0.80 NE7 0.39 0.24 0.94 2.62 SM7 0.29 0.52 1.87 2.51 SM8 0.32 0.31 1.59 2.44 WF4 0.40 0.30 1.14 2.32

185

Appendix 2.3 NO2 as N Dry Deposition (kg ha-1 a-1) site 2009 2010 2011 2012 AH3 0.30 0.18 0.17 0.17 AH7 0.38 0.23 AH8 0.26 0.18 0.12 0.15 AMS1 1.06 0.89 0.64 0.70 AMS 6 1.24 0.88 0.70 0.78 AMS14 0.33 0.36 0.22 0.28 BM10 0.07 0.07 0.07 0.05 BM11 0.11 0.06 0.03 0.02 BM7 0.06 0.02 0.01 0.01 JP101 0.21 0.12 0.11 0.12 JP102 0.70 0.45 0.49 0.47 JP104 1.17 0.86 0.73 0.96 JP107 0.40 0.29 0.21 JP205 0.17 0.09 0.08 0.05 JP210 0.18 0.09 0.06 0.09 JP212 0.83 0.69 0.40 0.56 JP213 0.11 0.06 0.05 NE10 0.20 0.11 0.13 0.15 NE11 0.16 0.05 0.05 0.03 NE7 0.52 0.32 0.27 SM7 0.15 0.07 0.06 0.05 SM8 0.17 0.09 0.07 0.06 WF4 0.23 0.14 0.10 0.12

186

Appendix 2.4 Total N Dry Deposition (kg ha-1 a-1) Average Total N (wet and dry) 2006 2007 2008 2009 2010 2011 2012 CALPUFF AH3 2.42 1.13 1.41 1.16 1.21 2.74 3.36 1.92 4.07 AH7 2.12 1.26 1.16 1.24 1.08 1.37 5.60 AH8 0.78 1.09 1.30 0.96 1.40 2.54 1.35 2.78 AMS1 3.55 3.07 2.44 2.45 2.25 3.00 3.41 2.88 AMS6 4.15 3.23 3.75 3.36 2.77 3.59 4.35 3.60 5 AMS14 1.76 2.17 2.87 4.97 2.94 BM10 0.98 0.70 2.54 3.15 1.84 1.65 BM11 0.80 0.99 2.30 3.56 1.91 2.15 BM7 1.00 1.34 1.87 4.00 2.05 1.93 JP101 2.10 1.06 1.14 1.03 1.03 1.63 2.53 1.50 2.25 JP102 3.61 1.73 1.70 2.08 1.72 2.55 2.69 2.30 6.83 JP104 3.11 1.95 1.97 2.19 1.68 2.44 3.20 2.36 6.63 JP107 2.04 1.13 1.22 1.76 1.25 2.08 1.58 2.04 JP205 1.73 0.92 0.81 1.02 1.06 1.94 2.07 1.36 2.17 JP210 2.24 1.01 1.08 1.27 1.03 2.12 2.30 1.58 2.46 JP212 1.54 1.88 1.73 3.90 2.42 2.29 8.44 JP213 3.02 0.91 1.31 1.99 1.18 2.35 1.79 2.43 NE10 0.82 0.93 2.14 2.62 1.63 2.66 NE11 1.23 1.37 2.33 1.64 3.35 NE7 1.38 1.06 1.70 4.32 2.12 3.05 SM7 1.09 1.21 4.29 4.65 2.81 2.63 SM8 1.06 0.87 2.26 3.88 2.02 3.79 WF4 1.04 0.86 2.80 4.18 2.22 3.70

187

Chapter 4 Appendices Appendix 4.1 Predicted annual average SO2 and NO2 concentrations at the forest health assessment sites.

Forest Annual Average SO2 Annual average NO2 health Concentrations (µg/m3) concentrations (µg/m3) Site Minimum Average Maximum Minimum Average Maximum JP101 0.91 1.07 1.34 1.70 1.89 2.37 JP102 3.87 4.35 5.54 10.84 13.01 15.45 JP103 2.00 2.25 2.46 11.03 11.69 12.46 JP104 3.30 3.46 3.75 12.72 13.68 14.68 JP106 0.80 1.01 1.26 2.99 3.33 3.66 JP107 0.57 0.71 0.81 1.78 1.94 2.09 JP108 0.78 0.86 0.95 0.97 1.00 1.02 JP109 0.98 1.24 1.36 5.48 5.75 5.92 JP201 0.37 0.43 0.52 0.72 0.81 0.91 JP205 0.74 0.83 0.96 1.16 1.28 1.44 JP210 1.34 1.46 1.56 1.66 1.73 1.89 JP212 3.85 4.00 4.50 15.46 16.57 18.01 JP213 0.96 1.13 1.25 1.05 1.11 1.16 JP303 1.68 2.00 2.30 4.78 5.05 5.37 JP304 3.17 3.45 3.98 12.83 14.07 15.83 JP307 1.29 1.38 1.50 4.05 4.30 4.63 JP308 0.52 0.60 0.68 1.86 2.00 2.21 JP310 1.88 2.16 2.57 3.79 4.48 5.57 JP311 1.24 1.44 1.73 2.26 2.66 3.33 JP312 1.36 1.47 1.57 1.64 1.73 1.80 JP313 1.18 1.25 1.32 1.33 1.39 1.44 JP315 0.78 0.90 1.07 3.01 3.35 3.79 JP316 1.18 1.26 1.35 1.42 1.48 1.55 JP318 1.33 1.47 1.67 1.51 1.62 1.69 JP317 2.34 2.54 2.82 3.67 4.01 4.26 AH3 1.65 1.85 2.04 4.64 4.87 5.13 AH7 2.55 3.02 3.74 6.96 8.02 10.07 AH8-R 0.77 0.90 1.06 2.99 3.34 3.79 BM7 0.53 0.63 0.82 0.54 0.62 0.78 BM10 0.68 0.76 0.89 1.44 1.65 1.97 BM11 1.09 1.28 1.55 1.69 2.00 2.27 NE7 1.93 2.36 2.65 3.23 3.47 3.73 NE10 1.26 1.30 1.36 1.38 1.47 1.54 NE11 1.57 1.71 1.87 6.62 7.00 7.44

188

R2 3.30 3.46 3.75 12.72 13.68 14.68 SM7 0.72 0.76 0.79 1.45 1.48 1.51 SM8 1.54 1.76 1.93 2.43 2.58 2.77 WF4 0.98 1.09 1.20 4.34 4.80 5.51 Notes: The predictions include non-LAR and LAR contributions. The non-LAR contributions are based on CMAQ predictions and represent one simulation year. The LAR contributions are based on CALPUFF predictions and represent five simulation years. The CALPUFF simulation was based on five years of meteorological data (2002 to 2006). The minimum, average, and maximum values provide one indicator of the uncertainty associated with the meteorological data.

189

Appendix 4.2 Predicted annual sulphur and nitrogen compound deposition at the forest health assessment sites. Sulphur compound deposition Nitrogen compound deposition (kg N/ha/a) Forest (kg S/ha/a) health Site Non-LAR LAR Total S Non-LAR LAR Reduced N Total N JP101 1.09 0.88 1.97 1.05 0.53 0.66 2.25 JP102 0.96 3.69 4.65 0.86 3.84 2.13 6.83 JP103 0.96 0.88 1.84 0.93 3.40 0.96 5.28 JP104 1.01 3.24 4.25 0.96 4.02 1.64 6.63 JP106 0.87 0.69 1.56 0.80 0.90 1.16 2.87 JP107 0.78 0.40 1.18 0.77 0.47 0.80 2.04 JP108 1.10 0.67 1.78 1.02 0.33 1.01 2.36 JP109 0.89 0.61 1.50 0.80 1.35 0.84 2.99 JP201 1.14 0.33 1.47 1.13 0.17 0.59 1.89 JP205 1.01 0.52 1.53 0.92 0.36 0.89 2.17 JP210 1.15 0.92 2.07 1.10 0.50 0.86 2.46 JP212 1.01 4.12 5.13 0.95 5.08 2.41 8.44 JP213 1.17 0.61 1.78 1.20 0.31 0.92 2.43 JP303 1.06 1.86 2.92 1.00 1.51 1.84 4.34 JP304 1.01 3.65 4.67 0.95 4.24 2.63 7.82 JP307 0.95 0.89 1.84 0.95 0.90 0.82 2.67 JP308 0.94 0.53 1.47 1.19 0.28 0.40 1.88 JP310 1.12 1.72 2.84 1.08 1.24 1.37 3.69 JP311 1.17 1.17 2.34 1.12 0.76 0.67 2.55 JP312 1.07 1.10 2.17 0.98 0.56 1.29 2.83 JP313 1.23 0.82 2.05 1.13 0.42 1.08 2.62 JP315 0.92 0.93 1.84 0.89 0.84 1.11 2.84 JP316 1.18 0.69 1.87 1.12 0.39 0.87 2.38 JP318 0.98 1.04 2.02 0.92 0.55 1.05 2.52 JP317 1.06 2.00 3.06 1.03 1.21 1.15 3.40 AH3 1.01 1.47 2.47 0.90 1.25 1.92 4.07 AH7 1.08 2.09 3.16 1.06 2.09 2.45 5.60 AH8-R 0.92 0.93 1.85 0.89 0.83 1.05 2.78 BM7 1.23 0.28 1.52 1.32 0.15 0.46 1.93 BM10 0.98 0.52 1.51 0.95 0.38 0.31 1.65 BM11 0.98 0.58 1.56 1.10 0.45 0.59 2.15 NE7 1.07 1.65 2.72 1.04 1.05 0.96 3.05 NE10 1.23 0.79 2.02 1.08 0.42 1.16 2.66 NE11 0.97 1.16 2.13 0.97 1.55 0.84 3.35 R2 1.01 3.24 4.25 0.96 4.02 1.82 6.80 SM7 1.29 0.44 1.73 1.38 0.36 0.89 2.63 SM8 1.33 1.04 2.36 1.35 0.70 1.73 3.79 190

WF4 0.87 1.09 1.96 0.86 1.18 1.66 3.70 Notes: The non-LAR contributions are based on CMAQ predictions and represent one simulation year. The LAR contributions are based on CALPUFF predictions and represent the average for the five simulation years. The Reduced N contributions are based on measurements. -2 Sulphur compounds = SO2 + SO4 -1 Nitrogen compounds = NO + NO2 + HNO3 + NO3 + N2O5 (CAMQ only) + Reduced N (measured)

191

Appendix 4.3 Predicted potential acid input (keq/ha/a) at the forest health assessment sites. Sulphur compound Nitrogen compound deposition Throughfall Bulk Assumption Forest deposition Assumption health Non- Reduced Total N BC PAI BC deposition PAI Site LAR Total S Non-LAR LAR LAR N deposition JP101 0.07 0.05 0.12 0.08 0.04 0.05 0.16 0.75 -0.47 0.37 -0.09 JP102 0.06 0.23 0.29 0.06 0.27 0.15 0.49 0.93 -0.15 0.34 0.44 JP103 0.06 0.05 0.11 0.07 0.24 0.07 0.38 1.16 -0.67 0.90 -0.41 JP104 0.06 0.20 0.27 0.07 0.29 0.12 0.47 2.28 -1.55 1.71 -0.97 JP106 0.05 0.04 0.10 0.06 0.06 0.08 0.20 0.96 -0.66 0.34 -0.04 JP107 0.05 0.02 0.07 0.05 0.03 0.06 0.15 0.36 -0.14 0.08 0.14 JP108 0.07 0.04 0.11 0.07 0.02 0.07 0.17 0.50 -0.22 0.65 -0.37 JP109 0.06 0.04 0.09 0.06 0.10 0.06 0.21 0.85 -0.54 0.56 -0.25 JP201 0.07 0.02 0.09 0.08 0.01 0.04 0.14 0.60 -0.37 0.25 -0.03 JP205 0.06 0.03 0.10 0.07 0.03 0.06 0.16 0.61 -0.36 0.22 0.03 JP210 0.07 0.06 0.13 0.08 0.04 0.06 0.18 0.45 -0.14 1.30 -1.00 JP212 0.06 0.26 0.32 0.07 0.36 0.17 0.60 1.61 -0.69 0.89 0.03 JP213 0.07 0.04 0.11 0.09 0.02 0.07 0.17 0.25 0.03 0.12 0.17 JP303 0.07 0.12 0.18 0.07 0.11 0.13 0.31 1.92 -1.43 1.09 -0.59 JP304 0.06 0.23 0.29 0.07 0.30 0.19 0.56 1.78 -0.93 1.08 -0.23 JP307 0.06 0.06 0.11 0.07 0.06 0.06 0.19 1.68 -1.38 1.31 -1.00 JP308 0.06 0.03 0.09 0.09 0.02 0.03 0.13 0.58 -0.35 0.23 0.00 JP310 0.07 0.11 0.18 0.08 0.09 0.10 0.26 0.71 -0.27 0.28 0.16 JP311 0.07 0.07 0.15 0.08 0.05 0.05 0.18 0.73 -0.40 0.30 0.03 JP312 0.07 0.07 0.14 0.07 0.04 0.09 0.20 1.09 -0.75 0.81 -0.47 JP313 0.08 0.05 0.13 0.08 0.03 0.08 0.19 0.60 -0.29 0.74 -0.43 JP315 0.06 0.06 0.12 0.06 0.06 0.08 0.20 1.07 -0.75 0.50 -0.18 JP316 0.07 0.04 0.12 0.08 0.03 0.06 0.17 0.43 -0.14 1.09 -0.80 JP318 0.06 0.07 0.13 0.07 0.04 0.08 0.18 1.05 -0.75 0.74 -0.43 192

JP317 0.07 0.12 0.19 0.07 0.09 0.08 0.24 1.92 -1.49 1.40 -0.96 AH3 0.06 0.09 0.15 0.06 0.09 0.14 0.29 1.44 -1.00 0.66 -0.22 AH7 0.07 0.13 0.20 0.08 0.15 0.18 0.40 0.81 -0.21 0.39 0.20 AH8-R 0.06 0.06 0.12 0.06 0.06 0.08 0.20 1.05 -0.74 0.50 -0.19 BM7 0.08 0.02 0.09 0.09 0.01 0.03 0.14 0.25 -0.02 0.08 0.15 BM10 0.06 0.03 0.09 0.07 0.03 0.02 0.12 0.51 -0.29 0.14 0.07 BM11 0.06 0.04 0.10 0.08 0.03 0.04 0.15 0.40 -0.15 0.14 0.11 NE7 0.07 0.10 0.17 0.07 0.08 0.07 0.22 1.78 -1.39 1.34 -0.95 NE10 0.08 0.05 0.13 0.08 0.03 0.08 0.19 0.67 -0.36 0.83 -0.51 NE11 0.06 0.07 0.13 0.07 0.11 0.06 0.24 2.08 -1.71 1.76 -1.39 R2 0.06 0.20 0.27 0.07 0.29 0.13 0.49 1.84 -1.09 1.19 -0.44 SM7 0.08 0.03 0.11 0.10 0.03 0.06 0.19 0.55 -0.26 0.61 -0.31 SM8 0.08 0.06 0.15 0.10 0.05 0.12 0.27 0.78 -0.36 0.52 -0.10 WF4 0.05 0.07 0.12 0.06 0.08 0.12 0.26 1.15 -0.77 0.46 -0.07 Note: The non-LAR contributions are based on CMAQ predictions and represent one simulation year. The LAR contributions are based on CALPUFF predictions and represent the average for the five simulation years. The Reduced N contributions are based on measurements. -2 Sulphur compounds S= SO2 + SO4 Nitrogen compounds N = NO + NO2 + HNO3 + NO3 + N2O5 (CAMQ only) + Reduced N (measured) The BC contributions are based on measurements BC = Ca+2 + Mg+2 + Na+ PAI = S + N – BC Shaded cells indicate positive PAI values.

193

Chapter 5 Appendices Appendix 5.1 Jack Pine Morphometric Measurements – Routine Monitoring Program (Plot and Off-Plot Trees) Measurement 1998 2001 2004 2011/2012  Height and DBH measured for all  Height and DBH  Height and DBH measured for all  Height and DBH measured trees of DBH>10cm (including measured for all trees of trees of DBH>10cm (including for all trees of DBH>10cm dead trees) in the vegetation plot DBH>10cm (including dead trees) in the vegetation plot (including dead trees) in the Height &  Height and DBH measured for dead trees) in the  Height and DBH measured on the vegetation plot Diameter the 10 numbered off-plot trees vegetation plot 4 largest-diameter, undamaged,  Height and DBH measured  Height and DBH dominant/co-dominant trees per for the 10 numbered off-plot measured for the 10 plot (vegetation plot, 4 soils plots), trees numbered off-plot trees total 20 trees measured  Tree morphology and condition  Tree morphology and  ARNEWS procedures not used  Tree morphology and per ARNEWS procedures for all condition per ARNEWS  Trees in the vegetation and soils condition per ARNEWS vegetation plot trees with procedures for all plots assessed visually for signs procedures for all vegetation DBH>10cm vegetation plot trees with of stem and crown damage plot trees with DBH>10cm Morphological  Tree morphology and condition DBH>10cm  Canopy photos not taken  Tree morphology and and Health per ARNEWS procedures for the  Tree morphology and condition per ARNEWS Measurements numbered 10 off-plot trees condition per ARNEWS procedures for the 10  Wide-angle photograph of the procedures for the 10 numbered off-plot tree tree canopy from ground level to numbered off-plot trees  Canopy photos not taken assess canopy cover  Canopy photos not taken  1 branch having a minimum of 8  1 branch having a  1 branch from each of 5 trees  1 branch from each of 5 years of growth obtained from minimum of 8 years of near the vegetation plot, but not numbered off-plot trees the upper third of the crown from growth was cut from the always from the numbered off-plot randomly chosen from the each of the 10 numbered off-plot upper third of the crown trees, cut from upper third of pool of 10 numbered off-plot trees, and two branches from the from each of the 10 crown on side of tree facing oil trees cut from upper third of crown of 7 other randomly numbered off-plot trees sands emission sources. Trees crown, on the side facing the selected trees (total of 24  ARNEWS procedures were in the dominant or co- oil sands emission sources branches) used to: dominant canopy class  Branches photographed Shoot Growth  ARNEWS procedures used to:  Measure annual shoot  Needle characteristics (chlorosis,  ARNEWS procedures used  Measure annual shoot growth growth for minimum of tip necrosis) and years of to: for minimum of 8 years 8 years (CAG+7 retention recorded (ARNEWS  Measure annual shoot (CAG+7 years) measured to years), to nearest procedures not used) growth for minimum of 8 nearest 0.2mm 0.2mm  Branches photographed years (CAG+7 years), to  Estimate defoliation in each  Estimate defoliation in  No measurement of lateral shoot 0.2mm (calipers) years needle growth class, as each years needle growth  Estimate defoliation in each percentages growth class, as years needle growth class, percentages as percentages

194

 1 increment core taken from  1 increment core taken  Increment cores obtained from the  Increment cores taken only each of the 10 numbered off-plot from each of the 10 4 largest-diameter, undamaged, from new numbered off-plot trees numbered off-plot trees dominant/co-dominant trees per trees at new sites and from Growth Rings  Cores mounted, sanded,  Cores mounted, sanded, plot (vegetation plot, 4 soils plots) replacement off-plot trees at measured with calipers to digitally-scanned and  Age determined by counting rings, existing sites nearest 0.02mm, under a 20x measured using using a dissecting microscope microscope computer software  Site Index derived from  No measurements  Site Index derived from  No measurements specifically measurements of and samples specifically driven by Site measurements of the 4 largest- driven by Site Index Site Index taken from the numbered off-plot Index calculations taken diameter, undamaged, calculations taken trees dominant/co-dominant trees per plot (vegetation plot, 4 soils plots)

195

Appendix 5.2 Jack Pine Tissue Sampling Sample 1998 2001 2004 2011/2012  Needle sampling in late  Needle sampling in  Needle sampling in late August  Needle sampling in late August August & early late August  Needle samples collected from branches and early September September  Three needle samples cut from the upper third of the crown, on  Three needle samples taken from  Current Annual Growth taken from each the side facing the oil sands emissions, of 1 branch cut from each of 5 (CAG) sampled from branch harvested from 5 trees in the dominant or co-dominant randomly chosen numbered off- each branch taken from each of the numbered canopy class (“top height” trees) to plot tree from the pool of 10 each of the 10 numbered off-plot trees: “…ensure that they are maximally exposed available numbered off-plot trees: off-plot trees, composited  CAG to potential emissions…” (frequently not  CAG into 1 needle sample per  1-year old (previous the identified off-plot trees). Sampled  1-year old (previous year) site year) branch contained at least 3 needle age  2-years old  Analysed for:  2-years old classes  No sample compositing  Total N & total C, by  No compositing, total  Three needle samples taken from each  From one branch, twice the CAG, LECO combustion of 30 samples branch harvested per off-plot tree: 1-yr-old & 2-yr-old sample mass is  Total elements analysed per site  CAG to be taken, and each divided into (including N, P, K, Ca,  Analysed for:  1-year old (previous year) two samples, providing one field Needles Mg, S, Fe, Mn, B, Zn,  Total N by micro-  2-years old duplicate sample per site. Total Cu, Mo, Na, Al, Pb, Ni, Kjeldahl  Compositing to create one sample per age number of samples per site = 18, Cd), by strong acid  Total S, by strong class per site. Total number of samples per all analysed digestion and ICP acid digestion and site = 3  Analyses:  Inorganic S by ICP, & by LECO  100-needle mass determined  Total S & N by LECO dry Johnson-Nishita combustion  Analyses: combustion distillation  Inorganic S by  Total S by LECO combustion  Inorganic S by Johnson-Nishita Johnson-Nishita  Inorganic S by Johnston-Nishita (Bi distillation) and by weak acid distillation  Total elements by ICP-MS extraction/ion chromatography  Total elements  Elemental concentrations by (including N, P, K, strong acid digestion and FLAA, Ca, Mg, S, Fe, Mn, CVAA, GFAA, ICP-AES or ICP- B, Zn, Cu, Mo, Na, MS. Focus on analysis for Al, Al, Pb, Ni, Cd), by Ca, Cu, Fe, Mg, Mn, Mo, Ni, P, strong acid K, Na, S & Zn digestion and ICP

196

Sample 1998 2001 2004 2011/2012  1 bark sample taken from  Not sampled  Not sampled  Not sampled each of 10 randomly selected trees near the vegetation plot (not the numbered off-plot trees), composited into 1 bark Bark sample per site  Analysed for 25 elements (including N, P, K, Ca, Mg, S, Fe, Mn, B, Zn, Cu, Mo, Na, Al, Pb, Ni, Cd) by ICP  1 twig sample (shoots up  Not sampled  Not sampled  Not sampled to 3 years old) taken from each of the 10 numbered off-plot trees, composited into 1 twig sample per site  Analysed for 25 elements Twig (including N, P, K, Ca, Mg, S, Fe, Mn, B, Zn, Cu, Mo, Na, Al, Pb, Ni, Cd) by ICP  Data used as FORSUST model input

197

Sample 1998 2001 2004 2011/2012  1 tree core taken from  Not sampled  Not sampled  Not sampled each of 10 randomly selected trees near the vegetation plot (not the numbered off-plot trees), composited into 1 stemwood sample per Stemwood site  Analysed for 25 elements (including N, P, K, Ca, Mg, S, Fe, Mn, B, Zn, Cu, Mo, Na, Al, Pb, Ni, Cd) by ICP  Data used as FORSUST model input Field Sampling (1 CAG Field Sampling (3 each Field Sampling (5 each of CAG, 1-yr and 2-yr Field Sampling (1 each of CAG, 1-yr Total sample from 10 trees at of CAG, 1-yr and 2-yr age classes, from 5 trees at each of 13 sites): and 2-yr age classes, from 5 trees at Number of each of 10 sites): age classes, from 10  Needles: 195 each of 25 sites): Individual  Needles: 100 trees at 1 site):  Needles: 375 Samples  Bark: 100  Needles: 30 Collected  Twigs: 100  Cores: 100 Lab Compositing: No compositing: Lab Compositing: No compositing: Samples  Needles: 1  Needles: 30  Needles: 3  Needles: 15 Analysed per  Bark: 1 Site  Twigs: 1  Cores: 1

198

Appendix 5.3 Soil Sampling in the TEEM Jack Pine Monitoring Program 1998 2001 2004 2011/2012 Field and Laboratory Compositing by Deposition Field Compositing by Deposition Field Compositing by No Compositing Zone Zone Deposition Zone Field Field Compositing Lab Compositing Field Field Compositing Field Field Compositing Field sampling (per site): Sampling (one composite (one composite Sampling (one composite Sampling (one composite  Existing sites: 4 (individual sample per depth, sample per depth, (individual sample per depth, (individual sample per depth, depths individually samples): per main plot): per site): samples): per main plot): samples): per site): sampled per subplot  SF: 48  SF: 12  SF: 3  SF: 48  SF: 12  TF: 32  TF: 4 = 64 samples  TF: 80  TF: 20  FF: 3  TF: 80  TF: 20  FF (new  FF (new sites):  New sites: 5 depths  FF: 48  FF: 12 No lab  FF: 48  SF: 12 sites): 24 1 individually sampled compositing for TF  FF  FF (existing per subplot = 80 samples: (existing sites): 1 samples  TF: 20 sites): 16 26 composite samples analysed per site: 44 composite samples analysed 6 (existing sites) or 7 (new sites) 80 (new sites) or 64  SF: 1 each for LFH, 0-5cm & 5-15cm per site: composite samples analysed per (existing sites)  TF: 4 each for LFH, 0-5cm, 5-15cm, 15-30cm & 30-  SF: 4 each for LFH, 0-5cm & 5- site: individual samples 50cm 15cm  TF: 1 each for LFH, 0-5cm, 5- analysed per site (no  FF: 1 each for LFH, 0-5cm & 5-15cm  TF: 4 each for LFH, 0-5cm & 5- 15cm & 15-30cm (all sites) compositing) 15cm, 15-30cm, 30-50 cm  FF: 1 each for LFH & 0-5cm (all  FF: 4 each for LFH, 0-5cm & 5- sites) & 5-15cm (new sites only) 15cm

199

Appendix 5.4 Soil Sample Laboratory Procedures in the TEEM Jack Pine Monitoring Program 1998 2001 2004 2011/2012  Air-dried samples, ground through  Air-dried samples, ground through  Air-dried samples, ground  Air-dried samples, ground 2mm screen (exceptions for LFH) 2mm screen (exceptions for LFH) through 2mm screen through 2mm screen (exceptions  Analyses:  Analyses: (exceptions for LFH) for LFH)  pH in CaCl2  pH in CaCl2  Analyses:  Analyses:  Total organic carbon by LECO  Total sulphur, carbon, nitrogen in  pH in CaCl2  pH in CaCl2 combustion LFH by LECO combustion  Total sulphur, carbon,  Soluble ions, extracted in  Total nitrogen by Dumas combustion  Al, Ba, Be, Cd, Ca, Cr, Co, Cu, Fe, nitrogen by LECO water & analysed by ICP & LECO detection Pb, Mg, Mn, Mo, P, K, Si, S, Ag, Na, combustion  CEC, extract in NH4Cl &  Al, Ba, Be, Cd, Ca, Cr, Co, Cu, Fe, Sr, Th, Ti, V & Zn by strong acid  CEC (NH4Cl) analysed by distillation of NH4 Pb, Mg, Mn, Mo, P, K, Si, S, Ag, Na, digestion and ICP  Exchangeable ions (Ca, Mg, or by NH4 electrode Sr, Th, Ti, V & Zn by strong acid  CEC (sum of Ca/Mg/K/Na & Na, K, Fe, Mn, Al, Si) by ICP  Exchangeable cations, digestion and ICP H/Al/Fe/Mn), extracted in NH4Cl, of the unbuffered CEC extracted in NH4Cl & analysed  CEC (sum of Ca/Mg/K/Na & measured by NH4 electrode or extract by ICP-AES H/Al/Fe/Mn), extracted in NH4Cl, distillation  Available nitrate, by  BC:Al calculated measured by NH4 electrode or  Exchangeable ions, by ICP of the extraction in KCl and  BS%, calculated (CEC & distillation unbuffered CEC extract colourimetric determination exchangeable cations  Exchangeable ions by ICP of the  Available nitrogen by extraction in  Available phosphorus, by extracted in NH4Cl) unbuffered CEC extract KCl & ion chromatography (NO3) or NH4F/H2SO4 extraction &  Total carbon, nitrogen &  Available nitrogen by extraction in electrode (NH4) colourimetric analysis sulphur by LECO dry KCl & ion chromatography (NO3) or  Available phosphorus by  Available potassium by combustion electrode (NH4) NH4F/H2SO4 extraction & acetate extraction & analysis  C:N calculated  Available phosphorus by colourimetric analysis by flame-emission  Soluble (available) N by KCl NH4F/H2SO4 extraction &  Available potassium,by ICP spectroscopy or ICP extraction, colourimetric, colourimetric analysis  SO4-S (LFH) by NH4Cl extraction &  SO4-S (LFH) and SO4-S autoanalyser or ion-selective  Available potassium by ICP ion chromatography (mineral soil) electrode analysis  SO4-S (LFH) by NH4Cl extraction &  SO4-S (mineral soil) by Ca3(PO4)2  Electrical Conductivity by  Soluble (available) P by NH4F ion chromatography extraction & ion chromatography electrode extraction, analysis by  SO4-S (mineral soil) by Ca3(PO4)2  Electrical Conductivity by electrode  Water-soluble ions autoanalyser extraction & ion chromatography  Water-soluble ions  BC:Al calculated  Inorganic sulphur by NH4Cl  Electrical Conductivity by electrode  Base saturation % calculated (ratio  Base saturation % (BSAM) (LFH) or Ca(H2PO4)2  Water-soluble ions of base cations to total cation calculated (ratio of base extraction, ICP-AES analysis  Base saturation % calculated (ratio exchange capacity) cations to total cation of base cations to total cation  Ca:Al ratio calculated exchange capacity) exchange capacity)  Ca:Al ratio calculated

200

Chapter 6 Appendices Appendix 6.1 Stand age and summary information on density, diameter, height, crown depth, and canopy closure for canopy (above or at the general canopy height) and sub-canopy (just below or completely below the general canopy height) trees at each site. Stand Diameter at 1.3 m Density (trees/ha) Height (m) Crown Depth (m) Crown Closure7 Age height (m) (in Sub- Sub- Sub- Sub- Sub- Site 2011)1 Canopy canopy Canopy canopy Canopy canopy Canopy canopy Canopy canopy JP 101 612 500 100 15.94 14.83 14.60 12.00 9.48 8.33 0.90 0.00 JP 102 683 375 300 20.95 13.73 18.07 13.96 10.31 6.52 1.85 1.71 JP 1035 792 1325 15.19 14.50 . . . . JP 104 752 600 175 19.99 15.04 18.60 16.23 8.07 5.68 2.41 2.92 JP 1066 82 1125 25 14.40 11.30 13.37 10.00 4.22 1.10 2.21 2.00 JP 1075 732 1375 12.21 13.49 . . . . JP 108 743 600 550 18.93 12.90 13.88 11.95 5.77 3.34 0.77 1.55 JP 1095 722 1075 13.57 16.15 . . . . JP 201 68 1300 25 14.22 11.90 14.23 10.80 6.46 6.10 2.10 . JP 2055 654 800 13.44 11.24 . . . . JP 210 83 875 175 15.76 11.81 15.14 12.04 8.66 5.93 2.03 3.25 JP 212 624 1100 425 16.14 12.06 15.93 12.51 8.02 6.07 1.82 2.00 JP 2136 67 400 350 14.95 12.00 12.29 9.70 7.24 4.74 1.00 1.07 JP 303 81 575 125 17.75 13.85 16.22 10.50 9.88 5.49 1.46 1.50 JP 304 71 675 175 15.20 11.04 13.99 11.16 8.20 6.36 1.87 1.50 JP 307 78 350 50 21.20 12.40 15.22 10.40 6.81 3.25 0.79 0.50 JP 308 65 875 275 12.97 11.19 15.11 13.44 6.71 6.31 1.94 2.00 JP 310 71 1325 75 13.15 10.47 14.61 13.07 4.82 3.83 1.68 2.00 JP 311 55 450 50 17.37 12.85 15.19 10.65 10.06 8.65 1.79 1.50 JP 312 88 425 250 15.58 12.61 10.01 7.51 5.92 2.98 0.87 0.94 JP 313 90 600 125 17.33 12.69 13.16 9.45 7.59 3.38 1.58 0.83 JP 315 80 175 175 28.45 17.42 18.49 10.13 14.91 4.97 1.00 0.83 JP 316 70 1100 0 14.13 . 14.62 . 7.04 . 2.34 . JP 317 93 450 325 18.93 12.09 14.39 9.19 10.86 5.84 1.43 0.80 JP 318 87 525 375 16.45 12.12 15.23 9.96 8.98 6.18 1.43 0.94

201

1 from tree cores taken in 2011/2012; 2from AMEC Earth & Environmental Ltd. 2000; 3 from AMEC Earth & Environmental Ltd. 2000 and tree cores taken in 2011; 4from Jones and Assoc. 2006. 5sampled after the 2011 wildfire, canopy status of trees not recorded; 6sampled after the 2011 wildfire but all data could be collected; 7 in classes: 0 = not touching adjacent trees; 1 = touching on 1 side, 2 = touching on 2 sides, 3 = touching on 3 sides, 4 = touching on 4 sides Appendix 6.2 Summary information on branch defoliation and internode length for the current year, and 1-, 2-, 3-, and 4-year old sections of the branch.

Defoliation1 Internode Length (cm) Current Current site current 1 yr old 2 yr old 3 yr old 4 yr old to 4 yrs current 1 yr old 2 yr old 3 yr old 4 yr old to 4 yrs JP101 0.00 0.20 0.20 2.20 3.60 1.24 5.06 4.10 4.38 5.46 5.76 4.95 JP102 1.60 1.60 2.20 2.80 3.20 2.28 3.10 2.94 2.96 2.84 3.28 3.02 JP104 1.60 1.60 2.20 2.20 2.80 2.08 3.96 3.78 3.80 3.42 4.74 3.94 JP1062 0.20 1.00 1.40 2.40 3.80 1.76 2.50 3.18 4.60 3.56 4.86 3.74 JP108 0.40 0.60 1.20 2.20 3.60 1.60 4.76 4.66 4.44 4.42 4.46 4.55 JP201 0.00 0.60 1.00 1.00 3.40 1.20 2.58 3.06 3.56 4.86 4.96 3.80 JP210 0.00 0.60 1.00 1.60 2.60 1.16 4.30 3.74 4.48 4.44 5.58 4.51 JP212 0.00 0.20 1.00 0.60 2.40 0.84 3.02 3.24 3.26 3.22 3.58 3.26 JP2132 1.00 0.40 2.40 2.20 3.00 1.80 2.02 2.38 2.38 2.46 3.64 2.58 JP303 0.60 0.80 1.60 1.00 3.00 1.40 3.68 3.62 3.90 4.16 3.64 3.80 JP304 0.60 0.60 0.20 0.80 2.40 0.92 3.80 3.94 3.86 3.80 3.98 3.88 JP307 1.80 0.00 1.20 1.60 3.00 1.52 4.26 4.42 4.56 4.58 5.00 4.56 JP308 0.40 0.80 1.40 2.00 2.80 1.48 3.18 3.02 3.18 3.22 2.66 3.05 JP310 0.60 0.40 1.60 1.60 1.60 1.16 3.34 3.32 3.68 4.26 4.70 3.86 JP311 0.00 0.00 1.40 2.60 3.00 1.40 5.08 5.58 6.42 5.64 7.45 6.03 JP312 0.60 0.20 1.20 1.60 2.40 1.20 2.64 2.44 2.98 2.58 3.82 2.89 JP313 0.20 1.20 1.60 2.00 3.40 1.68 3.32 3.54 3.34 3.76 3.44 3.48 JP315 0.00 0.20 1.20 1.40 3.20 1.20 4.52 4.42 3.80 4.40 4.00 4.23 JP316 0.00 0.00 0.60 1.80 2.80 1.04 4.88 4.82 4.56 4.08 4.72 4.61 JP317 0.60 0.80 1.40 1.60 3.40 1.56 5.70 3.92 4.80 5.62 5.92 5.19 JP318 0.20 1.60 1.20 3.00 3.40 1.88 5.18 4.00 4.52 4.78 5.12 4.72 1 in classes: 0 = no defoliation, 1: ≤ 25%, 2: ≤ 50%, 3: ≤ 75%, 4: ≤ 100% defoliation; 2stand sampled after the 2011 wildfire;

202

Chapter 8 Appendices Appendix 8.1 Concentrations of jack pine needles at the routine monitoring sites sampled in 2011 (sites 317 and 318 sampled in 2012). Due to large size of the data set, data are available in Excel file format from [email protected]

Appendix 8.2 Soil chemical analyses at the routine monitoring sites sampled in 2011 (sites 317 and 318 sampled in 2012).

Due to large size of the data set, data are available in Excel file format from [email protected]

203

Chapter 9 Appendices Appendix 9.1 Results of regression analyses (significance, R2, and slope estimate) for each understory response variable as a function of air quality predictor variables calculated based on emissions from sources using the CALPUFF model (from Davies, Bajwa and Person). The understory vegetation response variables were regressed on: predicted sulphur (S) and nitrogen (N) compound concentrations, S and N and base cation (BC) deposition, and potential acid input (PAI). Deposition values were calculated based on just sources in the Lower Athabasca Region (LAR) and for all sources (total). Base cation values were calculated using a throughfall assumption and a bulk assumption. Likewise Total PAI predictions were calculated for both the bulk and throughfall BC assumptions. Shaded cells indicate a non-significant regression. The following understory vegetation response variables were not significantly related to any of the predictor variables: Total cover, shrub cover, bryophyte cover, lichen cover, lichen richness. Total Richness Vascular Cover Vascular Richness

P R2 estimate P R2 estimate P R2 estimate

1 SO2 concentration 0.028 0.253 2.19 0.008 0.345 4.27 0.049 0.210 0.05

1 NO2 concentration 0.020 0.281 0.52 0.002 0.443 1.09

S deposition (LAR)2 0.030 0.247 33.45 0.006 0.363 67.81

Total S deposition2 0.030 0.249 34.64 0.007 0.356 69.24

N deposition (LAR)2 0.028 0.255 22.85 0.003 0.411 48.54

Reduced N deposition2 0.012 0.319 107.46

Total N deposition2 0.033 0.241 16.65 0.004 0.401 35.89

Base Cation (throughfall)2

Total PAI (throughfall)2

Base Cation (bulk)2

Total PAI (bulk)2 0.029 0.252 5.83 0.040 0.226 9.25 0.017 0.293 6.05

204

Forb Cover Shrub Richness Forb Richness Bryophyte Richness P R2 estimate P R2 estimate P R2 estimate P R2 estimate

1 SO2 concentration 0.016 0.296 3.34 0.024 0.265 0.39

1 NO2 concentration 0.005 0.376 0.85 0.023 0.267 0.09

S deposition (LAR)2 0.014 0.307 52.69 0.043 0.220 5.46

Total S deposition2 0.017 0.293 53.01 0.046 0.214 5.55

N deposition (LAR)2 0.009 0.339 37.25 0.033 0.240 3.83

Reduced N deposition2

Total N deposition2 0.014 0.305 26.47 0.042 0.221 2.76

Base Cation (throughfall)2

Total PAI (throughfall)2

Base Cation (bulk)2 0.025 0.261 1.50

Total PAI (bulk)2 0.008 0.350 9.71 0.013 0.311 5.07 0.018 0.289 -1.63 1 µg/m3 2 keq/ha/yr

205

Appendix 9.2 Results of regression analyses (significance, R2, and slope estimate) for each understory response variable as a function of distance from center of industrial activity and atmospheric deposition variables modelled based on data from ion exchange resins at a network of sites in the region (from Fenn). The understory vegetation response variables were regressed on: predicted deposition of ammonium (NH4-N), nitrate (NO3-N), total dissolved inorganic nitrogen (DIN), sulfate (SO4-S), and base cations (BC). For each of these, predicted values were modelled for bulk input and for throughfall through the forest canopy. Shaded cells indicate a non-significant regression. The following understory vegetation response variables were not significantly related to any of the predictor variables: Total cover, shrub cover, bryophyte cover, lichen cover, lichen richness. Total Richness Vascular Cover Vascular Richness

P R2 estimate P R2 estimate P R2 estimate

Distance from source1 0.023 0.269 -0.07 0.013 0.314 -0.12

2 NH4-N deposition (bulk) 0.015 0.301 5.34 0.001 0.470 11.16

2 NH4-N deposition (throughfall) 0.024 0.267 3.71 0.001 0.469 8.23

NO3-N deposition (bulk) 0.015 0.302 11.66 0.001 0.459 24.01

2 NO3-N deposition (throughfall) 0.018 0.288 4.87 0.001 0.475 10.46

2 SO4-S deposition (bulk) 0.015 0.302 2.51 0.001 0.467 5.23

2 SO4-S deposition (throughfall) 0.015 0.302 0.97 0.001 0.467 2.02

Dissolved Inorganic N (bulk)2 0.015 0.302 3.64 0.001 0.466 7.56 Dissolved Inorganic N (throughfall)2 0.021 0.276 2.03 0.001 0.473 4.45

Base Cations (bulk)3 0.015 0.303 14.0 0.001 0.466 29.04

Base Cations (throughfall)3 0.014 0.304 7.45 0.001 0.463 15.38 0.050 0.208 5.92

206

Forb Cover Shrub Richness Forb Richness Bryophyte Richness

P R2 estimate P R2 estimate P R2 estimate P R2 estimate

Distance from source1 0.011 0.325 -0.11 0.029 0.252 -0.011

2 NH4-N deposition (bulk) 0.003 0.424 8.95 0.017 0.292 0.909 NH4-N deposition (throughfall)2 0.003 0.414 6.53 0.026 0.260 0.634

NO3-N deposition (bulk) 0.003 0.419 19.38 0.017 0.292 1.979 NO3-N deposition (throughfall)2 0.003 0.422 8.33 0.020 0.281 0.831

2 SO4-S deposition (bulk) 0.003 0.423 4.20 0.017 0.293 0.428 SO4-S deposition (throughfall)2 0.003 0.422 1.62 0.017 0.293 0.165 Dissolved Inorganic N (bulk)2 0.003 0.421 6.07 0.017 0.293 0.619 Dissolved Inorganic N (throughfall)2 0.003 0.419 3.54 0.023 0.269 0.346

Base Cations (bulk)3 0.003 0.422 23.36 0.016 0.294 2.385

Base Cations (throughfall)3 0.003 0.421 12.39 0.016 0.295 1.268 1 km, 2kg/ha/yr, 3keq/ha/yr

207

Appendix 9.3 Complete list of species encountered during sampling of the sites in 2011/2012; the seven letter code in brackets is for those species shown in Figure 9.2. Nomenclature follows the USDA Plants database. (http://plants.usda.gov/java/ access September 2014) Herbaceous Vascular plants Woody Vascular plants Anemone multifida Poir. Alnus viridis (Chaix) DC. Anemone patens (L.) P. Mill. Amelanchier alnifolia (Nutt.) Nutt. ex Roemer Apocynum androsaemifolium L. Betula papyrifera Marsh. Aralia nudicaulis L. Picea mariana (Mill.) Britton, Sterns & Poggenb. Arctostaphylos uva-ursi (L.) Spreng. (ARCTUVA) Pinus banksiana Lamb. (seedlings) Aster ciliolatus (Lindl.) Löve Populus tremuloides Michx. Campanula rotundifolia L. Prunus pensylvanica L. f. Carex siccata Dewey Prunus virginiana L. Carex spp. (various) Rosa acicularis Lindl. Chamerion angustifolium (L.) Holub Salix spp. (various) Comandra umbellata (L.) Nutt. Shepherdia canadensis (L.) Nutt. Cornus canadensis L. (CORNCAN) Vaccinium myrtilloides Michx. (VACCMYR) Dracocephalum parviflorum Nutt. Vaccinium vitis-idaea L. (VACCVIT) Equisetum arvense L. Ledum groenlandicum Oeder Equisetum hyemale L. Erigeron acris L. Mosses Festuca saximontana Rydb. Aulocomnium palustre (Hedw.) Schwägr. Galium boreale L. Dicranum polysetum Sw. Galium trifidum L. Dicranum scoparium Hedw. Geocaulon lividum (Richards.) Fern. Hylocomium splendens (Hedw.) Schimp. Geranium bicknellii Britt. Orthotrichum speciosum Nees Goodyera repens (L.) R. Br. Pleurozium schreberi (Brid.) Mitt. Hierochloe odorata (L.) Beauv. Pohlia nutans (Hedw.) Lindb. Leymus innovatus (Beal) Pilger Polytrichum commune Hedw. Linnaea borealis L. (LINNBOR) Polytrichum juniperinum Hedw. Lycopodium annotinum L. Polytrichum strictum Brid. Lycopodium complanatum L. Ptilidium pulcherrimum (Weber) Vain. Maianthemum canadense Desf. (MAIACAN) Ptilium crista-castrensis (Hedw.) De Not. Melampyrum lineare Desr. Abietinella abietina (Hedw.) Fleisch. Tortula ruralis (Hedw.) G. Gaertn., Meyer & Mitella nuda L. Scherb. Oryzopsis asperifolia Michx. Sibbaldiopsis tridentata (Ait.) Rydb. Lichens Symphyotrichum laeve (L.) Löve Cladina mitis (Standst.) Hustich Pyrola asarifolia Michx. Cladina rangiferina (L.) Nyl. Pyrola secunda (L.) House Cladina stellaris (Opiz) Brodo Maianthemum stellatum (L.) Link Cladonia gracilis (L.) Willd. Maianthemum trifolium (L.) Sloboda Cladonia spp. (various)

208

Solidago simplex var. simplex Knuth Cladonia uncialis (L.) Weber ex Wigg. Viola adunca Sm. Flavocetraria cucullata (Bellardi) Karnefelt & Thell Viola canadensis L. Flavocetraria nivalis (L.) Karnefelt & Thell Peltigera aphthosa (L.) Willd. Peltigera leucophlebia (Nyl.) Gyelnik Peltigera neopolydactyla (Gyelnik) Gyelnik one unknown

Chapter 10 Appendices Appendix 10.1 Elements analysed in Hypogymnia physodes in 2011. [Only those in bold are considered in this report] Aluminum Lead Rubidium Antimony Lithium Samarium Arsenic Magnesium Silicon Barium Manganese Sodium Beryllium Mercury Strontium Bismuth Molybdenum Sulphur Cadmium Neodymium Tantalum Calcium Nickel Thallium Chromium Niobium Thorium Cerium Nitrogen Tin Cesium Palladium Titanium Cobalt Phosphorus Uranium Copper Platinum Vanadium Iron Potassium Wolfram Lanthanum Praseodymium Zinc

209