UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Relative bulk density as an index of soil compaction and forest productivity in British Columbia Zhao, Yihai (Simon) 2009

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2009_fall_zhao_yihai.pdf [ 1.04MB ]
Metadata
JSON: 24-1.0070825.json
JSON-LD: 24-1.0070825-ld.json
RDF/XML (Pretty): 24-1.0070825-rdf.xml
RDF/JSON: 24-1.0070825-rdf.json
Turtle: 24-1.0070825-turtle.txt
N-Triples: 24-1.0070825-rdf-ntriples.txt
Original Record: 24-1.0070825-source.json
Full Text
24-1.0070825-fulltext.txt
Citation
24-1.0070825.ris

Full Text

  RELATIVE BULK DENSITY AS AN INDEX OF SOIL COMPACTION AND FOREST PRODUCTIVITY IN BRITISH COLUMBIA   by   YIHAI (SIMON) ZHAO   B.Sc., Gansu Agricultural University, 1992 M.Sc., Gansu Agricultural University, 1995   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF   DOCTOR OF PHILOSOPHY   in   THE FACULTY OF GRADUATE STUDIES (Forestry)   THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)   June 2009   © Yihai (Simon) Zhao, 2009  ii ABSTRACT Soil compaction often limits conifer regeneration on sites degraded by construction of landings and roads, but inadequate understanding of compaction characteristics has sometimes led to inappropriate rehabilitation efforts. This warrants development of new methods to assess compaction and its relation to tree growth. The objective of this study was to develop a high- level integration indicator that will characterize compaction of forest soils and that could be correlated to tree height growth.  Mineral particle density of soils from interior British Columbia (BC) forests varied significantly among the geographic locations. Oxalate-extractable Fe- and Al-oxides and particle size distribution (PSD) were related to soil and mineral particle densities, while soil organic matter (SOM) and Al- and Fe-oxides were important soil properties in relation to soil particle density. The significance of levels of single soil properties in predicting maximum bulk density (MBD) were in the order: plastic and liquid limits, organic matter content, oxalate- extractable oxide, and PSD. Stratification of the sample according to Atterberg limits improved the predictability of MBD, and variation in particle density was included in the prediction by other soil properties used in the models. Height growth of interior Douglas-fir (Pseudotsuga menziesii var. glauca [Bessin] Franco) was restricted when relative bulk density (RBD) was > 0.72. For lodgepole pine (Pinus contorta Dougl. ex Loud. var. latifolia Engelm.) and hybrid white spruce (Picea glauca [Moench] Voss × engelmannii Parry ex Engelm.), an RBD of 0.60 - 0.63 corresponded to maximum height growth, while that of 0.78 - 0.84 appeared to limit height growth. The presence of surface organic material mitigated compaction and was often associated with lower RBD. Interior Douglas-fir and lodgepole pine planted in low elevation sites in north-central BC did not grow well and their height growth was weakly related to RBD. The results suggest that soil rehabilitation should be considered on disturbed sites where soil RBD is > 0.80. Findings in this study have implications in assessing forest soil compaction and its effect on site productivity. The results will help predict soil behaviour and associated tree growth in response to timber harvesting and site rehabilitation.  iii TABLE OF CONTENTS Abstract ..........................................................................................................................................ii Table of contents .......................................................................................................................... iii List of tables .................................................................................................................................vii List of figures ................................................................................................................................ix List of abbreviations .....................................................................................................................xi Acknowledgements .................................................................................................................... xiii Co-authorship statement.............................................................................................................xv 1 Introduction ................................................................................................................................1 1.1 Compaction of forest soils .....................................................................................................1 1.2 Impacts of soil compaction on tree growth............................................................................2 1.3 Soil properties commonly used to characterize soil compaction...........................................5 1.4 Methods used for the assessment of soil resistance to compaction.......................................6 1.4.1 Proctor compaction test ..................................................................................................7 1.4.2 Uniaxial compression test...............................................................................................7 1.5 Soil properties that affect soil susceptibility to compaction..................................................8 1.5.1 Soil organic matter..........................................................................................................8 1.5.2 Particle size distribution ...............................................................................................10 1.5.3 Water content and water potential ................................................................................13 1.5.4 Mineral composition.....................................................................................................15 1.6 High-level integration indicators of soil compaction ..........................................................17 1.7 Study objectives...................................................................................................................20 1.8 References ...........................................................................................................................26 2 Use of a Constant-Volume, Small-Sample Gas Pycnometer to Test Variation in Particle Density of Forest Soils in Interior British Columbia................................................................39 2.1 Introduction .........................................................................................................................39  iv 2.2 Materials and methods.........................................................................................................42 2.2.1 Description of the constant-volume, small-sample gas pycnometer ............................42 2.2.2 Operation of the pycnometer ........................................................................................42 2.2.3 Calibration of the small-sample gas pycnometer..........................................................43 2.2.4 Test of samples .............................................................................................................45 2.2.5 Soil organic matter, oxalate-extractable oxides, and particle size distribution ............45 2.2.6 Calculation of mineral particle density.........................................................................46 2.3 Statistical analysis................................................................................................................46 2.4 Results and discussion .........................................................................................................47 2.4.1 Calibration curve ..........................................................................................................47 2.4.2 Accuracy (uncertainty) analysis ...................................................................................47 2.4.3 Test of commercial sands and soil samples..................................................................49 2.5 Conclusions .........................................................................................................................53 2.6 References ...........................................................................................................................68 3 Maximum Bulk Density of British Columbia Forest Soils from the Proctor Test: Relationships with Selected Physical and Chemical Properties ..............................................72 3.1 Introduction .........................................................................................................................72 3.2 Materials and methods.........................................................................................................76 3.2.1 Study sites.....................................................................................................................76 3.2.2 Soil analysis ..................................................................................................................77 3.2.2.1 Maximum bulk density ..........................................................................................77 3.2.2.2 Particle density ......................................................................................................78 3.2.2.3 Soil organic matter.................................................................................................79 3.2.2.4 Soil oxides .............................................................................................................79 3.2.2.5 Particle size distribution ........................................................................................79 3.2.2.6 Plastic and liquid limits .........................................................................................80 3.2.3 Statistical analysis.........................................................................................................80 3.3 Results and discussion .........................................................................................................81 3.3.1 Relationships between MBD, WMBD and other soil properties.....................................81 3.3.1.1 Particle density ......................................................................................................81  v 3.3.1.2 Soil organic matter.................................................................................................82 3.3.1.3 Soil oxides .............................................................................................................84 3.3.1.4 Particle size distribution ........................................................................................84 3.3.1.5 Plastic and liquid limits .........................................................................................85 3.3.2 Predicting MBD by a set of soil properties ..................................................................86 3.3.3 Proposed method for using MBD as a reference in forest soil compaction studies .....90 3.4 Conclusions .........................................................................................................................91 3.5 References .........................................................................................................................107 4 Characterization of Soil Compaction and Tree Height Growth by Relative Bulk Density113 4.1 Introduction .......................................................................................................................113 4.2 Materials and methods.......................................................................................................115 4.2.1 Site description ...........................................................................................................115 4.2.2 Field and laboratory methods .....................................................................................116 4.2.3 Data analysis...............................................................................................................118 4.3 Results ...............................................................................................................................119 4.3.1 Relative bulk density and surface organic material....................................................119 4.3.2 Relative bulk density and relative height growth index: north-central BC ................121 4.3.3 Relative bulk density and relative height growth index: hybrid white spruce ...........121 4.3.4 Relative bulk density and relative height growth index: interior Douglas-fir ............122 4.3.5 Relative bulk density and relative height growth index: lodgepole pine....................123 4.4 Discussion..........................................................................................................................123 4.5 Conclusions .......................................................................................................................130 4.6 References .........................................................................................................................143 5 Concluding Chapter ...............................................................................................................148 5.1 Synthesis of results ............................................................................................................148 5.2 Strengths of the thesis........................................................................................................150 5.3 Limitations of the thesis.....................................................................................................151 5.4 Future research...................................................................................................................153  vi 5.5 References .........................................................................................................................156 Appendices .................................................................................................................................158 Appendix 1: Relationship between mineral particle density and total sand for forest soils from interior British Columbia.........................................................................................................158 Appendix 2: Maximum bulk density, WMBD obtained by the Proctor test, and other soil physical and chemical properties tested for the MBD Study. .................................................159 Appendix 3: Relationship between maximum bulk density and particle density for forest soils collected in British Columbia. .................................................................................................169 Appendix 4: Relationship between maximum bulk density and total C, and oxidisable organic matter for forest soils collected in British Columbia...............................................................170 Appendix 5: Relationship between critical water content and total C, and oxidisable organic matter for forest soils collected in British Columbia...............................................................171 Appendix 6: Relationship between maximum bulk density and Al-oxide, and Fe-oxide for forest soils collected in British Columbia................................................................................172 Appendix 7: Relationship between critical water content and Al-oxide, and Fe-oxide for forest soils collected in British Columbia................................................................................173 Appendix 8: Relationship between maximum bulk density and liquid limit, and plastic limit for forest soils collected in British Columbia. .........................................................................174 Appendix 9: Relationship between critical water content and liquid limit, and plastic limit for forest soils collected in British Columbia................................................................................175 Appendix 10: Regression analysis between relative height growth index and relative bulk density when data were grouped by presence of the surface organic material........................176 Appendix 11: Relationship between relative height growth index and bulk density for Douglas-fir at experiment 2, hybrid white spruce at experiment 4, and lodgepole pine at experiment 1. ...........................................................................................................................178 Appendix 12: Glossary ............................................................................................................180  vii LIST OF TABLES Table 1.1 Relationship between maximum bulk density and soil organic matter reported in the literature..........................................................................................................................23 Table 1.2 Relationships between maximum bulk density and critical water content at which MBD was reached...........................................................................................................24 Table 2.1 Properties of the small-sample gas pycnometer. ...........................................................54 Table 2.2 Uncertainty of the small-sample gas pycnometer with the change of volume of the solids. ..............................................................................................................................55 Table 2.3 Comparison of particle density values obtained by the liquid pycnometer and the small-sample gas pycnometer for three types of commercial sands...............................56 Table 2.4 Selected properties for soils collected from eight forest sites in interior British Columbia. .......................................................................................................................57 Table 2.5 Grouping of the mineral particle density for samples collected from eight forest sites in interior British Columbia............................................................................................58 Table 2.6 Relationships between mineral particle density and selected soil properties. ...............59 Table 2.7 Relationships between soil particle density and selected soil properties. .....................60 Table 3.1 Site description, biogeoclimatic zones, and annual precipitation for 33 study sites throughout British Columbia. .........................................................................................93 Table 3.2 Soil properties for 33 study sites in British Columbia. .................................................95 Table 3.3 Correlation coefficients for comparisons among 17 selected variables. .......................96 Table 3.4 Relationships among maximum bulk density and particle size properties obtained at 33 study sites in British Columbia..................................................................................98 Table 3.5 Principal component analysis loadings for the first three components of individual variables..........................................................................................................................99 Table 3.6 Regression constants and correlation coefficients for relationships between maximum bulk density as the dependent variable and selected soil properties as the independent variable. ........................................................................................................................100 Table 4.1 Location, elevation, annual precipitation, temperature, and soil texture for the experiments...................................................................................................................131 Table 4.2 Establishment time, treatments, number of soil samples, tree species, and year of tree height measurements for the experiments included into this study..............................132 Table 4.3 Four regression models used to derive maximum bulk density. .................................133  viii Table 4.4 Ranges of relative bulk density and site indices for interior Douglas-fir, lodgepole pine, and hybrid white spruce of the experiments. .......................................................134 Table 4.5 Regression analysis of thickness of surface organic material and relative bulk density on the relative height growth index. .............................................................................135  ix LIST OF FIGURES Figure 1.1 General bulk density - moisture curve obtained with the standard Proctor test for a coarse -textured soil.....................................................................................................25 Figure 2.1 Diagram of a constant-volume gas pycnometer...........................................................61 Figure 2.2 Photo of the small-sample, constant-volume gas pycnometer. ....................................62 Figure 2.3 Discs used for the calibration process of the small-sample gas pycnometer. ..............63 Figure 2.4 Calibration model obtained for the small-sample gas pycnometer. .............................64 Figure 2.5 Sensitivity analysis of the small-sample pycnometer. .................................................65 Figure 2.6 Relationship between mineral particle density and soil organic matter, and soil particle density and soil organic matter with soil organic matter content < 120 g kg-1 for forest soils from the interiror British Columbia.....................................................66 Figure 2.7 Relationships between soil particle density and soil organic matter content grouped by sampling locations. .................................................................................................67 Figure 3.1 Location of 33 study sites in British Columbia..........................................................102 Figure 3.2 Change of soil bulk density and porosity,with water content in the standard Proctor test..............................................................................................................................103 Figure 3.3 Plasticity of soils from the study areas, showing highly plastic soils, and soils with moderate and low plasticity as plotted on the Casagrande chart. ..............................104 Figure 3.4 Relationships among maximum bulk density and clay+silt for all samples, nonplastic samples, moderate and low plastic samples, highly plastic samples, and samples from oilfield rehailitation sites and roads and landings with total C < 30 g kg-1. ..............105 Figure 3.5 Relationships between clay and total C for all samples, nonplastic samples, moderate and low plastic samples, and highly plastic samples.................................................106 Figure 4.1 Relationship between relative bulk density and field bulk density for cohesive and non-cohesive soils, and cohesive soils with and without surface organic material...137 Figure 4.2 Relationship between the relative height growth index of interior Douglas-fir, lodgepole pine, and hybrid white spruce and relative bulk density for the seventh growing season at experiment 4. ...............................................................................139 Figure 4.3 Relationship between relative height growth index of hybrid white spruce and relative bulk density at experiment 4 with all data, and with outliers being removed.140 Figure 4.4 Relationship between relative height growth index of interior Douglas-fir and relative bulk density at experiment 2, and at experiment 4. ...................................................141  x Figure 4.5 Relationship between relative height growth index of lodgepole pine and relative bulk density at experiment 1, and at experiments 1, 3, and 5....................................142  xi LIST OF ABBREVIATIONS ASTM – American Society for Testing Materials. BEC – Biogeoclimatic ecosystem classification, a multi-scaled, ecosystem-based classification system that groups ecologically similar site based on climate, soils and vegetation. This classification is widely used throughout British Columbia as a framework for resource management and scientific research. BWBS – Boreal White and Black Spruce BEC zone. C – compression index, the slope of the linear portion of the relationship between porosity or bulk density and the logarithm of applied pressure obtained by the uniaxial compression test. CDF – Coastal Douglas-fir BEC zone. CWH – Coastal Western Hemlock BEC zone. D – degree of compactness, the field bulk density divided by the reference bulk density that is tested by the uniaxial compression test. fMBD – porosity at MBD. GP – gas pycnometer, a laboratory device that employs gas displacement method and that is used for measuring the density or volume of solids. ICH – Interior Cedar-Hemlock BEC zone. IDF – Interior Douglas-fir BEC zone. LLWR – least limiting water range, the range of water contents associated with field capacity, wilting point, soil resistance, and soil aeration at which plant growth would cease or be dramatically reduced. LP – liquid pycnometer, a laboratory device that employs liquid displacement method and that is used to determine the density of a liquid or a solid. LTSP – Long-Term Soil Productivity study, a North American network research program set up in 1990 that studies effects of compaction and organic matter removal on soil productivity in the long term. The network has over 60 installations in the US and Canadian Forests. Participants include U.S. Department of Agriculture, U.S. Forest Service, Canadian Forest Service, British  xii Columbia Ministry of Forests and Range, and various universities and industry groups. MBD – maximum bulk density obtained by the Proctor test. PSD – particle size distribution, the amounts of the various soil separates in a soil sample, usually by sedimentation, sieving, micrometry, or combination of these methods. RBD – relative bulk density, the field bulk density divided by MBD of the Proctor test. RHGI – relative height growth index, the height growth of the undamaged trees on disturbed sites divided by the height growth of the same species at the same age estimated on undisturbed sites with similar ecological characteristics using the SiteTools software. R ratio – relaxation ratio, the bulk density of the test material under specified stress divided by the bulk density after the stress has been removed. SBS – Sub-Boreal Spruce BEC zone. SI – site index, a relative measure of forest site quality based on the height of a site tree at breast height age of 50, where a site tree is the largest diameter tree of the target species in a 0.01 ha plot and the breast height age is the number of annual growth rings at breast height (1.3 m). Site index information helps estimate future returns and land productivity for timber and wildlife. WMBD – critical water content, the water content at which MBD is obtained in the Proctor test.  xiii ACKNOWLEDGEMENTS This work has been supported by two FIA-FSP grants, and I would like to extend a sincere thank-you to the FIA-FSP for the generous logistical support. My most sincere thank you goes to my supervisor Dr. Maja Krzic: thank you for your unfailing support and encouragement throughout my Ph.D.; thank you for giving me guidance to pursue my interests, for teaching me how to write more effectively; thank you for being energetic when I was tired; and thank you for your contributions to this thesis, for pushing me to improve each and every chapter. Thank you for the many social activities we spent at your home and in the field. My deepest thank you goes to Dr. Chuck Bulmer: thank you for inspiring me with your passion for soil compaction and tree growth; thank you for providing me opportunities to experience BC forests; thank you for the time and efforts you have spent with me in the field and the lab, and at your home, discussing my research and improving my understanding of BC forest soil. Thank you for your contributions to this thesis. I cherish those times I have spent with you and your family each time I visited Vernon, BC. It has been a pleasure to work with other members of my supportive committee: Drs. Suzanne Simard and Margaret Schmidt. Thank you for your commitment and support; thank you for inspiring discussions during and outside of committee meetings; thank you for taking time and efforts reviewing each manuscript I prepared and providing me with constructive guidance. I am particularly grateful for your encouragement to pursue my research interests. My thesis would not have been possible without generous financial support. A huge thank-you goes out to NSERC, UBC Ph.D. Tuition Fee Award, UBC International Partial Tuition Scholarship, UBC TAship, and UBC Donal S. McPhee Fellowship. Thank you Dr. Les Lavkulich for your generosity and encouragement toward my research and academic improvement; thank you for your zeal about science, especially your knowledge in the field of Soil Science and BEYOND☺, which has been inspiring me to purse the better in my academic career. Thank you Dr. Sietan Chieng, without your kind introduction, I would not have met my supervisor and would not have started this meaningful and memorisable journey at this famous university in the beautiful Vancouver. I would like to acknowledge many people who helped collect MBD samples - I did not get a chance to meet most of you when I started this work but what you had done for the pilot  xiv study facilitated my following-up research. I also owe gratefulness to Dr. Shannon Berch and Dr. Graeme Hope for part of the tree growth data; George Franssen and Peter Staffeldt for the data collection in the field and logistic help on the road; Clive Dawson and Carol Dyck for the lab analysis. I have enjoyed my time being with my lovely colleagues: Stephanie, Leslie, Peter, Andrea, Shannon, Mary, Brian, Sarah, Melissa, Rachel, Phraba, and Christian. Thank you folks! Thank you Will for the help in the MBD and the Atterberg Limits tests. My friends and family have been incredibly supportive. Thanks to my parents for their unfailing love and encouragement and to my sisters and brother for their continuous care about my life in Canada. In addition to those already listed elsewhere, special thanks goes out to Alice, Harry, and Feng for the many social activities and trips; Art Bomke and Mike Novak for the Graduate Seminars I enjoyed quite a lot; and Martin Hilmer for keeping our SOIL 200 teaching lab always neat and tidy. Thanks, everyone, for helping me keep life fun. And thanks to all the Soil People in the McMillan Building for the mutual support and fun times. Last, but definitely not least, I owe the world to wife Lin Lin and daughter Zihan (Maria) Zhao. Lin and Maria, thanks for your ever-lasting love and encouragement, for keeping the joy in my life, for sticking with me through the ups and downs during this journey, and for making my every struggling worthwhile and my life meaningfully balanced. I will be forever grateful.  xv CO-AUTHORSHIP STATEMENT Chapters 2 – 4 have been prepared as stand-alone, peer-reviewed publications. Chapter 3 has been published in the Soil Science Society America Journal, chapter 4 has been submitted to a peer-reviewed journal, and chapter 2 will be submitted to a peer-reviewed journal for publication. I am the senior author on all papers. I took primary responsibility for the design, implementation, analysis, and writing of all co-authored chapters. Details of co-authorship contributions for Chapters 2, 3, and 4 are outlined below. Chapter 2: Dr. Chuck Bulmer contributed to the design of the gas pycnometer, helped in soil sampling and testing, as well as pycnometer calibration. Dr. Maja Krzic provided multiple comments on the manuscript and contributed to the manuscript writing. Chapter 3: My co-authors Dr. Maja Krzic, Dr. Chuck Bulmer, and Dr. Margaret Schmidt developed the original idea of relating soil physical and chemical properties to the maximum bulk density (MBD). I embraced the opportunity to include more soil physical and chemical properties, developed four sets of models with improved accuracy in predicting the MBD. Dr. Maja Krzic instructed on proposal development, helped with field sampling and laboratory tests, and commented on multiple versions of the manuscript. Dr. Chuck Bulmer helped in sampling and soil tests, the development of the research proposal, and contributed to the manuscript writing. Dr. Margaret Schmidt helped in the development of the research proposal and commented on the manuscript writing. Chapter 4: Drs. Maja Krzic and Chuck Bulmer contributed to the initial idea of this chapter through stimulating discussions and reviewed numerous drafts of the manuscript, and provided guidance with methodology development. Dr. Chuck Bulmer also contributed with soil sample collection and providing access to archived tree growth data. Dr. Suzanne Simard provided guidance on relative height growth and edits of the manuscript. Dr. Margaret Schmidt was involved in the initial concept of this chapter through her work on the maximum bulk density, provided insight on factors affecting tree growth in a variety of ecosystems and provided comments on the manuscript.  1 1 INTRODUCTION 1.1 Compaction of forest soils Forest soils are prone to degradation caused by the application of heavy machinery during timber harvesting and mechanical site preparation. The soil degradation can include compaction (Greacen and Sands, 1980; Froehlich and McNabb, 1984), soil erosion (Miles et al, 1984; Carr 1987), nutrient displacement (Ballard and Hawkes, 1989), unsuitable moisture, thermal, and aeration regimes (Standish et al. 1988; Sutton, 1991; Day and Bassuk, 1994), and unbalanced nutrient cycles (Dick et al. 1988). Soil compaction is a process of soil densification caused by the application of stresses, usually of short duration. This process is associated with changes of several soil properties, such as bulk density, porosity, water retention, soil strength, and aggregate stability. Compaction has been recognized as a major factor affecting forest production (Mckee et al. 1985; Firth and Murphy, 1989). Froehlich (1973) reviewed compaction in western USA forest and concluded that 50% of the harvested area is disturbed and 25% can be considered as compacted under tractor logging. In British Columbia (BC), over 20% of the harvested area has been compacted based on the estimation by Utzig and Walmsley (1988). The widespread use of heavy machinery in BC’s plantations has led to the concern that compaction is reducing long-term soil productivity (Holcomb, 1996). In BC 25 million ha forest (out of a total of 59 million ha forest) are timber production forests (Ministry of Forests and Range 2007). Less than 1% of the timber production forest is harvested annually (e.g., around 170,000 ha were harvested in 2006). In Canada, about one million ha of forest is harvested annually (Natural Resources Canada 2008), which brings about 0.25 million ha of harvested area annually under the question of compaction and the consideration of rehabilitation. Even though the soil conservation provisions of the Forest Practices Code of British  2 Columbia required that the amount of dispersed soil disturbance in the net area to be reforested is limited to < 10% for non-sensitive soils, and < 5% for sensitive soils (Bulmer 1998), soil compaction is still a problem in BC, especially for those permanent and temporary access structures (i.e., landings and roads) that are considered to be severely compacted. For example, in BC permanent access comprises up to 7% of the harvested area and in certain cases temporary roads and landings occupy more. The potential reduction of productivity resulting from soil degradation due to forestry practices in BC between 1976 and 1986 was estimated at about 80 million-dollar losses per year (Utzig and Walmsley 1988). The situation of soil compaction caused by logging is severe in developing countries. For example, in Brazil, illegal logging accounted for 43% of the total wood consumption (Gutierrez-Velez and MacDicken 2008), the annual deforestation rate in the Brazilian Amazon was 2.38 million ha in 2003 (Fearnside 2003), and soil compaction is among the most obvious impacts of deforestation. Appraisal of soil susceptibility to compaction and degree of compaction would be necessary to establish the likely effects of forestry operations on soil compaction and tree growth in BC and other regions of the world. 1.2 Impacts of soil compaction on tree growth Generally, soil compaction leads to concentration of roots in the top layer (e.g., 0-20 cm) and decreased rooting depth and root elongation (Veihmeyer and Hendrickson 1948; Taylor and Gardner 1963; Willatt 1986; Whalley et al. 1995). Due to the loss of most macro-pores, roots growing in severely compacted soils are also characterized by greater diameter and tortuous growth (Lipiec et al. 1991). Decreased tree growth or increased mortality associated with root deformation has been reported for several species (Rudolph 1939; Halter and Chanway 1993; Harrington and Howell 1998). Reduction in seedling height growth on compacted soil throughout the U.S. was ranging from 5 to 50% (Adams and Froehlich 1981), while survival  3 could decrease by 83% (Andrus and Froehlich 1983). In BC, Utzig and Walmsley (1988) estimated a 50% reduction in volume (m3/ha) that is associated with soil degradation. Interior Douglas-fir (Pseudotsuga menziesii var. glauca (Beissn.) Franco), lodgepole pine (Pinus contorta Loudon var. latifolia Engelm.), and interior spruce (including Picea glauca (Moench) Voss, P. engelmannii Parry ex. Engelm., P. glauca engelmannii) are the most important commercial species in the interior of BC (Nigh et al. 2004), in 2000, they comprised 73% of the total volume harvested and about 90% of the total number of seedlings planted (Ministry of Forests 2001). Lodgepole pine root systems do not develop with a definite symmetry (Preston 1942). Although the root habit varies considerably with soil type and conditions (Lotan and Critchfield 1990), root deformity in planted trees is 44% higher than in naturally regenerated trees (Robert and Lindgren 2006). Root growth is particularly critical during the first year: a root growth of 12.7 - 15.2 cm was observed for seedlings growing on prepared seedbeds in Montana and Idaho, while that of only 9.6 cm was observed on scarified, un-shaded seedbeds in the central Rocky Mountains (Lotan 1964). In compacted soils, lodgepole pine developed a dual root system, with the lateral roots growing largely within the top 30-60 cm and vertical roots extending to the rock layer at a depth of 100-120 cm (Berndt and Gibbons 1958; Bishop 1962). On blade-scarified sites with forest floor incorporated into the mineral soil, lodgepole pine develops a root system larger than spruce by the second growing season, and this trend continued during the rest of the five-year long experiment (McMinn 1978). It has been reported that growth of both lodgepole pine and hybrid spruce were less sensitive to soil disturbance than other species such as Douglas-fir (Smith and Wass 1991; Page- Dumroese et al. 1998). Hybrid spruce is a shallow-rooted species that forms over 87% of its root mass in the top 15 cm of soil (including both forest floor and an A horizon; Safford and Bell 1972; Kimmins and Hawkes 1978; Strong and La Roi 1983), and it can develop different forms  4 of taproots and layered roots under soil conditions that vary in soil density, texture, aeration, and water table. For example, lateral roots extending 3 to 7 m were observed in spruce growing in the Humic Gleysols and Terric Organic soils in the boreal forest of Canada (Wagg 1967; Strong and La Roi 1983), while lateral roots extending to over 10 m were observed in spruce growing in silt loam soil in Europe (Nadezhdina et al. 2006). Douglas-fir is a deep-rooting species and thrives on well-aerated, deep soils. In the absence of obstructions, it develops a radially symmetric root system: initially a taproot forms and grows rapidly during the first few years, while the main lateral roots develop during the first or second growing season. In deep soils (69 to 135 cm), the taproot could grow to about 50% of its final depth in three to five years, and to 90% in six to eight years (Hermann 1977). In compacted soils, however, Douglas-fir would form a dense, shallow rooting system, with a limited number of rope-like lateral roots (Kuiper and Coutts 1992). This root distribution pattern observed in compacted soils was partly attributed to the horizontal orientation of pores (Slowinska-Jurkiewicz and Domzal 1991). Heilman (1981) showed that height growth of 45-day old Douglas-fir seedlings was not affected by compaction, but found that root penetration declined linearly with increasing soil bulk density. The existence of long lateral roots allowed the seedlings to access more favorable soil conditions, which explained the lack of relationship between height growth and compaction. Since certain plant nutrients (e.g., P, K, Ca, Cu) have limited mobility in soils, reduced root growth due to compaction will further restrict the uptake of these nutrients by plants. As a result, confined root elongation and limited nutrient availability are often associated with reduced seedling survival, height and root growth, and net photosynthesis, and increased rate of shoot respiration (Steve and Thomas 1981; Corns 1988; Conlin and van den Driessche 1996; Stone and Kabzems 2002; Bulmer and Simpson 2005; Kabzems and Haeussler 2005; Fleming et al. 2006).  5 1.3 Soil properties commonly used to characterize soil compaction Soil properties commonly used to characterize soil compaction include bulk density, total porosity, air-filled porosity, and soil strength. However, these properties have limited usefulness for comparing compaction among soils that vary in mineralogy, organic matter content, texture, and in situ soil conditions. Soil bulk density is the mass of dry soil per unit of bulk volume, including the air space. Soil bulk density can be employed to indicate soil compaction within a given soil type (Schoenholtz et al. 2000), but it varies among soils of different textures, structures, organic matter content, and mineralogy. Although bulk density is the most commonly used soil property to indicate soil compaction, it is difficult to compare the management impacts on soil compaction among different soils using bulk density, and determining a critical value for bulk density impeding plant growth is challenging (Bulmer 1998). Various studies have reported different critical limits depending on the soil type and plant species tested. For example, Heilman (1981) set the critical root growth-limiting bulk density level at 1.7–1.8 Mg/m3 for Douglas-fir seedlings growing in sandy loam to loam textured soils. On the other hand, Heninger et al. (2002) reported that sandy loams with an artificially created bulk density of 1.59 Mg/m3 stopped root penetration of potted two-year-old Douglas-fir. Daddow and Warrington (1983) reported that root growth of Pitch pine, Austrian pine, and Norway spruce was limited when the average bulk density was around 1.40 Mg/m3 for silt loam and 1.60- 1.80 Mg/m3  for sandy loam. Lousier (1990) considered bulk density values of 1.2-1.4 Mg/m3 to be growth limiting for most ecosystems. Soil pore geometry describes the volume of the various sizes of pores in a soil. Soil pore geometry is subject to temporal variation by the influences such as drying and wetting, freezing and thawing, therefore it usually reflects the soil condition at the time of sampling rather than  6 some final or permanent condition (Carter 1990), thus making it difficult to use soil porosity to indicate soil compaction even for the same soil type. Mechanical resistance measures soil strength or resistance to deformation, and it could be used to approximate the resistance encountered by plant roots. However, soil mechanical resistance is strongly affected by soil water content and the nature of that relationship varies with soil texture. Operating conditions (e.g., speed of insertion) and instrument set-up (e.g., cone size and shape) also affect measurements of soil mechanical resistance (Greacen and Sands 1980). Interpretation of these data depends on the site characteristics, like soil texture, as well as water content at the time of measurement and penetrometer specifications. All of these lead to question the usefulness of mechanical resistance in describing soil compaction and linking it to plant growth (Bulmer 1998). Previous field-based attempts to quantify the relationship between soil physical properties (bulk density, porosity, and soil mechanical resistance) and tree growth at the site level have had mixed success. Bulk density is only weakly correlated with tree growth (Froehlich and McNabb, 1984; Miller et al. 1996; Krzic et al. 2003), and foresters are mainly concerned with the critical value above which tree growth is reduced. There is a need to establish a parameter that not only accurately indicates soil compaction but also provides a link between the degree of soil compaction and tree growth. 1.4 Methods used for the assessment of soil resistance to compaction A wide variety of methods have been used to measure soil compaction resistance or soil strength, which could be grouped into indirect (e.g., water stability test, Atterberg test) and direct methods (e.g., confined compression test, triaxial test, direct shear test) (Horn and Lebert 1994). The most frequently used methods include the standard Proctor test and uniaxial unconfined compression test.  7 1.4.1 Proctor compaction test The Proctor test, standardized by ASTM (2000), has been employed on agricultural, rangeland, and forest soils. Soils are first sieved to 4.75 mm to eliminate coarse particles, and then are placed into a standard mould (943 to 2121 cm3 volume; 10.2 or 15.2 cm diameter; 11.6 cm height). A 24.4-N (2.5-kg) rammer is then dropped from a height of 30.5 cm to produce a total of 600 kN-m/m3 of compactive effort, and then bulk density is determined. For the same soil type, bulk density obtained after the compactive effort varies with soil water content. The water content at which bulk density reaches the highest value (i.e., maximum bulk density - MBD) is referred to as the critical water content - WMBD (Fig. 1.1). Maximum bulk density determined by the standard Proctor test has been defined as the highest compaction value encountered in the field for a particular soil (Stengel et al. 1984). Both MBD and WMBD are parameters traditionally used in soil engineering to assess soil susceptibility to compaction. 1.4.2 Uniaxial compression test A static loading, uniaxial compression test has been considered appropriate to test compaction resistance of agricultural soils, since in the field, loading by heavy machinery is mainly of a static type, though the loading time is short, and impact and vibration or kneading also may occur (Hakansson 1990). The volume of the oedometer cylinder used in the uniaxial test varies from 126 to 11540 cm3 (Soane 1990), with the largest cylinder (11540 cm3, inner diameter 35 cm, depth 12 cm) commonly used in Europe to carry out a soil compression tests (Hakansson 1990). The disturbed soil samples are sieved through 2.5 cm mesh and thoroughly wetted by spraying with water and mixing, then stored for at least one week for swelling. In the test, the oedometer is filled with loose, wet soil to the rim and is subjected to 200 kPa pressure for one week. The volume of the sample is recorded 15 minutes after pressure removal, water content is determined, and the bulk density is calculated. Hakansson (1990) regarded 200 kPa as the standard pressure in testing European soil compressibility since it imitated pressure applied  8 by machinery during tillage or trafficking. This test is called the uniaxial compression test because the loading is uniaxial and the apparent movement of particles during the test is in one direction (i.e., vertical movement) (Koolen 1987). Both MBD from the Proctor test and the bulk density from the uniaxial test have been regarded as the maximum values of soil compaction encountered in the field as a result of traffic by heavy machinery (Hakansson 1990; Stengel et al. 1984). Therefore, these two parameters have been considered the reference bulk density to determine how far the field bulk density is from such a value. The Proctor test is an indirect compaction test for homogenized soil samples, while the uniaxial unconfined compression test is an indirect consolidation test for homogenized or structured soil samples. Since activities such as timber harvesting and site preparation commonly take place when the site is dry or moist, the Proctor test is more representative of such management activities. Consequently, the reference bulk density in forest ecosystems is commonly determined using the Proctor method. 1.5 Soil properties that affect soil susceptibility to compaction Soil strength depends on the applied stress that reflects soil cohesiveness - the bonding of the soil particles, and the angle of internal friction - the resistance when soil is forced to slide over soil (Hillel 1998). Both cohesiveness and the angle of internal friction are controlled by soil physical and chemical properties, such as soil organic matter content, texture, structure, and surface charge. 1.5.1 Soil organic matter Variations in MBD have been widely attributed to changes in soil organic matter (Saini, 1966; Adams, 1973; Howard et al. 1981). Not only the quantity of organic matter but also its quality has been identified to affect susceptibility to soil compaction (Soane, 1990). The relationship between MBD and organic matter content was expressed as:  9 OMbaMBD ×−= where a is the MBD for the same soil without organic matter, b is the effectiveness of organic matter in reducing soil compressibility, and OM is the organic matter content (Zhang et al. 1997). When comparing a silt loam soil at two organic matter contents (2.8 and 4.1 %, w/w), Free et al. (1947) found that MBD was higher for the low organic matter content. The soil reached the MBD at higher WMBD when organic matter content was also high. Increased organic matter content in sandy soils under radiata pine forests in South Australia was associated with reduced compaction under a given load (Greacen and Sands 1980). In a study carried out by Ball et al. (1988) on Gleysol and Cambisol soils from Scotland, a reduction in MBD of 0.175 Mg m-3 per increase of 1% organic carbon was observed. Felton and Ali (1992) mixed animal manure with soils and immediately performed the Proctor test. They found reductions in MBD ranging from 0.056 to 0.077 Mg m-3 per 1% increase in organic carbon. Significant relationships between MBD and organic matter have been observed in several studies, and are summarized in Table 1. A strong curve linear relationship exists between compressibility parameters and soil organic matter content. For example, for 35 South African forest soils ranging from 80 to 660 g kg-1 clay and from 2.6 to 57.7 g kg-1 organic carbon, Smith et al. (1997b) obtained a highly significant quadratic relationship (R2=0.65) between compression index (C) and organic matter content (obtained by LOI, loss-on-ignition, the loss in mass in a soil after ignition at 450 oC). The C values were highest when LOI values were between 5 and 9%. Incorporation of the same amount of un-decomposed and fully decomposed organic matter has very different effects on soil compressibility (Soane 1990). Guerif (1979) found that from 100 to 1000 kPa stress, the R ratio (the bulk density of the test material under specified stress divided by the bulk density after the stress has been removed) of un-decayed organic  10 matter-soil mixture was nearly twice as that decayed organic matter-soil mixture. The difference in the R ratio was due to higher elasticity of the un-decayed straw under compression forces than those of decayed particles. Since quality of soil organic matter has a profound effect on compressibility, parameters such as readily oxidisable organic matter (i.e., the loss-on pretreatment with H2O2) (Day 1965; Ball et al. 2000) and organic carbon (Donkin 1991; Smith et al. 1997b) have been used to determine soil compaction susceptibility. Soane et al. (1972) found that the oxidisable organic matter fraction was better correlated with soil compressibility (R2=0.66) than was total organic matter content (R2=0.52). In studying Gleysols and Eutric Cambisols in the UK, Ball et al. (1996) showed that soil carbohydrate content was more significantly related (P < 0.001) to bulk density than soil organic carbon content (P < 0.01). Carbohydrate decreased the compressibility and increased the plastic limits of the soil. Organic matter content may have different effects on MBD versus C. Guerif (1979) reported that addition of 20 g kg-1 organic matter to 150 g kg-1 clay had a considerable effect on C at any given water content, and addition of wheat straw increased the void ratio, while C was not changed. Gupta et al. (1985) reported that varying amounts of maize residues (0 – 34 g kg-1 ) added to clay loam, silt loam, and sand, had almost no effect on C, but bulk density was lowest (0.9 Mg m-3 ) at 100 kPa stress when the highest amount of residue (34 g kg-1) was applied. On the other hand, Smith et al. (1997b) found that compressibility was significantly correlated (R2=0.648) with organic matter content (loss-on-ignition) for 35 South African forest soils. In addition to the effect of its content on MBD, the dimension of organic matter particles also influences soil compressibility. For example, when maximum compaction was set to 49 MPa, increasing the ratio of length to diameter (ranged from 0 -1) of straw increased soil compressibility (O’Dogherty and Wheeler 1984). 1.5.2 Particle size distribution Particle size distribution (PSD) also affects MBD. Larson et al. (1980) investigated  11 relationships between C to clay content of 36 agricultural soils collected worldwide belonging to eight soil orders (Mollisols, Inceptisols, Spodosols, Vertisols, Entisols, Alfisols, Ultisols, and Oxisols). The authors found that C was more strongly related to clay content (R2=0.635) when soils were less weathered. Maximum C (0.55) was observed when clay content was 450 g kg-1 and the C decreased with increasing clay content over that value. For highly weathered soils (dominated by kaolinite or iron oxides in the clay fraction), a stronger relationship between C and clay content (R2=0.902) was obtained; the maximum C was 0.50 when clay content was 450 g kg-1. Using 42 expanding and nonexpanding soil samples that Larson et al. (1980) had collected, Gupta et al. (1985) found that C increased with increasing clay content to 330 g kg-1, and the C curve levelled off when clay content was > 330 g kg-1. The change in C with clay content may be ascribed to the soil tensile strength with the presence of clay. For 17 soils from France (Vertisol, Gleysol, Fluvisol, and Cambisol) with a clay content from 57 - 519 g kg-1, the overall dry tensile strength was linearly related to clay content (Guerif 1990). Using clay and silt content together may improve predictability of soil compressibility as compared to using individual soil particle classes. For example, the MBD for 20 glaciofluvial soils in Finland was predicted by clay and silt content (Heinonen 1977), as: MBD = 1.42 – 0.0016 %Clay + 0.0021 %Silt (R=0.79). For the compression feature of 35 Oxisol, Ultisol, Alfisol, and Entisol forest soils from South Africa, clay plus silt, clay, coarse silt, fine silt, medium sand and fine sand were each significantly correlated with MBD and C, and clay plus silt was better related to both C and MBD (R2=0.914) than was clay alone (R2=0.862) (Smith et al. 1997a, 1997b). The effects of sand content on compressibility were also studied. For example, it was reported that sandy loams and loamy sands with high fine sand fractions were highly susceptible to compaction (Milford et al. 1961; Bodman and Constantin 1965; Bennie and Krynauw 1985). The MBD was best explained by the percentage of very coarse sand, silt, and clay for 40 coarse-  12 to-medium textured soils from South Africa and 33 artificial mixtures (Van der Watt 1969), while MBD was best related to very coarse sand for 13 Australian coarse-to-medium textured sand and loamy sand soils (Henderson et al. 1988). Maximum bulk density increased with increasing content of very coarse sand and clay, the reason is that soils were broadly graded with the high proportions of very coarse sand or clay, which gave higher MBD values than soils were narrowly graded (Henderson et al. 1988). Peakedness (kurtosis) and symmetry (skewness) of particle size are sometime used to describe PSD (Webster 1990; Shirazi and Boersma 1984). These two parameters are especially useful when the specific surface area of particles is low and their surface charge density is weak, because soil structure or aggregate stability will not be the main factor influencing soil compressibility under such circumstances. For example, a well-graded sand (i.e., low kurtosis) is more compressible than a uniformly-graded (i.e., high kurtosis) sand at the same initial void ratio (Oda 1972). Moolman (1981) reported a linear relationship between MBD and kurtosis; where kurtosis explained 82% of the variation in the MBD. However, kurtosis is not always positively related to MBD, especially when particle sizes were not normally distributed (i.e., lack of one or two fractions of the particle size). For example, Smith et al. (1997b) showed that MBD decreased as the degree of kurtosis increased. In their study, the skewness varied (i.e., from -5.88 for sand to 6.96 for clay) far away from a value of zero under the normal distribution, and the high value of kurtosis under this condition indicated a high frequency of one class of particle instead of a well graded distribution of particles. Particle size distribution did not always have a significant effect on compressibility. For 93 forest soil samples with various soil textures from four biogeoclimatic zones in BC, MBD was weakly related to PSD (R2 < 0.03) (Krzic et al. 2004). The authors ascribed this poor relationship to the uneven representation of sand, silt, and clay, especially since the majority of the samples had low clay content.  13 1.5.3 Water content and water potential Soil water is an important factor affecting the degree of soil compaction. As Proctor (1933) has shown, under a certain compactive effort, soil bulk density increases with soil water content up to a maximum point, after which, it decreased. For 36 world agricultural soils with clay and silt or silt loam textures, soils became further compacted under the same stress as soil water content increased from 15.7, 18.7, 23.3, to 27.6% (v/v) for Aquic Hapludoll soils, and from 11.6, 17.6, 24, to 28.8% (v/v) for Eutrorthox soils (Larson et al. 1980). On the other hand, Jakobsen and Greacen (1985) reported that water content was of little importance for the compaction of sandy soils (90% sand content) below saturation. In assessing soil susceptibility to compaction, Gupta and Allmaras (1987) reported that, at soil water content below field capacity, different compacting loads had limited effect on volume reduction of a sandy loam soil. At soil water content above field capacity, however, soil volume reduction was more pronounced with increasing loading. The difference in water content also caused different compaction patterns (i.e., increase of bulk density) for the same soil subjected to high and low compacting loads. The effect of soil water content on MBD may vary among soils with different parent material. For example, high MBD values (1.65-2.0 Mg m-3) were reached at relatively low water content for aeolian sand, sandstone, and illite derived soils, while relatively low MBD values (1.2-1.5 Mg m-3) were reached at high water content for dolerite, diabase, and shale derived soils from South African forests (Smith et al. 1997b). The effects of soil water content on compaction may also vary among soils with different organic matter levels. For example, at low water content, the penetration resistance of compacted soils decreased with increasing peat content, while the penetration resistance increased with increasing peat content at high water content (Ohu et al. 1985). For aggregate samples containing varying organic matter contents that had pre-equilibrated at water suctions  14 of 1.9 and 4.2 m, the compression strength increased as organic matter content increased at low water suction but not at high water suction (Kuiper 1959). In a compressibility study of German silt loam, clay soil, and sandy soil, Zhang et al. (1997) found that the effectiveness of organic matter in reducing soil compressibility remained stable up to a water content close to the WMBD for silt loam and clay soil, then decreased with increasing water content. However, for sandy soil, such effectiveness increased with increasing water content and reached the maximum at the WMBD, then subsequently decreased with additional increase in water content. Water content also influences the effect of aggregate size on compressibility (Willatt 1987). For a Stagnogleyic Cambisol with sandy clay loam texture at 0-250 mm depth and clay loam texture at 250-350 mm depth, water content had a limited effect on compaction of 0-2 mm aggregates: only the wettest sample (0.296 w/w) showed a marked increase in compaction after the first compacting pressure of 35 kPa. A similar trend was found for the 0-10 mm aggregate sample composed of equal proportions of 0-2 mm, 2-5 mm, and 5-10 mm aggregates. For 2-5 and 5-10 mm aggregates, soil was further compacted at any given applied pressure with an increase of water content. However, the 2-5 and 5-10 mm aggregates were less readily compacted than smaller aggregates at the lowest water content. Maximum bulk density decreases with WMBD. Diâz-Zorita and Grosso (2000) reported a wide range of MBD values (1.10-1.72 Mg m-3) with corresponding WMBD ranging from 138 to 275 g kg-1. Krzic et al. (2004) found that MBD was strongly related to WMBD for 93 forest soils in BC. Aragon et al. (2000) studied 30 Argentinean soils and found that there was a strong negative correlation between MBD and WMBD, and their results were in agreement with other research conducted in that area (Table 1.2). Soil friction and cohesion were related to Atterberg limits in undisturbed soils (Baver 1930), and plastic and liquid limits have been considered important parameters in relations to  15 MBD (Kirby 1991; Thacker et al. 1994; Mapfuno and Chanasyk 1998). For the loam to clayey textured soils in Alberta, the plastic limit was within 3% of the critical water content and could be used to predict MBD (Thacker et al. 1994); while Ball et al. (2000) used liquid limit and loss- on-pretreatment organic matter to predict MBD of UK cultivated soils. Since water potential is related to pore size distribution, the degree of soil compaction can be reflected by the change in pore size distribution. Koolen (1974) reported a hyperbolic relationship between compaction pressure and total porosity, while Lipiec et al. (1991) reported that relative bulk density (RBD) was related to air-filled porosity. Macroporosity is generally considered to be more sensitive to compaction than total porosity because macropores are responsible for aeration and drainage (da Silva and Kay 1997), and are destroyed the most by compaction (Douglas 1986; Carter 1990; Richard et al. 2001). For example, the degree of compaction was significantly related to macroporosity (R2=0.85) of a fine sandy loam Orthic Humo-Ferric Podzol at Prince Edward Island (Cater 1990). 1.5.4 Mineral composition The type of soil minerals present in a particular soil has also been reported to influence soil susceptibility to compaction (Soane et al. 1972). Soils with predominantly 2:1 type clay minerals had a higher maximum C value (i.e., 0.55) than soils with predominantly kaolinite or Fe oxides (i.e., C of 0.50) in the clay fraction (Larson et al. 1980). Similar results were reported by Smith et al. (1997) for South African forest soils: soils derived from base-rich parent materials such as dolerite, diabase, and shale had a low MBD, while those derived from sandstone, granites, and illite had a high MBD (Smith et al. 1997b). In addition to organic matter and clay, other cementing agents that enhance aggregate formation, including Fe-, Al-, Si-oxides, or carbonates also affect susceptibility to compaction. Oxides have the potential to stabilize micro- and macroaggregates through electrostatic interactions between positive charges associated with oxides and negative charges of crystalline  16 clay minerals (Pinheiro-Dick and Schwertmann 1996). It has been reported that oxides of Fe and Al were the main cementing agents that enhanced aggregate stability in BC soils (McKeague and Sprout 1975; McKeague and Wang 1980). The role of Fe- and Al-oxides as the main soil cementing agents has been reported in other places. For example, Fe-, Al-oxides, Ca-carbonate, and interstratified minerals increased the stability of aggregates for Italian soils with soil textures ranging from clay, clay loam, and sandy loam to sandy clay loam (Nwadialo and Mbagwu 1991). Fe-, Al-, and Mn-oxides were the major factors that contributed to the formation and stability of microaggregates for five floodplain soils from the Niger River in Nigeria (Igwe and Stahr 2004). Aggregate stability is related to the ability of the bulk soil to resist applied compressive forces; this is because aggregation increases the number of contact points per unit volume of bulk soil, which increases the shear resistance (Dexter, 1975). Consequently, well-aggregated soils are stronger than structureless materials (Horn and Lebert 1994). An inverse linear relationship (R2=0.94) between MBD and the aggregate stability index (wet sieving) of the disturbed Gleysols and Cambisols was reported by Ball et al. (1988). Soil MBD was well predicted by including oxides (like Fe) in equations with organic C, liquid limit, and total sand (R2=0.99) for 14 forest soils from California ranging from sand, clay loam, sandy loam, loamy sand, to loam texture (Howard et al. 1981). For three Indiana calcareous soils, soil strength was positively related to extractable Si- and CaCO3-equivalent but negatively correlated with clay and extractable Fe and Al, (McBurnett and Franzmeier 1997). Soils in BC are developed from many different rock types that contain various forms and proportions of minerals (e.g., Fe-, Al-oxides and hydroxides, Ca- and Mg-carbonates) (Valentine et al. 1978). A high content of Fe and Mn oxides, which are heavy minerals with a particle density ranging from 3.0 to 5.5 Mg m-3, increases soil particle density. Displacement of mineral soil particles by erosion or tillage can lead to variations in particle density, too (Ball  17 2000). On the other hand, particle density of soil organic matter is around 1.3 - 1.5 Mg m-3 (Caldwell et al. 2007; Redding et al. 2005). Soil particle density may deviate considerably from 2.65 Mg m-3 for soils with high organic matter content. When using MBD to describe soil compaction, it is worth considering the influence of particle density on MBD. The Proctor test not only requires a large sample, but it is also time-consuming and expensive to conduct (Horn and Lebert 1994). Consequently, efforts have been made to predict MBD, a parameter of soil compaction susceptibility, by using other soil properties (e.g., organic matter content, PSD, liquid limit). The preferable soil properties in this prediction process are those that are simpler and cheaper to determine than MBD, can be carried out on small soil samples, and/or are readily obtained from existing documents such as soil surveys. 1.6 High-level integration indicators of soil compaction Several soil properties have been used to characterize the state of soil compaction, such as bulk density, air-filled porosity at a certain matric water tension, soil strength or penetration resistance. Because compaction is a process characterized by volume change, volumetric properties such as bulk density or porosity are frequently used to indicate soil compaction (Hakansson 1990). However, critical limiting values for bulk density have not been defined for the wide range of soil conditions typical in forests, primarily because such limiting values are different for soils with varying texture, organic matter, and other properties. A bulk density indicating an extremely compact state in one soil may imply a very loose state in another soil (Hakansoon 1990). Establishment of limiting values characterizing soil compaction would be beneficial for soil scientists, foresters, and land managers. Efforts have been made to identify high-level soil parameters, other than bulk density, that can integrate several soil properties and relate them to plant growth. One of these parameters was the least limiting water range (LLWR) introduced by da Silva et al. (1994). The  18 LLWR describes the range of soil water contents where water availability, soil mechanical resistance, and air-filled porosity do not exceed assigned values associated with growth limitations. The use of LLWR to evaluate soil conditions in forest productivity studies would be advantageous because it focuses on soil conditions that have been shown to affect plant growth. In addition, LLWR can be used in conjunction with soil water content measurements, which are relatively easy to obtain for detailed studies, but practical limitations also arise. For example, determining the soil water retention curve can be costly and time consuming: in some soil types, such as fine textured soils, it may take as long as four weeks to equilibrate the sample to the permanent wilting point (Zou et al. 2000). Because changes in LLWR for a particular soil type are driven in large part by changes in soil compaction (da Silva et al. 1994), measuring the LLWR may not always be necessary to determine conditions affecting plant growth in compacted soils. Another approach for evaluating the state of soil compaction involves expressing the actual bulk density as a percentage of some reference compaction state (Lipiec et al. 1991, Topp et al. 1997; Lipiec and Hatano 2003). For example, Joosse and McBride (2003) proposed comparisons based on the void ratio to evaluate soil quality of agricultural sites. Such comparisons would allow conditions from a wide range of soil types to be evaluated using a single threshold limit, much as the critical limits of soil mechanical resistance and air-filled porosity appear to be relatively independent of soil type (Hakansson and Lipiec 2000; Zou et al. 2001). Therefore, use of a reference state could potentially enhance interpretations in soil compaction studies. Various parameters for a reference compaction state have been proposed (Carter 1990; da Silva et al. 1994; Hakansson and Lipiec 2000). The MBD determined by the standard Proctor compaction test (ASTM 2000) is rigorously defined, readily determined with standard test equipment, and has been used in several studies (Carter 1990; Smith et al. 1997b; Aragon et al.  19 2000). The potential advantages of using MBD as a reference compaction state can only be realized if the soil samples reliably represent site conditions, and this can create challenges in forest soils. Unlike agricultural soils, where soil type is often relatively consistent within a particular field, the properties of forest soils are known to vary widely across short distances (Courtin et al. 1983) due to the variable topography and the absence of tillage to mix and homogenize surface layers. Such variation would require a large number of samples for determining MBD, and therefore some alternative method to predict MBD would be beneficial. Relative bulk density (RBD), which is the field bulk density divided by MBD, and the degree of compactness (D), which is the field bulk density divided by a reference bulk density that is tested by the uniaxial compression test with a stress of 200 kPa, are suggested as high- level integrating parameters to describe compressibility of agricultural soils (Topp et al. 1997; Hakansson and Lipiec, 2000). Relative bulk density and D provided useful indexes to assess changes in soil bulk density and have been proven to be biologically meaningful for annually disturbed agricultural soils (Carter 1990; Hakansson and Lipiec 2000). For example, RBD of a fine sandy loam Orthic Humo-Ferric Podzol at Prince Edward Island was related (R2=0.69) to the relative grain yield of cereals (spring barley and spring wheat). A RBD range of 77 - 84% was associated with a relative grain yield ≥ 95%, while a relative compaction range of 84-89% corresponded to a macro-pore volume of 13.5 - 10.0%, indicating the need for regular soil loosening to maintain optimum soil aeration (Carter 1990). The degree of compactness was correlated with spring barley yield over a wide range of soil types in Sweden with a clay content between 2 - 60% and an organic matter content between 1 - 11% (Hakansson 1990). The author found that the optimal degree of compactness (Dopt, corresponding to highest grain yield) was independent of soil PSD and organic matter content (R2=0.00) and that it was consistently at 0.87. Since the reference BD obtained by the uniaxial test in the study of Hakansson (1990) was 7 - 17% lower than that obtained by the  20 Proctor test in the study of Carter (1990), the Dopt of 0.87 corresponded to an optimal RBD of 0.74 - 0.81. 1.7 Study objectives The RBD has been used in several studies to relate soil compaction to growth of annual plant species, but its usefulness has not yet been tested for assessment of tree growth in forest ecosystems. Development of a high-level integrating parameter of soil compaction that can be related to tree growth will be helpful to guide operational practices associated with soil disturbance on areas affected by forest management, and to assess the viability of rehabilitation to restore productivity to degraded areas. This thesis develops a method to test the variation in particle density of BC forest soils, introduces models to predict the reference bulk density (i.e., MBD), and examines the usefulness of RBD in relation to tree growth. Variation in particle density among sites makes it necessary to develop a new indicator (i.e., RBD) of forest soil compaction rather than bulk density in relation to soil productivity. The overall goal of this thesis project was to develop a high-level integration indicator that will characterize compaction of forest soils and that could be correlated to tree height growth. The specific objectives and hypotheses were: Objective 1: Assess the accuracy of the volume test for the constant-volume gas pycnometer that we have developed for small sample sizes (< 36 cm3). Objective 2: Examine the range of variation of particle density for forest soils in interior BC. Objective 3: Evaluate the relationship between particle density and selected soil properties (i.e., soil organic matter - SOM, oxides, and PSD). Hypotheses tested were: constant-volume gas pycnometry was a precise and quick method to test soil particle density; variation in particle density was explained by SOM, oxides,  21 and PSD; and mineral particle density of forest soils in interior BC varied from 2.65 Mg m-3. In this study, a new method was developed to analyze the accuracy of a custom constant-volume gas pycnometer. The evaluation of relationships between the particle density and other soil properties revealed the main sources that contributed to the variation in particle density. Objectives 1 - 3 are addressed in Chapter 2. Objective 4: Evaluate the relationships between MBD and WMBD, determined by the standard Proctor test, and other soil properties for a wide range of forest soils in BC. Objective 5: Identify the soil properties most important for predicting MBD. Objective 6: Develop a method for using MBD as a reference bulk density in forest soil compaction studies. The hypothesis tested was: soil MBD and WMBD could be predicted by soil physical and chemical properties. The models derived during this study were based on soil properties (e.g., liquid limit, total C, PSD) that are commonly determined in forest productivity studies to predict the reference bulk density. The models make it possible to avoid the Proctor test, which is labour intensive and time consuming to conduct. This is particularly useful when a large number of soil samples are collected. Objectives 4 - 6 are addressed in Chapter 3. Objective 7: Determine RBD for soils on heavily disturbed timber-growing sites such as forest landings and roads. Objective 8: Assess the relationship between RBD and tree height growth. Objective 9: Evaluate the influence of the thickness of surface organic materials on tree height growth. Hypotheses tested were: Presence of organic matter could mitigate the severity of soil compaction on tree growth; optimal tree height growth was expected within a certain RBD range; and there existed an RBD threshold limiting tree height growth, which warranted the  22 necessity of soil decompaction or rehabilitation. In the study of RBD and its relationship to tree growth, the relative height growth index (RHGI) was used to eliminate variation in growth caused by differences in climate and other site conditions among study sites. This approach allows the determination of soil compaction effects on tree growth. Objectives 7 - 9 are addressed in Chapter 4. Chapter 2 focused on the determination of soil particle density of forest soils in interior BC and found a significant variation in soil particle density among study sites, which indicated that mass-related soil compaction indicators such as bulk density will vary significantly with particle density. Consequently, the same bulk density of two soils does not necessarily imply the same degree of compaction when particle densities of these two soils are different. This study also determined that there was a strong correlation between soil particle density and SOM, oxides, and PSD. Hence, soil particle density was not selected from the multiple regression analysis carried out to predict MBD in Chapter 3. Soil properties such as Atterberg limits, SOM, oxides, and PSD that were important for predicting MBD were identified in this chapter. Development of the MBD prediction models made it possible to avoid labour-intensive and time-consuming Proctor test during the next part of this thesis research (addressed in Chapter 4), allowing me to use MBD as a reference bulk density in the determination of RBD. Relative bulk density, calculated as quotient of field bulk density to the MBD, was chosen since it is not affected by soil mass and particle density. In addition, RBD has been successfully used in indicating soil compaction and plant yield in agriculture, but its usefulness has not been tested in forest soils; in Chapter 4, I selected three economically important tree species (i.e., lodgepole pine, hybrid white spruce, and Douglas-fir) in BC and explored the usefulness of RBD in relation to the tree height growth.  23 Table 1.1 Relationship between maximum bulk density (MBD, Mg m-3) and soil organic matter reported in the literature. Equation* R2 (n) Source MBD = 1.845 - 0.0138 OCC 0.75 (36) Thomas et al. 1996 MBD = 1.86 - 0.0055 OM 0.52 (58) Soane 1975 MBD = 1.619 - 0.0065 OCC 0.66 (30) Aragon et al. 2000 MBD-1 = 0.611+0.0032 OCC 0.70 (30) Aragon et al. 2000 MBD = 1.74 - 0.0015 TOC 0.75 (26) Diaz and Grosso 2000 MBD = 1.71 - 0.0011 TC 0.70 (93) Krzic et al. 2004 *OCC, organic carbon content; OM, readily oxidisable organic matter; TOC, total organic matter; and TC, total carbon (in g kg-1).  24 Table 1.2 Relationships between maximum bulk density (MBD, Mg m-3) and critical water content at which MBD was reached (WMBD, 100 kg kg-1) as reported in the literature. Model R2 (n) Sources MBD = 2.152 - 0.027 WMBD MBD-1 = 0.4032 + 0.0109 WMBD 0.84 (36) 0.86 (36) Thomas et al. 1996 MBD = 2.151 - 0.027 WMBD MBD-1 = 0.4032 + 0.0109 WMBD 0.87 (39) 0.89 (39) Wagner et al. 1994 MBD = 2.009 - 0.024 WMBD MBD-1 = 0.4032 + 0.0109 WMBD 0.92 (30) 0.93 (30) MBD = 2.087 - 0.024 WMBD MBD-1 = 0.4308 + 0.0097 WMBD 0.91 (105) 0.93 (105) Aragon et al. 2000 MBD = 2.09 - 0.0031 WMBD 0.75 (26) Diâz-Zorita & Grosso 2000 MBD = 2.0509 – 0.0256 WMBD 0.90 (93) Krzic et al. 2004  25 Water content (kg kg-1) 0.18 0.20 0.22 0.24 0.26 0.28 0.30 B ul k de ns ity  (k g m -3 ) 1680 1700 1720 1740 1760 1780 1800 1820 1840 1860 MBD WMBD  Figure 1.1 General bulk density - moisture curve obtained with the standard Proctor test for a coarse-textured soil.  26 1.8 References Adams, P.W., and Froehlich, H.A. 1981. Compaction of forest soils. Pacific Northwest Extension publication. Oregon State Univ., Corvallis, O.R. PNW No. 217. Adams, W.A. 1973. The effect of organic matter on the bulk density and true densities of some uncultivated podzolic soils. J. Soil Sci. 24:10-17. Andrus, C.W., and Froehlich, H.A. 1983. An evaluation of four experiments used to till compacted forest soils in the Pacific northwest. For. Res. Lab, Oreg. State Univ., Corvallis, O.R. Res. Bull. No. 45. Aragon, A., Garcia, M.G., Filgueira, R.R., and Pachapsky, Y.A. 2000. Maximum compactibility of Argentina soils from the proctor test: the relationship with organic matter and water content. Soil Tillage Res. 56:197-204. American Society for Testing Materials. 2000. Design D698-00a. Standard test methods for laboratory compaction characteristics of soil using standard effort (12,400 ft-lbf/ft3 (600 kN- m/m3)). ASTM, West Conshohocken, P.A. Ball, B. C., R. W. Lang, M. S. O'Sullivan, and M. F. Franklin. 1988. Cultivation and nitrogen requirements for continuous winter barley on a gleysol and a cambisol. Soil Tillage Res. 13:333-352. Ball, B.C., Campbell, D.J., and Hunter, E.A. 2000. Soil compactability in relation to physical and organic properties at 156 sites in UK. Soil Tillage Res. 57:83-91. Ball, B.C., Cheshire, M.V., Robertson, E.A.G., and Hunter, E.A. 1996. Carbohydrate composition in relation to structural stability, compactibility and plasticity of two soils in a long-term experiment. Soil Tillage Res. 39:143- 160. Ballard, T.M., and Hawkes, B.C. 1989. Effect of burning and mechanical site preparation on growth and nutrition of planted white spruce. Can. For. Serv., Victoria, B.C. Inf. Rep. BC-  27 X-309. Baver, L.D. 1930. The Atterberg consistency constants: factors affecting their values and a new concept of their significance. J. Am. Soc. Agron. 22:935-948. Bennie, A.T.P., and Krynauw, G.N. 1985. Causes, adverse effects and control of soil compaction. S. Afr. J. Plant Soil 2:109-114. Berndt, H.W., and Gibbons, R.D. 1958. Root distribution of some native trees and understory plants growing on three sites within ponderosa pine watersheds in Colorado. USDA For. Serv., Fort Collins, C.O. Rocky Mountain Forest and Range Experiment Station Paper 37. Bishop, D.M. 1962. Lodgepole pine rooting habits in the Blue Mountains of northeastern Oregon. Ecology 43:140-142. Bodman, G.B., and Constantin, G.K. 1965. Influence of particle size distribution on soil compaction. Hilgardia 36:567-591. Brady, N.C., and Weil, R.R. 2004. Elements of the nature and properties of soils. 2nd ed. Pearson-Prentice Hall. N.J. Bulmer, C.E. 1998. Soil rehabilitation in British Columbia: a problem analysis. Land Management Handbook 44. B.C. Min. For. Research program, Victoria, B.C. 45 p. Bulmer and Simpson 2005 Bulmer, C., and Simpson, D.G. 2005. Soil compaction and water content as factors affecting the growth of lodgepole pine seedlings on sandy clay loam soil. Can. J. Soil Sci. 85:667–669. Caldwell, P.V., Vepraskas, M.J., and Gregory, J.D. 2006. Physical properties of natural organic soils in Carolina Bays of the southeastern United States. Soil Sci. Soc. Am. J. 71:1051-1057. Carr, W.W. 1987. Restoring productivity on degraded forest soils: two cases studies. For. Can. and B.C. Min. For., Victoria, B.C. FRDA Rep. No 002. Carter, M.R. 1990. Relative measures of soil bulk density to characterize compaction in tillage studies on fine sandy loams. Can. J. Soil Sci. 70:425-433.  28 Corns I.G.W. 1988. Compaction by forestry equipment and effects on coniferous seedling growth on 4 soils in the Alberta foothills. Can. J. For. Res. 18:75-84. Conlin, T.S.S., and van den Driessche, R. 1996. Short term effect of soil compaction on growth of Pinus contorta seedlings. Can. J. For. Res. 26:727-739. Courtin, P., Feller, M.C., and Klinka, K. 1983. Lateral variability in some properties of disturbed forest soils in southwestern British Columbia. Can. J. Soil Sci. 63:529–539. Da Silva, A.P., and Kay, B.D. 1997. Effect of soil water content variation on the least limiting water range. Soil. Sci. Soc. Am. J. 61:884–888. Daddow, R.L., and Warrington, G.E. 1983. Growth-limiting soil bulk densities as influenced by soil texture. USDA Agric. For. Serv., Ft. Collins, C.O. Rep. WSDG-TN-00005. Day, P.R. 1965. Particle fractionation and particle size analysis. In Methods of soil analysis, Part 1, Physical and mineralogical properties, including statistics of measurement and sampling. Edited by C.A. Black. Agronomy Monograph No. 9. Am. Soc. Agron., Madison, W.I. pp.545-577. Day, S.D., and Bassuk, N.L. 1994. A review of the effects of soil compaction and amelioration treatments on landscape trees. J. Arbor. 20:9-16. Dexter, A.R. 1975. Uniaxial compression of ideal brittle tilths. J. Terramech. 13:3-14. Diâz-Zorita, M., and Grosso, G.A. 2000. Effect of soil texture, organic carbon and water retention on the compactability of soils from the Argentinean pampas. Soil Tillage Res. 54:121-126. Dick, R.P., Myrold, D.D., and Kerle, E.A. 1988. Microbial biomass and soil enzyme activities in compacted and rehabilitated skid trail soils. Soil Sci. Soc. Am. J. 52:512-516. Donkin, M.J. 1991. Loss-on-ignition as an estimator of soil organic carbon in A-horizon forestry soils. Communications in Soil Science and Plant Analysis. 22:233-241. Douglas, J.T. 1986. Macroporosity and permeability of some soils cores from England and  29 France. Geoderma. 37:221-231. Fearnside, P.M. 2005. Deforestation in Brazilian Amazonia: history, rates, and consequences. Conserv. Biol. 19:680-688. Felton, G.K., and Ali, M. 1992. Hydraulic parameter response to incorporated organic matter in the B horizon. Trans. Am. Soc. Agric. Eng. 35:1153-1160. Firth, I., and Murphy, G. 1989. Skidtrails and their effect on the growth and management of young Pinus rudiata. New Zealand J. For. Sci. 19:22-28. Fleming R.L., Powers R.F., Foster N.W Kranabetter, J.M., Scott, D.A, Felix Ponder, F.Jr., Berch, S., Chapman, W.K., Kabzems, R.D., Ludovici, K.H., Morris, D.M., Page-Dumroese, D.S., Sanborn, P.T., Sanchez, F.G., Stone, D.M., and Tiarks, A.E. 2006. Effects of organic matter removal, soil compaction, and vegetation control on 5-year seedling performance: a regional comparison of Long-Term Soil Productivity sites. Can. J. For. Res. 36:529-550. Free, G.R., Lamb, J., and Carleton, E.A. 1947. Compactability of certain soils as related to organic matter and erosion. J. Am. Soc. Agron., 39:1068-1076. Froelich, H.A. 1973. The impact of even-aged forest management on physical properties of soils. In Even-age management. Edited by R.K. Herman and D.P. Lavender. School of Forestry, Oreg. State Univ., Corvallis. pp. 199-220. Froehlich, H.A., and McNabb, D.H. 1984. Minimizing soil compaction in Pacific Northwest forests. In Forest soils and treatment impacts. Proc. 6th N. Am. For. Soils Conf. Edited by E.L. Stone.  Dept. For., Univ. Tenn., Knoxville, T.N. pp. 159-192. Greacen, E.L., and Sands, R. 1980. Compaction of forest soils: a review. Aust. J. Soil Res. 18:163-189. Guerif, J. 1979. Mechanical properties of straw: the effects on soil. In Straw decay and its effect on disposal and utilization. Edited by E. Grossbard. Wiley, Chichester. pp.73-81. Guerif, J. 1990. Factors influencing compaction-induced increases in soil strength. Soil Tillage  30 Res. 16:167-178. Gupta, S.C., and Allmaras, R.R. 1987. Models to assess the susceptibility of soil to compaction. Adv. Soil Sci. 6:65-100. Gupta, S.C., Hada, A., Voorhees, W.B., Wolf, D., Larson, W.E., and Schneider, E.C. 1985. Development of guides for estimating the ease of compaction of world soils. Res. Rep. Submitted to BARD Fund on completion of Project Grant No. US-337-80. Gutierrez-Velez, V.H., and MacDicken, K. 2008. Quantifying the direct social and governmental costs of illegal logging in the Bolivian, Brazilian, and Peruvian Amazon. Forest Policy Econ. 10:248-256. Hakansson, I., 1990. A method for characterizing the state of compactness of the plough layer. Soil Tillage Res. 16:105-120. Hakansson, I., and Lipiec, J. 2000. A review of the usefulness of relative bulk density values in studies of soil structure and compaction. Soil Tillage Res. 53:71-85. Halter, M.R., and Chanway, C.P. 1993. Growth and root morphology of planted and naturally- regenerated Douglas-fir and lodgepole pine. Ann. Sci. For. 50:71-77. Harrington, T.B., and Howell, K.D. 1998. Planting cost, survival, and growth one to three years after establishing loblolly pine seedlings with straight, deformed, or pruned taproots. New For. 15:193-204. Heilman, P. 1981. Root penetration of Doulas-fir seedlings into compacted soil. For. Sci. 27:660-666. Heinonen, R. 1977. Towards ‘normal’ soil bulk density. Soil Sci. Soc. Am. J. 41:1214-1215. Henderson, C., Leveett, A., and Lisle, D. 1988. The effects of soil water content and bulk density on the compactibility and soil penetration resistance of some Western Australian sandy soil. Aust. J. Soil Res. 26:391-400. Heninger, R., Scott, W., Dobkowski, A., Miller, R., Anderson, H., and Duke, S. 2002. Soil  31 disturbance and 10-year growth response of coast Douglas-fir on nontilled and tilled skid trails in the Oregon Cascades. Can. J. For. Res. 32:233–246. Hermann, R.K. 1977. Growth and production of tree roots. In The below-ground ecosystem: a synthesis of plant-associated processes. Edited by J.K. Marshall. Colo. State Univ., Range Sci. Dept., Science Series 26. Fort Collins, C.O. 351 p. Hillel, D. Environmental soil physics. Academic Press, San Diego, C.A. 771 p. Holcomb, R.W. 1996. The long-term soil productivity study in British Columbia. FRDA report no. 256. Can. For. Serv. and B.C. Min. For., Victoria, B.C. p.23. Horn, R., and Lebert, M. 1994. Soil compactability and compressibility. In Soil compaction in crop production. Edited by B.D. Soane and C. van Ouwerkerk. Elsevier, AE Amsterdam, the Netherlands. pp. 45-70. Howard, R.F., Singer, M.J., and Frantz, G.A. 1981. Effects of soil properties, water content, and compactive effort on the compaction of selected California forest and range soils. Soil Sci. Soc. Am. J. 45:231–236. Igwe, C.A., and Stahr, K. 2004. Water-stable aggregates of flooded Inceptisols from South- eastern Nigeria in relation to mineralogy and chemical properties. Aust. J. Soil Res. 42:171- 179. Jakobsen, B.F., and Greacen, E.L. 1985. Compaction of sandy forest soils by forwarder operations. Soil Tillage Res. 5:55-70. Joosse, P.J., and R.A. McBride. 2003. Assessing physical quality of plastic soils of differing mineralogy and pre-stress history using mechanical parameters: I. Saturated compression tests. Can. J. Soil Sci. 83:45–63. Kabzems, R., Haeussler, S., 2005. Soil properties, aspen, and white spruce responses 5 years after organic matter removal and compaction treatments. Can. J. For. Res. 35:2045–2055. Kimmins, J.P., and Hawkes, B.C. 1978. Distribution and chemistry of fine roots in a white  32 spruce-subalpine fir stand in British Columbia: implications for management. Can. J. For. Res. 8:265-279. Kirby, J.M.. 1991. Critical-state soil mechanics parameters and their variation for Vertisols in eastern Australia. J. Soil Sci. 42:487-499. Koolen, A.J. 1974. A method for soil compactibility determination. J. Agri. Eng. Res. 19:271- 278. Koolen, A.J. 1987. Deformation and compaction of element soil volumes and effects on mechanical soil properties. Soil Tillage Res. 10:5-19. Krzic, M., Bulmer, C.E, Teste, F., Rahman, S., and Dampier, L. 2003. Relative measure of bulk density to characterize compaction of forest soils caused by harvest. Final report. Forestry Innovation Investment (Reference No. R2003-0219). Krzic, M., Bulmer, C.E., Teste, F., Dampier, L., and Rahman, S. 2004. Soil properties influencing compactability of forest soil in British Columbia. Can. J. Soil Sci. 84:219-226. Kuiper, H. 1959. Confined compression tests on soil aggregate samples. Medad. Landbouwgesch. Opzoeekingsstn. Staat Gent 24: 349-357. Kuiper, L.C., and Coutts, M.P. 1992. Spatial disposition and extension of the structural root system of Douglas-fir. For. Ecol. Manage. 47:111-125. Larson, W.E., Gupta, S.C., and Useche, R.A. 1980. Compression of agricultural soils from eight soil orders. Soil Sci. Soc. Am. J. 44:450-457. Lipiec, J., Hakansson I., Tarkiewicz, S., and Kossowski, J. 1991. Soil physical properties and growth of spring barley as related to the degree of compactness of two soils. Soil Tillage Res. 19:307-317. Lipiec, J., and Hatano, R. 2003. Quantification of compaction effects on soil physical properties and crop growth. Geoderma. 116:107-136. Lotan, J.E. 1964. Initial germination and survival of lodgepole pine on prepared seedbeds.  33 USDA Forest Service, Research Note INT-29. Intermountain Forest and Range Experiment Station, Ogden, U.T. Lotan, James E. and William B. Critchfield. 1990. Pinus contorta ssp. murrayana – lodgepole pine forest. In Silvics of North America, Volume 1. Conifers. Coordinated by R.M. Burns and B. H. Honkala. Agriculture Handbook 654. USDA For. Serv., Washington, D.C. pp. 302-313. Lousier, J.D. 1990. Impacts of forest harvesting and regeneration on forest soils. B.C. Min. For., Victoria, B.C. Land Management. Rep. No. 67. Mapfumo, E., and Chanasyk, D.S. 1998. Guidelines for safe trafficking and cultivation, and resistance-density-moisture relations of three disturbed soils from Alberta. Soil Tillage Res. 46:193-202. McBurnett, S.L., and Franzmeier, D.P. 1997. Pedogenesis and cementation in Calcareous till in Indiana. Soil Sci. Soc. Am. J. 61:1098-1104. McKeague, J.A., and Sprout, P.N. 1975. Cemented subsoils (duric horizons) in some soils of British Columbia. Can. J. Soil Sci. 55:189-203. McKeague, J.A., and Wang, C. 1980. Micromorphology and energy dispersive analysis of orstein horizons of podzolic soils from New Brunswick and Nova Scotia, Canada. Can. J. Soil Sci. 60:9-21. McKee Jr. W.H., Hatchell, G.E., and Tiarks, A.E. 1985. Managing site damage from logging. A loblolly pine management guide. General technical Report, SE-32, USDA For. Serv., pp. l- 21. McMinn, R.G. 1978. Root development of white spruce and lodgepole pine seedlings following outplanting. In Proceedings of the root form of planted trees symposium, May 16-19, Victoria, BC. Edited by E. Van Eerden and J. M. Kinghorn. Victoria, B.C. pp. 186-190. Miles, D.W.R., and Swanson, F.J., and Youngberg, C.T. 1984. Effect of landslide erosion on  34 subsequent Douglas-fir growth and stocking levels in the western Cascades, Oregon. Soil Sci. Soc. Am. J. 48:667-671. Milford, M.H., Kunze, G.W., and Bloodworth, M.E. 1961. Some physical, chemical and mineralogical properties of compacted and adjacent soil layers in coarse textured soils Soil Sci. Soc. Am. J. 25:511-515. Miller, R.E., Scott, W., and Hazard, J. 1996. Soil compaction and conifer growth after tractor yarding at three coastal Washington locations. Can. J. For. Res. 26:225-236. Ministry of Forests. 2001. Annual report. B.C. Min. For., Victoria, B.C. Ministry of Forests and Range. 2007. The State of British Columbia’s forests, 2006. B.C. Min. For. Range, Victoria, B.C. Moolman, J.H. 1981. Soil textural properties influencing compactibility (abstract). Agrochemophysica. 13:13-19. Nadezhdina N., Cermak, J., Neruda, J., Prax A., Ulrich R., Nadezhdin, Gasparek J., and Pokorny, E. 2006. Roots under the load of heavy machinery in spruce trees. Eur. J. Forest Res. 125:111-128. Natural Resources Canada. 2008. The state of Canada’s forests, Annual report 2008. Natural Resources Canada, Ottawa, O.N. Nigh, G.D., Ying, C.C., and Qian, H. 2004. Conifer species in the interior of British Columbia, Canada. For. Sci. 50:659-671. Nwadialo, B.E., and Mbagwu, J.S.C. 1991. An analysis of soil components active in microaggregate stability. Soil technology, 4:343-350. O’Dogherty, M.J., and Wheeler, J.A. 1984. Compression of straw to high densities in closed cylindrical dies. J. Agric. Eng. Res. 29:61-72. Oda, M. 1972. Initial fabrics and their relations to mechanical properties of granular material. Soil Found. 12:17-36.  35 Ohu, J.O., Raghavan, G.S.V., and McKeyes, E. 1985. Peatmoss effect on the physical and hydraulic characteristics of compacted soils. Trans. ASAE 28:420-424. Page-Dumroese, D.S., Harvey, A.E., Jurgensen, M.F., and Amaranthus, M.P. 1998. Impacts of soil compaction and tree stump removal on soil properties and outplanted seedlings in northern Idaho, USA. Can. J. Soil Sci. 78:29-34. Pinheiro-Dick, D., and Schwertmann, U. 1996. Microaggregates from Oxisols and Inceptisols: dispersion through selective dissolutions and physicochemical treatments. Geoderma. 74:49–63. Preston, R.J. 1942. Growth and development of root systems of juvenile lodgepole pine. Ecol. Monogr. 12:449-468. Proctor, R.R. 1933. Fundamental principles of soil compaction. Description of field and laboratory methods. Eng. News Record. 111:286-289. Redding T.E., Hannam, K.D., Quideau, S.A., and Devito, K.J. 2005. Particle density of aspen, spruce, and pine forest floors in Alberta, Canada. Soil Sci. Soc. Am. J. 69:1503-1506. Richard, G., Cousin, I., Sillon, J. F., Bruand, A., and Guerif, J. 2001. Effect of compaction on the porosity of a silty soil: influence on unsaturated hydraulic properties. Euro. J. Soil Sci. 52:49-58. Robert, J.A., and Lindgren, B.S. 2006. Relationships between root form and growth, stability, and mortality in planted versus naturally regenerated lodgepole pine in north-central British Columbia. Can. J. For. Res. 36:2642-2653. Rudolph, P.O. 1939. Why forest plantations fail. J. For. 37:377-383. Safford, L.O., and Bell, S. 1972. Biomass of fine roots in a white spruce plantation. Can. J. For. Res. 2:169-172. Saini, G.R. 1966. Organic matter as a measure of bulk density of soil. Nature. 210:1295-1296. Schoenholtz, S.H., van Miegroet, H., and Burger, J.A. 2000. A review of chemical and physical  36 properties as indicators of forest soil quality: challenges and opportunities. For. Ecol. Manage. 138:335-356. Shirazi, M.A., and Boersma, L. 1984. A unifying quantitative analysis of soil texture. Soil Sci. Soc. Am. J. 48:142-147. Slowinska-Jurkiewicz, A., and Domżal, H. 1991. The structure of the cultivated horizon of a Stagnogleyic Luvisol developed from loam, within the zone affected by tractor wheel traffic. Soil Tillage Res. 19:245-253. Smith, C.W., Johnston, M.A., and Lorentz, S. 1997a. Assessing the compaction susceptibility of South African forestry soils. I. The effects of soil type, water content and applied pressure on uni-axial compaction. Soil Tillage Res. 41:53-73. Smith, C.W., Johnston, M.A., and Lorentz, S. 1997b. Assessing the compaction susceptibility of South African forestry soils. II. Soil properties affecting compactibility and compressibility. Soil Tillage Res. 43:335-354. Smith, R.B., and Wass, E.F. 1991. Impacts of two stumping operations on site productivity in interior British Columbia. For. Can. Pac. For. Cent. Victoria, B.C. Inf. Rep. BC-X-327. Soane, B.D. 1975. Studies on some soil physical properties in relation to cultivation and traffic. In Soil physical conditions and crop production. Min. Agric. Food Fish, Tech. Bull. 29. pp.160-183. Soane, B.D. 1990. The role of organic matter in soil compactibility: a review of some practical aspects. Soil Tillage Res. 16:179-201. Soane, B.D., Campbell, D.J., and Herkes, S.M. 1972. The characterization of some Scottish arable topsoils by agricultural and engineering methods. J. Soil Sci. 23:93-104. Standish, J.T., Commandeur, P.R., and Smith, R.B. 1988. Impacts of forest harvesting on physical properties of soils with reference to increased biomass recovery: a review. Can. For. Serv., Victoria, B.C. Inf. Rep. BC-X-301.  37 Stengel, P., Douglas, J.T., Guerif, J., and Goss, M.J. 1984. Factors influencing the variation of some properties of soil in relation to their suitability for direct drilling. Soil Tillage Res. 4:35–53. Steve, W., and Thomas B.R.. 1981. Effects of skid roads on diameter, height, and volume growth in Douglas-fir. Soil Sci. Soc. Am. J. 45:629-632. Stone, D.M., Kabzems, R., 2002. Aspen development on similar soils in Minnesota and British Columbia after compaction and forest floor removal. For. Chron. 6, 886–891. Strong, W.L., and La Roi, G.H. 1983. Root system morphology of common boreal forest trees in Alberta, Canada. Can. J. For. Res. 13:1164-1173. Sutton, R.F. 1991. Soil properties and root development in forest trees: a review. For. Can., Sault Ste. Marie, O.N. Inf. Rep. O-X-413. Taylor, H.M., and Gardner, H.R. 1963. Penetration of cotton seedling taproot as influenced by bulk density, moisture content, and strength of soil. Soil Sci. 96:153-156.  Thacker, D.J., Campbell, J.A., and Johnson, R.L. 1994. The effect of soil compacton on root penetration, mechanical impedance and moisture-density relationships of selected soils of Alberta. A.B. Land Conserv. Reclam Manag. Edmonton, A.B. Group Report #RRTAC OF-9. Thomas, G.W., Haszler, G.R., and Blevins, R.L. 1996. The effects of organic matter and tillage on maximum compactability of soils using the Proctor test. Soil Sci. 161:502-508. Topp, G.C., Reynolds, W.D., Cook, F.J., Kirby, J.M., and Carter, M.R, 1997. Physical attributes of soil quality. In Soil quality for crop production and ecosystem health. Development in Soil Science. Edited by E.G. Gregorich and M.R. Carter. Elsevier, AE Amsterdam, the Netherlands. pp. 21-58. Utzig, G.F., and Walmsley, M.E. 1988. Evaluation of soil degradation as a factor affecting forest productivity in British Columbia: a problem analysis Phase I. For. Can. and B.C. Min. For., Victoria, B.C. FRDA Rep. No. 025.  38 Valentine, K.W.G., Sprout, P.N., Baker, T.E., and Lavkulich, L.M, 1978. The soil landscapes of British Columbia. B.C Min. Environ., Victoria, B.C. 197 p. Van Der Watt, H.v.H. 1969. Influence of particle size distribution on soil compactibility. Agrochemophysica. 1:79-86. Veihmeyer, F.J., and Hendrickson, A.H. 1948. Soil density and root penetration. Soil Sci. 65:487-493. Wagg, J.W.B. 1967. Origin and development of white spruce root forms. Can. Dept. For. Rural Dev., Forestry Branch, Pub. 1192. Ottawa, O.N. 45 p. Webster, R., and Oliver, M.A. 1990. Statistical methods in soil and land resources survey. Oxford University Press, Oxford, U.K. 316 p. Whalley, W.R., Dumitru, E., and Dexter, A.R. 1995. Biological effects of soil compaction. Soil Tillage Res. 35:53-68. Willatt, S.T. 1986. Root growth of winter barley in a soil compacted by the passage of tractors. Soil Tillage Res. 7:41-50. Willatt, S.T. 1987. Influence of aggregate size and water content on compactibility of soil using short-time static loads. J. Agric. Eng. Res. 32:107-115. Zhang, H., Hartge, K.H., Ringe, H. 1997. Effectiveness of organic matter incorporation in reducing soil compactability. Soil Sci. Soc. Am. J. 61:239-245. Zou, C., Sands, R., Buchan, G., and Hudson, I. 2000. Least limiting water range: a potential indicator of physical quality of forest soils. Aust. J. Soil Res. 38:947-958.  39 2 USE OF A CONSTANT-VOLUME, SMALL-SAMPLE GAS PYCNOMETER TO TEST VARIATION IN PARTICLE DENSITY OF FOREST SOILS IN INTERIOR BRITISH COLUMBIA1 2.1 Introduction Accurate assessment of soil particle density is important for the calculation of heat- capacity and soil porosity and determination of particle size distribution (PSD). A particle density of 2.65 Mg m-3 is often assumed for the mineral soil since quartz, feldspars, micas, and the colloidal silicates that usually comprise the major matrix of mineral soils have densities around this value (Hillel 1980; Smettem 2006). However, soil particle density may deviate from this value when soil organic matter (SOM) content is high or when a soil has developed on parent material that contains minerals different from quartz or feldspars. For example, particle density of humus is close to 1.50 Mg m-3, while opal has a particle density of 1.90 Mg m-3 (Flint and Flint 2002); ferromagnesian minerals have particle densities between 2.90 and 3.50 Mg m-3, while iron oxides and other heavy minerals have particle densities that exceed 4.00 Mg m-3 (Smettem 2006). Some studies have also shown that accurate determination of particle density is needed when a data set encompasses soil types developed on parent materials of varying mineralogy (Smith et al. 1997; McBratney et al. 2002; Zhao et al. 2008). Standard methods for the determination of soil particle density are based on measuring mass and volume. Soil mass is directly and easily measured by weighing, while volume is indirectly measured by liquid or gas pycnometry with a greater degree of difficulty. A liquid  1 A version of this chapter will be submitted for publication. Zhao, Y.H., Bulmer, C.E., and Krzic, M. Use of a constant-volume, small-sample gas pycnometer to test variation in particle density of forest soils in Interior British Columbia.  40 pycnometer (LP) is a specific gravity flask (or a volumetric flask for large samples) used to calculate the volume of a fluid (i.e., water) displaced by the sample (Flint and Flint 2002). This method involves careful weighing to obtain the volume of a soil sample (vs) from w assww ws wwwwvv ρ ))(( −−−==                                                                                    (1) where vw is the volume of water displaced by the sample at the observed temperature; ww is the weight of the pycnometer filled with water at the observed temperature; wsw is the weight of the pycnometer filled with the soil sample and water; ws is the weight of the pycnometer and soil sample; wa is the weight of the pycnometer filled with air; and ρw is the density of water at the observed temperature. Soil particle density is then calculated from s as s v ww −=ρ                                                                                                                   (2) The liquid pycnometry method requires complete saturation of the soil sample to obtain an accurate estimation of the volume of soil solid, which is achieved by grinding the sample to break the aggregates before testing (Van Keulen 1973) and evacuating or heating to de-air the sample and the liquid during the test (Heiskanen 1992). This method is time consuming (e.g., four to five samples can be done in an hour) and it also requires good technique (i.e., constant sample handling through all procedures and for all samples) to control numerous sources of error (Sojka 1988). Gas displacement methods used to measure particle density of porous solids have been considered faster, easier, and more accurate relative to liquid pycnometry (Flint and Flint 2002). Gas pycnometry (Say 1797) is based on Boyle-Mariotte’s law of volume-pressure relationships. This law describes the inversely proportional relationship between the absolute pressure and the volume of a gas within a closed system, if the temperature is kept constant. The gas  41 displacement method does not have the problem of air entrapment, it is highly accurate, and is easy and quick to carry out (e.g., 10-12 samples can be done in an hour) (Kummer and Cooper 1945; McIntyre et al. 1965; Bielders et al. 1990; Marinder 1996). A commercial gas pycnometer (GP) has been recommended over a custom GP for accurate volume determination (ASTM 2000). However, commercial GPs are expensive and have been known to have low accuracy in determining the volume of solid particles (Tamari 2004; Geddis et al. 1996). With the development of optimization methods (Tamari 2004), it is possible to design and construct a custom GP with high accuracy and at low cost. Soil particle density value tested by the gas pycnometer in this study is the average particle density for the entire fine fraction, including organic and mineral components, which is referred to as soil particle density, while specific particle density for these two components are referred to as particle density of SOM, and mineral particle density, respectively. Some soil chemical and physical properties have been shown to be correlated with soil particle density. Ruhlmann et al. (2006) used the mass proportion of organic matter to the mass of soil solids (mOM) to describe the influence of SOM on soil particle density, and the authors found that mOM was significantly (R2=0.96) related to soil particle density. Soils with a high content of Fe- and Mn- oxides tend to have high particle density, while presence of Al-oxide contributed to the low value of particle density. For example, magnetite (Fe3O4) has particle density of 5.2 Mg m-3, and kaolinite (Al4SiO10(OH)8) and K-feldspar (KAlSi3O8) have a particle density around 2.6 -2.7 Mg m-3 (Flint and Flint 2002), indicating that Fe- and Mn- oxides are positively and Al-oxide negatively related to soil particle density. Since most clay minerals have a higher particle density compared to quartz (Ruhlmann 2006), clay could be positively and sand could be negatively related to soil particle density. This study described a procedure to obtain the pressure-volume model, took a new method to analyze the accuracy of constant-volume GP, and evaluated the particle density of  42 forest soils in interior BC, The objectives of this study were to (1) assess the accuracy of the volume test for the constant-volume GP that we developed for small sample sizes (< 36 cm3); (2) examine the range of variation of particle density for forest soils in interior BC; and (3) evaluate the relationship between particle density and selected soil physical and chemical properties (i.e., SOM, oxides, and PSD). Hypotheses tested were (1) constant-volume gas pycnometry was a precise and accurate method to test soil particle density; (2) variation in soil particle density was explained by SOM, oxides, and PSD; and (3) mineral particle density of forest soils in interior BC varied from 2.65 Mg m-3. 2.2 Materials and methods 2.2.1 Description of the constant-volume, small-sample gas pycnometer Generally, a constant-volume GP is composed of a sample chamber, a gas tank, and a pressure transducer (Fig. 2.1). The tank and the chamber can be connected pneumatically through a tube with a two-way valve ‘Z’. The tank is also connected to the pressure transducer and the air supply through a tube with a main two-way valve ‘M’ (Fig. 2.1). For the small- sample, constant-volume GP (small-sample GP hereafter) that we have developed (Fig. 2.2), the volume of both the tank and chamber was adjustable to accommodate the size of the sample. We used a gauge pressure transducer instead of an absolute pressure transducer. All the walls of the chamber, the tank, and the tubes were made of aluminum, which conducts heat quickly, so that heat change generated during the test was neutralized by the surrounding heat when the equipment and sample were put in the lab with a constant temperature (i.e., 25oC). 2.2.2 Operation of the pycnometer The following steps described the procedure to run a test in our study: (1) place the oven-dried sample into the chamber; (2) open valves ‘Z’ and ‘M’ to fill the pycnometer with atmospheric air;  43 (3) adjust the reading of the gauge P to zero; (4) close valve ‘Z’ to isolate the chamber, while the gauge is connected to the tank; (5) fill the tank with pressurized gas until gauge reading is 20.00 PSI; (6) close valve “M’. For each test run, the tank is always filled with the pressurized gas to 20.00 PSI at this step; (7) open valve ‘Z’ to allow gas to expand from the tank to the chamber (or vice versa); and (8) record the gauge reading at the point when it stops changing (i.e., gas expansion is finished; this usually occurred within 20 to 30 seconds). 2.2.3 Calibration of the small-sample gas pycnometer For the accurate application of the Boyle-Mariotte’s law, it is necessary to measure the internal volume of the tank, the chamber, the pressure gauge, valves ‘M’ and ‘Z’, and the tubes between them. These volumes were difficult to measure directly and were not measured in the development of our small-sample GP; instead, aluminum discs of a known volume (Fig. 3) were used to obtain a series of volume-pressure paired data. Precise measurement of the volume of these discs is the key in the calibration, which was done using the Archimedes’ method (Flint and Flint 2002). An empirical equation between volume and pressure was derived based on the Boyle-Mariotte’s law. Pycnometer equations The following assumptions were made before the equations were derived: (1) the gas inside the pycnometer behaves ideally (i.e., its compressibility is negligible and gas is not adsorbed on solids), (2) the sample and the pycnometer’s components are rigid, (3) the pycnometer is gas-tight and the expanding gas quickly reaches a static equilibrium, and (4) the mass effect of the initial air that stays in the chamber on the test, as of step no. 2 in the “Operation of the pycnometer”, is negligible.  44 For the calibration, when the tank is filled with gas (i.e., pti  = 20.00 PSI), the following equations were obtained using the Boyle-Mariotte’s law: pti × vt = nti × R × T                                                                                                         (3) pci × (vc - vs) = nci × R × T = 0 (pci = 0.00 PSI)                                                               (4) When gas expansion between the tank and the chamber stops, ptf × vt = ntf × R × T                                                                                                         (5) pcf × (vc - vs) = ncf × R × T (pcf = ptf  = pf)                                                                        (6) where subscriptions i and f refer to steps 5 and 8 in the ‘Operation of the pycnometer’, respectively; v is the volume; p is the gauge reading in PSI; n is the number of moles of gas; R is the gas constant; T is the temperature of pycnometer (i.e., 25oC). Subscriptions t and c refer to tank and chamber; and s refers to solid of either the calibration discs or the soil sample. The mass of the gas enclosed in the pycnometer does not change after the tank is filled to 20.00 PSI. Consequently, nti + nci = ntf + ncf                                                                                                             (7) then, rearrange (3) – (6) , ti tti n TR vp =× ×                                                                                                                     (3)’ ci scci n TR vvp =× −× )(                                                                                                          (4)’ tf tf n TR vp =× ×                                                                                                                    (5)’ cf scf n TR vvp =× −× )(                                                                                                          (6)’ and replace n in equation (7) with corresponding components from equation (3)’ - (6)’, the volume of soil sample can be expressed as  45 ct f tti s vvp vpv ++×−=                                                                                                     (8) Since the initial pressure (pti) in equation (8) was set to a constant value (i.e., 20.00 PSI), equation 8 takes the form of c x ay +−= , where a and c are constants. We obtained a series of pressure-volume paired data through calibration and used these data to solve the constants of a and c, then vt and vc can be solved through vt = a/20 and vc = c - vt. 2.2.4 Test of samples Two sets of samples were analyzed during this study – the first set included two commercially available white (quartz) sands and one black sand that were commercially available, while the second set included 283 soil samples collected at eight forest sites in the interior BC. Both LP and small-sample GP were used to evaluate particle density of the three sand types, while small-sample GP was used to test the particle density of the soil samples. Sampling sites were located within the Interior Plateau physiographic region (Holland 1964). Soil samples were collected immediately below the interface between the surface organic material (if present) and the underlying mineral material. On sites where coarse fragments (diameter > 2 mm) occupied less than 25% by volume, soil samples were collected in 518 cm3 cores, while the excavation method was used on sites where coarse fragment content exceeded 25% by volume (Grossman and Reinsch 2002). Samples were sieved through a 2-mm sieve and used for the tests. 2.2.5 Soil organic matter, oxalate-extractable oxides, and particle size distribution Soil total C was determined by the dry combustion method (Nelson and Sommers 1996) using a LECO analyzer. Soil oxides of Al, Fe, Mn, and Si were extracted by 0.2 M ammonium oxalate solution. This method extracts active Al-, Fe-, Mn-, and Si-oxides, including a fraction of organic bound oxides (Loeppert and Inskeep 1996). The extracted ions were measured by  46 inductively coupled plasma (ICP) spectrometer. Soil PSD was determined by the hydrometer method and wet sieving (Gee and Or 2002). Samples were pre-treated with hydrogen peroxide (30%) and heat. Particle size distribution was described using the Canadian System of Soil Classification (Sheldrick and Wang 1993) in terms of percentage of clay (<0.002 mm), fine silt (0.002-0.005 mm), medium silt (0.005-0.02 mm), coarse silt (0.02-0.05 mm), very fine sand (0.05-0.10 mm), fine sand (0.10-0.25 mm), medium sand (0.25-0.50 mm), coarse sand (0.50-1.00 mm), and very coarse sand (1.00-2.00 mm), while sand fraction was reported as very fine to coarse (0.05-1.00 mm) and very coarse sand in this study. 2.2.6 Calculation of mineral particle density Since particle density determined by the small-sample GP includes both organic matter and mineral particles, we rearranged the equation that Adams (1973) used in calculating particle density of SOM to calculate particle density of soil mineral components: o s s m SOM SOM ρρ ρρ ×− ×−= 1 )1(                                                                                                       (9) where ρm is the mineral particle density; ρs is the soil particle density; and ρo is the particle density of SOM, which was set at 1.5 Mg m-3 (Kellogg and Wangaard 1969). The SOM content was calculated by multiplying the total C by a factor of 1.724 (Baldock and Skjemstad 1999). 2.3 Statistical analysis We used SAS REG procedure (SAS Institute 1990) to carry out multiple regression analysis, with particle density as the dependant variable; SOM, soil oxalate-extractable oxides (i.e., Al- and Fe-oxides), and PSD (i.e., clay, medium silt, coarse silt, and very coarse sand) as independent variables. A stepwise method was used to exclude any independent variables that may have overlapping effects on the dependent variable. The χ2 significance level was set at  47 0.25 for entry of variables and 0.10 for retention of variables, respectively. The SAS GLM (SAS Institute 1990) with Scheffe’s test was used for multi-comparison of mineral particle densities among groups, with α value set at 0.05 for the test of significance. 2.4 Results and discussion 2.4.1 Calibration curve The small-sample GP showed a strong relationship between the volume of solids and the reciprocal of the gauge reading (Fig. 2.4). This strong relationship implies the possibility of running a two-point only calibration - one when the sample chamber is empty and the other when the chamber is filled with sample, to derive the relationship between the volume and the gauge reading. Values of the geometric factor (volumetric ratio of the tank to the sample chamber), the filling factor (volumetric ratio of solids to the sample chamber), and the pressure factor (ratio of the maximum pressure to the minimum pressure) of the small-sample GP (Table 2.1) all were within the ranges recommended by Tamari (2004) who pointed out that a filling factor ranging from 0.4 - 0.7, with a geometry factor range of 0.30 - 0.60, and a pressure factor range of 2.0 - 7.5 are good solutions for minimizing errors during the test. Due to the porosity of the soil, the filling factor obtained using the actual soil samples decreased by 22% relative to the filling factor obtained using the calibration discs (Table 2.1). 2.4.2 Accuracy (uncertainty) analysis When the volume-pressure equation was obtained, change of gauge readings with the change of solid volume can be expressed by the derivative of vs to pf in equation 8, as )()()()()( 22 , f f f f tti fct f tti s pdp cpd p vppdvv p vpvd =×=++×−=                                       (10) If other conditions, such as being tight, thermostatic, and rigid are satisfied, the product of pti and vt is a constant (c). Under such a condition, the uncertainty of the test - minimum  48 volume change of the solid that could be tested by the pycnometer - is directly related to the precision of the gauge (i.e., the part d(pf )). For example, the minimum change of gas pressure tested by the gauge we used was 0.01 PSI. If its reading precision improved to 0.001, the uncertainty of the volume estimation would decrease by 10 folds. On the other hand, the higher precision would put more restrictions on conditions such as being tight and thermostatic. Since precision of a commercial pressure transducer was a constant when it was manufactured, its influence on the volume estimation was often neglected in the uncertainty analysis (Tamari 2004).  Danielson and Sutherland (1986) also pointed out that accurately calibrated pressure transducer is not required if the transducer’s response varies linearly with pressure. Our analysis showed that the precision of the pressure transducer is an important factor in determining the accuracy or uncertainty of a volume-constant GP. With the precision of our gauge at 0.01 PSI, a strong relationship was observed between the uncertainty of the small-sample GP and the volume of solids tested (Fig. 2.5). The maximum test error (0.096 cm3) was recorded when the sample chamber was empty, while the minimum test error (0.006 cm3) was when the chamber was completely filled with solids (Table 2.2). When tested with soils, the average volume of a solid was 14.102 cm3, which corresponded to an uncertainty of 0.050 cm3 or 0.35% of the volume of the solid (Table 2.2), which proved the hypothesis that constant-volume gas pycnometry was a precise method to test soil particle density. Tamari (2004) used the law of propagation of uncertainty to prove that the pycnometer’s accuracy is greatly improved when the sample chamber is filled with as much of the solid particles as possible. We achieved the same result while the new analysis method used in this study was simpler and more straightforward than the method the above author used. Our results showed that the precision of gauge and volume of the sample are two factors directly influencing the accuracy of the test. It has been reported that the relative uncertainty of volume estimation was more sensitive to pressure than to temperature. Tamari (2004) pointed  49 out that variation in air pressure may exceed 10 Pa within 1200 seconds, and such a change in air pressure during a test would cause substantial inaccuracy for a constant-volume GP with an absolute pressure transducer. With a gauge pressure transducer being used, the influence of naturally changing air pressure on the test was negligible and no calibration was needed for each value of the atmospheric pressure in the development of a constant-volume GP (Tamari 2004). In addition, it took less than 200 seconds from filling the chamber with pressurized gas to getting the final pressure reading. On the other hand, temperature would not influence the result when it is maintained relatively constant during each test (Flint and Flint 2002). It was, however, necessary to zero the gauge each time when the weather condition, such as air temperature, air pressure, and humidity, changed. The minimum error when the chamber was filled with solids was not zero for the small- sample GP (Table 2.2), which indicated that air stayed in the chamber before step 7 in the “operation of the pycnometer” and mass of gas in the tank may have had effects on the volume estimation. The Boyle-Marriote’s law describes gas behaviour under absolute pressure, while we measured the gas pressure relative to the atmospheric pressure, which would have caused error in the modeling. However, when all other conditions (e.g., being rigid, gas-tight, thermostatic) are satisfied, the error caused by the above reasons was negligible (0.01%, Table 2.2). 2.4.3 Test of commercial sands and soil samples Particle density values of each of the three sand types tested with both LP and the small- sample GP corresponded very close, with values obtained by the LP 0.4-0.6% higher than those by the small-sample GP (Table 2.3). The particle density of the black abrasive sand specified by the company was around 2.8 Mg m-3, while we measured it as 3.0 Mg m-3. This product was not a well-defined material from an industrial process, which could have resulted in a particle density different from the company specifications. Air entrapment generally does not occur in  50 structureless sands (i.e., sand with no aggregation), therefore values obtained by the LP would be closer to the actual particle density values for these sands. For the GP test, the uncertainty was 0.21% when the sample chamber was filled with sand (Table 2.2), which would be responsible for the lower particle density values obtained by the small-sample GP. Liquid pycnometer is preferred to GP for testing structureless materials. The GP, on the other hand, has been proven a faster and more accurate device than a LP for well-aggregated, porous materials (Flint and Flint 2002). These considerations are especially important for soil testing, since most soil types are characterized by some degree of aggregation. Even though most macro-aggregates (> 1 mm) are destroyed during sample preparation (i.e., sieving and drying), meso- and micro- aggregates still persist in samples under various conditions (Oades 1993; Six et al. 2004). Variation in SOM content and soil mineralogy accounts for variation in soil particle density. Cellulose and lignin, which are the primary constituents of cell walls in trees, have a particle density of 1.50 Mg m-3 (Kellogg and Wangaard 1969), while the particle density of SOM varies narrowly between 1.3 and 1.5 Mg m-3 (Caldwell et al. 2006). In our study, deviation in estimating particle density of SOM from 1.3 to 1.5 Mg m-3 accounted for 1% variation, on average, for the determination of mineral particle density. On the other hand, variability in SOM content was much larger, ranging from 4.6 to 214.1 g kg-1 (Table 2.4), which influenced the variation of mineral particle density to a larger extent. Mineral particle density decreased weakly with the increase of SOM in this study (Fig. 2.6a), indicating that particle density of SOM would be slightly less than the value (i.e., 1.5 Mg m-3) considered in the calculation of mineral particle density, especially when SOM was > 120 g kg-1. With three out of 283 soil samples with high SOM (i.e., SOM > 120 g kg-1) being removed, soil particle density decreased by 0.0028 Mg m-3 per g kg-1 increase of SOM (Fig. 2.6b). Soil organic matter explained 40% of the variation in soil particle density while more variation was due to soil mineralogy. In a study that focused on the determination of soil particle density from  51 SOM and mineral matrix for various soil types in Germany, Ruhlmann et al. (2006) found that the effectiveness of soil organic carbon increase per g kg-1 in reducing soil particle density was more pronounced (i.e., 0.0052 Mg m-3) at lower range (i.e., 3.7 – 68.4 g kg-1), and this relationship was independent of the soil type. Soils included into the study mentioned above consisted of K-feldspars, smectite, dolomite, muscovite, chlorite, and pyrite, all of which have particle densities different than quartz. However, the average of soil particle density from each site was relatively similar to each other (2.68±0.007 Mg m-3), indicating that the soil organic matter was the major source of variation in soil particle density. In our study, grouping soils by their geographical locations substantially improved the relationship between soil particle density and SOM (Fig. 2.7). This was especially true for locations 2, 3, and 4. The similar regression coefficients (Fig. 2.7) obtained among locations 1, 2, and 4 indicated that strength of SOM in decreasing the particle density was not substantially changed, which provided insight into the larger variations in soil mineralogy among groups than within the group. In soil types where SOM accounted for a large portion of the variation in particle density, a relatively uniform parent material is indicated. This is illustrated well for location 4, where the soil samples were all collected from glaciofluvial terraces with parent materials that were sorted by running water. Mineral particle density of soils collected in interior BC was often different than particle density of quartz, which was considered to be around 2.65 Mg m-3, ranging from 2.36 to 2.94 Mg m-3 (Table 2.4). Such variation may arise from variation in the types of heavy minerals in soils derived from glacial till (Holland 1964). Soil samples from five geographic locations were stratified into three groups that were different (P < 0.05) in the particle density, with one group having particle density significantly greater than 2.65 Mg m-3 and two groups having particle density significantly lower than this value (Table 2.5). It appeared that particle densities were consistently lower for samples from three sites in the southern interior BC (i.e., locations 1, 4,  52 and 5) than for the other sites (Table 2.5). The wide variation in mineral particle density values between sites reflected the intricate arrangement of many different rock types in the province (Valentine et al. 1978). In studying mineralogy and soil genesis of soils from central and northeastern BC, Arocena and Sanborn (1999) found that mica, chlorite, and kaolinite were the most common minerals in the clay fraction, while quartz and feldspar were predominant in the sand fraction. Multiple regression analysis showed that oxalate-extractable oxides, silt, very coarse sand, and clay explained 62% of the variation in the mineral particle density for the whole data set (Table 2.6). For a subset of samples, except location 5, Al- and Fe-oxides were important variables related to mineral particle density with Al-oxide being negatively and Fe-oxide positively related to the mineral particle density (Table 2.6). Our samples were relatively coarse textured, with the sand fraction varying from 0 to 975 g kg-1 with an average of 504 g kg-1. However, there was a weak relationship between mineral particle density and sand fraction (R2=0.02, Appendix 1), which indicated that heavy minerals, such as chlorite (particle density 3.4 Mg m-3, Ruhlmann et al. 2006) and muscovite (particle density 2.8 Mg m-3, Ruhlmann et al. 2006) that were present in clay and silt fractions, have contributed to the variation in mineral particle density. The hypothesis that variation in soil particle density was explained by SOM, oxides, and PSD was accepted in this study: about 77% variation in soil particle density was explained by SOM, soil oxides, and soil PSD, while the relationship was further improved when samples were stratified according to their geographic locations (Table 2.7); inclusion of soil oxides and PSD substantially improved the relationship between soil particle density and SOM, which was especially true for Al- and Fe-oxides. McKeague and Day (1966) studied the quantitative use of Fe and Al in characterizing soil processes for samples collected throughout nine provinces and territories in Canada and showed that oxalate extraction dissolved most of the amorphous Fe- or  53 Al-oxides but very little of crystalline Fe- and Al-oxides. The relationship between mineral particle density and soil oxides would be further improved when the dithionite extraction method is used, since this method extracts both amorphous and crystalline Fe- and Al-oxides (McKeague and Day 1966). Carrying out this test would also help distinguish the extent to which variation in particle density arises from inherent characteristics of the soil parent material, compared to active processes of soil formation. 2.5 Conclusions Precision of the gauge was an important factor influencing the accuracy of the test for the volume-constant GP used in this study. The uncertainty of the pycnometer decreased with increasing volume of the solid: the minimum uncertainty was achieved when the sample chamber was filled with solids, and the uncertainty of the test with the porous soils was around 0.35% (v/v) for the small-sample GP we used. Soil organic matter accounted for 40% of the variation in soil particle density, while more variation was caused by differences in soil mineralogy. Soil particle density for samples collected from five geographic locations in interior BC was related to SOM with varying significance. Oxalate-extractable Fe- and Al-oxides and PSD were related to mineral particle density with variable accountability, while stratifying the samples according to their geographic locations did not improve the relationship. Soil organic matter and Fe- and Al-oxides were useful soil properties in relation to soil particle density, while stratification of samples according to their geographic locations improved the relationship. Particle density of mineral soils collected in interior BC varied substantially (i.e., from 2.36 to 2.94 Mg m-3), and samples from the five geographic locations were grouped into three groups with significantly different particle density values. Findings of this study will enhance the precision of determination of soil porosity, PSD, and other particle density-related properties of forest soils in BC.  54 Table 2.1 Properties of the small-sample gas pycnometer.  Property†  vt (cm3) vc (cm3) ϕa ϕb λ μ Value 12.39 36.89 < 0.64 <0.50 0.34 2.04 †vt, the volume of the tank; vc, the volume of the sample chamber; ϕ, filling factor, the ratio of volume of (a) calibration discs and (b) soil sample to vc; λ, geometric factor, the ratio of vt to vc; μ, pressure factor, the ratio of the minimum pressure to the maximum pressure encountered in the test.  55 Table 2.2 Uncertainty of the small-sample gas pycnometer with the change of volume of the solids.                                                                                 Volume of solids (cm3)  0 1.812 7.211 9.543† 14.102‡ 18.485§ 36.891¶ Uncertainty - volume (cm3) 0.096 0.089 0.070 0.064 0.050 0.038 0.006 Uncertainty - percentage (%) ∞ 4.92 0.98 0.67 0.35 0.21 0.01 †minimum volume of soil samples tested. ‡average volume of soil samples tested. §maximum volume of soil samples tested. ¶chamber was filled with solids.  56 Table 2.3 Comparison of particle density values obtained by the liquid pycnometer (LP) and the small-sample gas pycnometer (GP) for three types of commercial sands (n=3).  Particle density (Mg m-3) Sand type† LP Small-sample GP White sand No.1 2.656 (0.003‡) 2.646 (0.005) Black abrasive sand 3.028 (0.013) 3.032 (0.019) White sand No.2 2.470 (0.002) 2.465 (0.003) †white sand No. 1 was produced by the Lane Mountain Company, Valley, Washington; it was 99% pure silica quartz which varied in size and had a particle density of 2.6 Mg m-3. White sand No.2 was a standard sand with uniform size and shape produced by the Ottawa Silica Sand Co., Ottawa, IL. The black abrasive sand was an unpurified sand type that varied in size produced by the Kleen Blast Abrasive Warehouse, S.W. Hayward, CA. ‡standard deviation.  57 Table 2.4 Selected properties for soils collected from eight forest sites in interior British Columbia (n=283). Property Mean Max Min SD† Soil particle density (Mg m-3) 2.60 2.89 2.17 0.10 Mineral particle density (Mg m-3)  2.67 2.94 2.36 0.08 Soil organic matter (g kg-1) 34.2 214.1 4.6 25.7 Fe-oxide (%) 0.41 1.95 0.12 0.22 Al-oxide (%) 0.25 2.47 0.06 0.24 Clay (g kg-1) 124.1 455.2 5.2 70.7 Fine silt (g kg-1) 70.5 303.2 1.9 37.0 Medium silt (g kg-1) 182.3 359.9 7.9 73.8 Coarse silt (g kg-1) 118.8 300.0 9.8 47.7 Very coarse sand (g kg-1) 356.8 858.0 0.0 191.8 †standard deviation.  58 Table 2.5 Grouping of the mineral particle density for soil samples collected from interior British Columbia forests. Loc. No. Geographic location Latitude and longitude No. observations Particle density (Mg m-3) 2 Prince George & Fort St. James 54°41' N 124°28' W 54°32' N 124°15' W 54°21' N 123°27' W 124 2.73a† (0.039‡) 3 Miriam Creek 50o24’N 118o57’W  27 2.66b (0.096) 4 Okanagan Falls 49o18’N 119o26’W 34 2.64b (0.028) 1 Kamloops  50o43’N 120o25’W 60 2.60c (0.059) 5 Will Lake 50o27’N 119o38’W 35 2.57c (0.047) †values with the same letter are not significantly different at P < 0.05. ‡standard deviation.  59 Table 2.6 Relationships between mineral particle density (Mg m-3) and selected soil properties. Loc. Intercept Coefficients and variables‡ R2 P Overall (283†) 2.402 -0.14 AlO + 0.18 FeO – 0.0004 MSI + 0.0015 FSI + 0.0004 VCS + 0.0003 CSI + 0.0002 CL 0.62 < 0.0001 1 (60) 2.866 0.0009 FSI – 0.0009 MSI – 0.0006 CSI -0.06 AlO 0.71 < 0.0001 2 (126) 2.709 0.05 FeO + 0.0001 CL – 0.0007 MSI+ 0.0007 CSI + 0.0003 FSI 0.27 < 0.0001 3 (27) 2.721 -0.14 AlO 0.85 < 0.0001 4 (35§) 2.677 - 0.25 AlO 0.43 < 0.0001 5 (35) 2.431 0.0003 VCS 0.16 0.0001 †number of samples. ‡AlO and FeO (Al- and Fe-oxides), in %; CL (clay), FSI (fine silt), MSI (medium silt), CSI (coarse silt), and VCS (very coarse sand), in g kg-1. See Table 2.5 for group information. §one sample had no data of particle size distribution.  60 Table 2.7 Relationships between soil particle density (Mg m-3) and selected soil properties. Loc. Intercept Coefficients and variables‡ R2 P Overall (280†) 2.435 -0.0025 SOM + 0.17 FeO – 0.11 AlO – 0.0003 VCS + 0.0012 FSI + 0.0002 CL– 0.0002 MSI + 0.0002 CSI 0.77 < 0.0001 1 (60) 2.888 - 0.0007 MSI – 0.0028 SOM – 0.0007 CSI + 0.0008 FSI 0.81 < 0.0001 2 (124) 2.722 -0.0027 SOM + 0.06 FeO + 0.0002 CL – 0.0004 MSI + 0.0004 CSI 0.87 < 0.0001 3 (26) 2.740 -0.0031 SOM - 0.09 AlO 0.94 < 0.0001 4 (34) 2.683 -0.0026 SOM – 0.22 AlO + 0.0001 CL 0.96 < 0.0001 5 (35) 2.409 -0.0012 SOM + 0.0003 VCS 0.46 < 0.0001 †number of samples; one sample from loc. 4 had no data of particle size distribution. ‡AlO and FeO (Al- and Fe-oxides), in %; SOM (Soil organic matter), CL (clay), FSI (fine silt), MSI (medium silt), CSI (coarse silt), and VCS (very coarse sand), in g kg-1. See Table 2.5 for group information.  61  Figure 2.1 Diagram of a constant-volume gas pycnometer.  62   Figure 2.2 Photo of the small-sample, constant-volume gas pycnometer.  63  Figure 2.3 Discs used for the calibration process of the small-sample gas pycnometer.  64 Reciprocal of gauge reading (PSI-1) 0.08 0.10 0.12 0.14 0.16 0.18 0.20 Vo lu m e of  s ol id  (c m 3 ) 0 5 10 15 20 25 30 ***000.1,280.49770.247 2 =+−= Rxy  Figure 2.4 Calibration model obtained for the small-sample gas pycnometer. ***significant at P < 0.001.  65 Volume of solid (cm3) 0 5 10 15 20 25 30 M in im um  v ol um e se ns ed  b y th e py cn om et er  (c m 3 ) 0.00 0.02 0.04 0.06 0.08 0.10 ***000.1,10839.310907.310749.9 22532 =×+×−×= −−− Rxxy  Figure 2.5 Sensitivity analysis of the small-sample pycnometer. ***significant at P < 0.001.  66 y = -0.0009x + 2.70 n=283, R2 = 0.08 2.30 2.40 2.50 2.60 2.70 2.80 2.90 3.00 0 50 100 150 200 Soil organic matter (g kg-1) P ar tic le  d ne si ty  (M g m -3 ) a y = -0.0028x + 2.69 n=280, R2 = 0.40*** 2.30 2.40 2.50 2.60 2.70 2.80 2.90 3.00 0 20 40 60 80 100 120 Soil organic matter (g kg-1) P ar tic le  d ne si ty  (M g m -3 ) b  Figure 2.6 Relationship between (a) mineral particle density and soil organic matter, and (b) soil particle density and soil organic matter with soil organic matter content < 120 g kg- 1 for forest soils from the interior British Columbia. ***significant at P < 0.001.  67 2.30 2.40 2.50 2.60 2.70 2.80 2.90 0 20 40 60 80 100 120 Soil organic matter (g kg-1) P ar tic le  d en si ty  (M g m -3 ) Loc. 1: y=-0.0032x + 2.64, n=60, R2=0.36*** Loc. 2: y=-0.0026x + 2.75, n=124, R2=0.80*** Loc. 3: y=-0.0042 + 2.73, n=26, R2=0.70*** Loc. 4: y=-0.0027x + 2.66, n=35, R2=0.87*** Loc. 5: y=-0.0012x +2.55, n=35, R2=0.33***  Figure 2.7 Relationships between soil particle density and soil organic matter content grouped by sampling locations. Loc. 1, Kamloops; loc. 2, Prince George and Fort St. James; loc. 3, Miriam Creek; loc. 4, Will Lake; loc. 5, Okanagan Falls. ***significant at P < 0.001.  68 2.6 References Adams, W.A. 1973. The effect of organic matter on the bulk and true densities of some uncultivated podzolic soils. J. Soil Sci. 24:10–17. American Society for Testing and Materials. 2000. D5550-00: Standard test method for specific gravity of soil solids by gas pycnometer. West Conshocken, PA. Baldock, J.A., and Skjemstad, J.O. 1999. Soil organic carbon/soil organic matter. In Soil Analysis – An Interpretation Manual. Edited by K. Peverill, D. Reuter, and L. Sparrow. CSIRO Publishing, Melbourne, Vic. Australia. pp. 159-170. Bielders, C.L., De Backer, L.W., and Delvaux, B. 1990. Particle density of volcanic soils as measured with a gas pycnometer. Soil Sci. Soc. Am. J. 54:822–826. Caldwell, P.V., Vepraskas, M.J., and Gregory, J.D. 2006. Physical properties of natural organic soil in Carolina Bays of the Southeastern United States. Soil Sci. Soc. Am. J. 71:1051-1057. Danielson, R.E., and Sutherland, P.L. 1986. Porosity. In SSSA book series 5, Methods of soil analysis. Part 1, Agron. Monogr. 9. 2nd ed. Edited by A. Klute. ASA and SSSA, Madison, WI. pp. 443–461. Flint, A.L., and Flint, L.E. 2002. Particle density. In SSSA Book Ser. 5, Methods of soil analysis. Part 4, Physical methods. Edited by J.H. Dane and G.C. Topp. SSSA, Madison, WI. pp. 229–240. Geddis, A.M., Guzman, A.G., and Bassett, R.L. 1996. Rapid estimate of solid volume in large tuff cores using a gas pycnometer. NUREG/CR-6457, Office of Nuclear Regulatory Research, US Nuclear Regulatory Commission. Washington, DC. Gee, G., and Or, D. 2002. Particle-size analysis. In Methods of soil analysis. Part 4. Physical methods. Edited by J.H. Dane and G.C. Topp. SSSA Book Series no. 5. SSSA, Madison,  69 WI. pp. 255-293. Grossman, R.B., and Reinsch, T.G. 2002. Bulk density and linear extensibility. In Methods of soil analysis. Part. 4. SSSA. Edited by J.H. Dane & G.C. Topp. Madison, WI. pp. 201-254. Heiskanen, J. 1992. Comparison of three methods for determining the particle density of soil with liquid pycnometers. Commun. Soil Sci. Plant Anal. 23:841-846. Hillel, D. 1980. Fundamentals of soil physics. Academic Press. NY. 413 p. Holland, S.S. 1964. Landforms of British Columbia, a physiographic Outline. B.C. Dept. Mines and Pet. Res. Bull. 48. Kellogg, R.M., and Wangaard, F.F. 1969. Variation in the cell-wall density of wood. Wood Fiber Sci. 1:180-204. Kummer, F.A., and Cooper, A.W. 1945. Soil porosity determination with the air pycnometer as compared with the tension method. Agric. Eng. 26:21-23. Loeppert, R.L., and Inskeep, W.P. 1996. Iron. In Methods of soil analysis. Part 3 - Chemical methods. Edited by J.M. Bigham. SSSA, Madison, WI. pp. 639-664. McIntyre, D.B., Welday, E.E., and Baird, A.K. 1965. Geologic application of the air pycnometer:a study of the precision of measurement. Geol. Soc. Am. Bull. 76:1055-60. Marinder, B.O. 1996. A simple apparatus for determining the density of solids. Meas. Sci. Technol. 7:1569–1573. McBratney, A.B., Minasny, B., Cattle, S.R., and Vervoort, R.W. 2002. From pedotransfer functions to soil inference systems. Geoderma. 109:41-73. Nelson, D.S., and Sommers, L.E. 1996. Total carbon, organic carbon, and organic matter. In Methods of soil analysis. Part 3 - Chemical methods. Edited by J.M. Bigham. SSSA, Madison, WI. pp. 961-1010.  70 Oades, J.M. 1993. The role of biology in the formation, stabilization, and degradation of soil structure. Geoderma. 56:377-400. Say, H. 1797 D’un instrument propre à mesurer le volume des corps, sans les plonger dans aucun liquide. Ann. Chim. 23:1–27. Sheldrick, B.H., and Wang, C. 1993. Particle size distribution. In Soil sampling and methods of analysis. Edited by M.R. Carter. CSSS, Lewis Publishers, Boca Raton, FL. pp. 499-512. Six, J., Bossuyt, H., Degryze, S., and Denef, K. 2004. A history of research on the link between (micro)aggregates, soil biota, and soil organic matter dynamics. Soil & Tillage Res. 79:7-31. Smettem, K.R.J. 2006. Particle density. In Encyclopedia of Soil Science. Edited by Lal, Rattan. 1:1, 1243-1244 (URL: http://www.informaworld.com/smpp/content~content=a740174511~db=all~order=title). Smith, C.W., Johnston, M.A., and Lorentz, S. 1997. Assessing the compaction susceptibility of South African forestry soils. II. Soil properties affecting compactibility and compressibility. Soil Tillage Res. 43:335-354. Sojka, R.E. 1988. A review:Measurement of root porosity (volume of root air space). Environ. Exp. Bot. 28:275-280. Tamari, S. 2004. Optimum design of the constant-volume gas pycnometer for determining the volume of solid particles. Meas. Sci. Technol. 15:549-558. Van Keulen, J. 1973. Density of porous solids. Materials and Structures. 6:181-183. Valentine, K.W.G., Sprout, P.N., Baker, T.E., and Lavkulich, L.M. (eds.) 1978. The soil landscapes of British Columbia. Resources Analysis Branch, B.C. Ministry of Environment, Victoria, BC. 197 p. Zhao, Y.H., Krzic, M., Bulmer, C.E., and Schmidt, M.G. 2008. Maximum bulk density of  71 British Columbia forest soils from the proctor test:Relationships with selected physical and chemical properties. Soil Sci. Soc. Am. J. 72:442-452.  72 3 MAXIMUM BULK DENSITY OF BRITISH COLUMBIA FOREST SOILS FROM THE PROCTOR TEST: RELATIONSHIPS WITH SELECTED PHYSICAL AND CHEMICAL PROPERTIES2 3.1 Introduction Mechanized forest harvesting operations apply heavy weights to soil, which often lead to compaction. Reduced tree volume and height growth caused by compaction have been reported in various parts of North America (Wert and Thomas, 1981; Page-Dumroese et al., 1998) and it can take decades (as long as 70 years) for compacted soils to naturally recover to their pre- disturbance conditions (Froehlich et al., 1985; Miller et al., 1996). Compaction is a process of increasing the soil bulk density (and decreasing porosity) by application of mechanical forces to the soil. Successful planning to minimize compaction depends on knowledge of the distribution of soil types in a given area, coupled with a knowledge of the behavior of the soils in response to compactive effort. Many regions of North America and elsewhere have extensive soil resource inventories, but work on the site-specific effects of compaction needs to be better developed. To maintain sustainable soil productivity, it is necessary to assess the ability of a soil to support plant growth after machines have traveled over it. Soil scientists studying sustainability have traditionally measured bulk density as an indicator of compaction, and this measurement is also made because the bulk density is a key soil property for determining site nutrient contents. Despite this, limiting values for bulk density have not been defined for the wide range of soil  2A version of this chapter has been published. Zhao, Y.H., Krzic, M., Bulmer, C.E., and Schmidt, M.G. (2008) Maximum Bulk Density of British Columbia Forest Soils from the Proctor Test: Relationships with Selected Physical and Chemical Properties. Soil Sci. Soc. Am. J. 72: 442-452.  73 conditions typical in forests, primarily because such limiting values are different for soils with varying texture, organic matter, and other properties. Establishment of limiting values would be beneficial for soil scientists and land managers. One approach to better evaluate the state of soil compaction among soil types involves expressing the actual bulk density as a percentage of some reference compaction state (Lipiec et al., 1991, Topp et al., 1997; Lipiec and Hatano, 2003). The idea of comparing soil physical conditions on field sites to a reference state was also proposed by Joosse and McBride (2003), who proposed comparisons based on the void ratio to evaluate soil quality of agricultural sites. Such comparisons, would allow conditions from a wide range of soil types to be evaluated using a single threshold limit, much as the critical limits of soil mechanical resistance and air-filled porosity appear to be relatively independent of soil type (Hakansson and Lipiec, 2000; Zou et al., 2001). Therefore, use of a reference state could potentially enhance interpretations in soil compaction studies. Various parameters for a reference compaction state have been proposed (Carter, 1990; da Silva et al., 1994; Hakansson and Lipiec, 2000), but the maximum bulk density (MBD) determined by the standard Proctor compaction test (ASTM, 2000) is rigorously defined, readily determined with standard test equipment, and has been used in several studies (Carter, 1990; Smith et al., 1997; Aragon et al., 2000). The potential advantages of using MBD as a reference compaction state can only be realized if the soil samples reliably represent site conditions, and this can create challenges in forest soils. Unlike agricultural soils, where soil type is often relatively consistent within a particular field, the properties of forest soils are known to vary widely across short distances (Courtin et al., 1983) in response to more variable topography on many forested sites, and the absence of tillage to mix and homogenize surface layers. Such variation would require large number of samples to be taken to determine MBD, and some alternative method to predict MBD would be beneficial.  74 The standard Proctor method (ASTM, 2000) evolved from studies by civil engineers (Proctor, 1933) on the compaction of soils for dam and road foundations. Two parameters are obtained from this method: MBD and the critical water content (WMBD) at which MBD is achieved for a given amount of energy. The compactive force applied in the Proctor test as it is used in engineering studies has evolved over the years to make it more applicable to changing needs. Despite this no information is currently available to determine whether different levels of applied force would improve interpretations of compaction effects on growth (Hakansson and Lipiec, 2000).  Therefore the standard Proctor test is commonly used in productivity studies (Carter, 1990). The variation in MBD as determined by the standard Proctor test for a range of soils has been attributed to changes in soil organic matter content, particle size distribution (PSD), Fe- and Al-oxides, or plastic and liquid limits. For example, quantity as well as quality of organic matter has been identified to have effects on MBD (Soane, 1990; Aragon et al. 2000), and both organic C (Donkin, 1991; Smith et al., 1997; Krzic et al., 2004) and readily oxidisable organic matter (Ball et al., 2000) have been used to predict MBD. Cementing agents, such as Fe-, Al-, or Mn-oxides (in acidic soils) and carbonates (in calcareous soils) enhance aggregate stability, contributing to high soil shear strength (Yee and Harr, 1977). Dorel et al. (2000) reported that Caribbean Andosols and Nitisols (ISSS/ISRIC/FAO, 1998) or Andisols and Alfisols (according to the Soil Survey Staff, 2006) were more resistant to compaction because of the presence of stable microaggregates containing halloysite and Fe-oxide. Larson et al. (1980) found that among 36 agricultural soils from around the world, soils with predominantly kaolinite or Fe- oxide in the clay fraction had lower MBD than soils with predominantly 2:1 type clays. The MBD was significantly correlated with clay, fine silt, coarse silt, medium sand and fine sand, and the clay+silt fraction had the strongest (inverse) correlation with MBD (R=0.79) on 26 South African forest soils (Smith et al., 1997), while Nhantumbo and Cambule (2006)  75 also showed a relationship between clay content and MBD. Peakedness (kurtosis) and symmetry (skewness) of the PSD curve have also been suggested as parameters in predicting MBD (Webster and Oliver, 1990). Well-graded soils, as indicated by a low coefficient of kurtosis, tend to have higher MBD.  A linear relationship between MBD and kurtosis (R2=0.82) was reported by Moolman (1981), while Smith et al. (1997) showed that MBD decreased as degree of kurtosis increased, but the relationship was not strong (R=-0.48). Particle density of mineral soils dominated by Fe-oxides and heavy minerals can range from 3.0 to 5.0 Mg m-3 (Padmanabhan and Mermut, 1995; Ruhlmann et al., 2006), while organic soil may have a particle density as low as 0.84 Mg m-3 (Redding and Devito, 2006). Our particle density study (Chapter 2) showed that forest soils collected from interior BC varied significantly in particle density of mineral, from 2.57 Mg m-3 in the south interior to 2.73 Mg m- 3 in the north-central interior. Since soil bulk density is influenced by particle density, consideration of particle density should be made when predicting MBD. This is particularly important when the soils examined have a wide variation in particle density, as they may for groups of soils developed on diverse geologic materials in mountainous terrain. Due to the complex inter-relationships among soil properties, attempts have also been made to combine several soil properties when predicting MBD. For example, variation in MBD was predicted well by liquid limit, organic C, and sand (R2=0.98) in a study carried out by Howard et al. (1981) on 14 forest and rangeland soils in California. Similarly, Ball et al. (2000) showed that MBD, WMBD, and total porosity at MBD were predicted (R2=0.49, 0.55, and 0.43, respectively) by a combination of liquid limit and readily oxidisable organic matter for a range of cultivated soils in Great Britain. Based on these findings, and as part of a larger study carried out throughout British Columbia to determine effects of compaction on forest soil productivity and tree growth, we evaluated the potential for predicting MBD based on properties that can be determined on  76 samples normally collected during field evaluations of bulk density on forested sites. Our objectives were to (1) evaluate relationships between MBD and WMBD, determined by the standard Proctor test, and other soil properties for a wide range of BC forest soils, (2) identify the soil properties most important for predicting MBD, and (3) describe a proposed method for using MBD as a reference bulk density in forest soil compaction studies. Hypothesis tested was that soil MBD and WMBD could be predicted by soil physical and chemical properties. 3.2 Materials and methods 3.2.1 Study sites A total of 147 soil samples were collected from 33 study sites (Table 3.1) located in timber-growing areas within the Boreal White and Black Spruce (BWBS), Sub-Boreal Spruce (SBS), Interior Douglas-fir (IDF), Interior Cedar-Hemlock (ICH), Coastal Douglas-fir (CDF), and Coastal Western Hemlock (CWH) biogeoclimatic zones of British Columbia (Meidinger and Pojar, 1991). Ninety-three samples from 16 of the study sites were included in a previous study by Krzic et al. (2004). The most common soil textural classes were silt loam and loam, with substantial variation often occurring within study sites. Soils were classified as Inceptisols or Brunisols and/or Gleysols (according to the Soil Classification Working Group, 1998), Alfisols or Luvisols (Soil Classification Working Group, 1998), and Spodosols or Podzols (Soil Classification Working Group, 1998), which covered the range of pedogenetic development in British Columbia. The majority of the soils were developed on glacial till, with the exception of one Inceptisol in the ICH developed on colluvium, two Alfisols in the SBS and ICH zones developed on lacustrine parent material, and seven Inceptisols in the CDF and CWH zones developed on glaciomarine parent material (Table 3.1). The sites sampled for this study included 12 long-term soil productivity (LTSP)  77 installations, three long-term landing rehabilitation trials, seven provincial park sites, five oil- exploration-disturbed sites, four road rehabilitation sites, and two stumping-disturbed sites (Fig. 3.1). The LTSP sites in British Columbia are part of the North American LTSP network that includes the Department of Agriculture, Forest Service, Canadian Forest Service, British Columbia Ministry of Forests and Range, and various universities and industry groups (Powers, 2006). Sample locations were selected to be representative of typical site conditions. For the LTSP sites, sample locations were harvested with minimal soil disturbance. Soils from landing and oilfield rehabilitation trials usually experienced some scalping of surface layers, while soils from the stumping trials were characterized by some mixing of surface soil layers. Samples (ca. 35 kg each) were collected at 0-0.1, 0.1-0.2, or 0-0.2 m depth after removal of the forest floor (if present). The number of samples collected at each site varied between two and 12. 3.2.2 Soil analysis 3.2.2.1 Maximum bulk density The MBD and WMBD were determined using the standard engineering Proctor test (Proctor, 1933; ASTM, 2000). Soil samples were air dried until friable and then passed through a 19-mm sieve followed by sieving through a 4.75-mm sieve and further air dried. To carry out the test, an initial estimate was made for each sample of the water content at which MBD would be achieved. Because WMBD is typically slightly less than the plastic limit, the initial estimate involves determining the water content of a sample that has been moistened to the point where a sample that is squeezed in the hand will remain in a lump when hand pressure is released, but breaks cleanly into two pieces when “bent” (ASTM, 2000). Water was then added to a 2.3 kg sub-sample until it reached the estimated water content, and then four more sub-samples were prepared, two with soil water content (W) below, and two with W above this value. The five sub- samples were then left in sealed plastic bags to equilibrate overnight. During the test, soil was  78 compacted in a standard mold (9.43×10–4 m3) using a 2.5-kg rammer falling freely from a height of 0.3 m. The soil was added to the mold in three layers and 25 blows of the rammer were applied to each layer. Total compactive effort applied to the sample was approximately 600 kN m m–3 (or 595 kJ m–3). The compacted sample was used to determine bulk density and corresponding water content. Soil water content was determined gravimetrically (w/w) by drying samples at 105°C for 16 h. Dry bulk densities vs. W values were plotted on a graph and the points were fitted with a best-fit curve (third order polynomial). From the resulting compaction curve, MBD was determined from either (1) the peak of the curve or (2) the highest sample value when the peak of the curve lay below that level. Approximately 0.5 kg from each sample was sieved through a 2-mm sieve to determine the percentage of fine fraction, which was used to correct MBD. The volume of mineral coarse fragments was determined from dry mass and assumed to have particle density of 2.65 Mg m-3. Fine fraction MBD was calculated as the mass of dry, coarse-fragment-free mineral soil per volume, where volume was also calculated on a coarse fragment free basis. All MBD values are reported on fine fraction basis. 3.2.2.2 Particle density Particle density was determined by the gas displacement method (Flint and Flint, 2002) that was modified so that the expansion chamber (instead of sample chamber) was pressurized to 239 kPa with room air. The expansion chamber was then opened to the sample chamber. Volumes of the two chambers, hose, and transducer were not measured directly: instead, a model describing the volume-pressure relationship was derived based on the changing volume of the sample chamber with known-volume plastic discs. Soil samples < 2 mm, oven dried at 60°C for 48 h, were used for the test. The samples used for determination of particle density were not treated to remove organic matter, so the reported particle density for each sample reflected an average particle density for the entire fine fraction, including organic and mineral components.  79 3.2.2.3 Soil organic matter Soil total C was determined by the dry combustion method (Nelson and Sommers, 1996) using a LECO analyzer on a sample that had passed through a 2-mm sieve. The estimate of the “readily oxidisable organic matter” was obtained as the weight difference before and after treatment of the sample with hydrogen peroxide (H2O2) and expressed as the gravimetric fraction of the original mass of soil. Hydrogen peroxide treatment involved heating a 50 g sample with 75 mL of H2O2 (30%) and 300 mL of water to 80oC and adding small increments of H2O2 until no further reaction was observed. 3.2.2.4 Soil oxides Soil oxides of Al, Fe, Mn, and Si were extracted by 0.2 mol L-1 acid ammonium oxalate solution. This method extracts active Al-, Fe-, Mn-, and Si-oxides including a fraction of oxides bound by organic matter (Loeppert and Inskeep, 1996). The extracted ions were measured by inductively coupled plasma (ICP) spectrometer. 3.2.2.5 Particle size distribution Soil PSD was determined by the hydrometer method (Gee and Or, 2002). Samples were pre-treated with hydrogen peroxide (30%) and heat, while samples from the IDF zone that may have contained carbonates were also treated with a sodium acetate buffer. Particle size distribution was described using the Canadian System of Soil Classification (Sheldrick and Wang, 1993) in terms of the percentage of clay (<0.002 mm), fine silt (0.002-0.005 mm), medium silt (0.005-0.02 mm), coarse silt (0.02-0.05 mm), very fine sand (0.05-0.10 mm), fine sand (0.10-0.25 mm), medium sand (0.25-0.50 mm), coarse sand (0.50-1.00 mm), and very coarse sand (1.00-2.00 mm). Peakedness (kurtosis) and symmetry (skewness) of the PSD curve were calculated using the equations proposed by Webster and Oliver (1990): skewness: Y1 = M3/(M23/2)                                                                   [1]  80 kurtosis: Y2 = (M4/M22)-3                                                                   [2] where: M2 = (1/n)∑(Xi –μ)2 is the 2nd moment of the distribution about the mean of the observation; M3 = (1/n)∑(Xi –μ)3 is the 3rd moment of the distribution about the mean of the observation; M4 = (1/n)∑(Xi –μ)4 is the 4th  moment of the distribution about the mean of the observation; n is the number of observations; Xi is the ith observation, and μ is the mean of the observations. 3.2.2.6 Plastic and liquid limits Plastic limit was determined as the gravimetric water content at which a soil sample could be rolled by hand into a thread of 3 mm diameter without breaking (McBride, 2002). Liquid limit was determined using the one-point Casagrande method (McBride, 2002). The soil water content, at which 20 to 30 blows of the cup are required to close a groove along a distance of 13 mm, was determined gravimetrically after drying at 105°C for 16 h. The liquid limit is calculated using the following equation: LL = W ×(N/25)0.12                                                                                                        [3] where LL is the liquid limit, N is the number of blows, and W is the gravimetric water content. 3.2.3 Statistical analysis Distribution of the data was summarized by principal component analysis (PCA) using the SAS PRINCOMP procedure (SAS Institute, 1990). Prior to the PCA, “missing values” in the data matrix (i.e., 34 samples without plastic limit and three without liquid limit) were filled by the SAS PRINQUAL procedure with the minimum generalized variance method (SAS Institute, 1990). Simple regression analyses between dependent (i.e., MBD, WMBD) and independent variables (i.e., soil properties) were run. Samples without plastic or liquid limit were not included in the regression. In addition, multiple regression analysis was used to select soil properties that were highly correlated with dependent variables. A stepwise method was selected in multiple regression analysis, because certain soil properties (e.g., plastic and liquid limit; total  81 C and oxidisable organic matter) may have multiple impacts on soil MBD. Significance levels of the Chi-square score for variable entry and stay were set at 0.25 and 0.10, respectively. 3.3 Results and discussion The minimum, maximum, and mean values of the soil properties are presented in Table 3.2 (Appendix 2 for MBD, WMBD, and PSD data), while a partial correlation matrix for the relationship between dependant variables (MBD, WMBD) and selected soil properties is presented in Table 3.3. 3.3.1 Relationships between MBD, WMBD and other soil properties 3.3.1.1 Particle density Particle density for soils from our study sites ranged from 2.33 to 2.97 Mg m-3 (Table 3.2), with relatively large variation observed among the sites (27 out of 33 sites had within-site particle density differences higher than 5%). The MBD was significantly (R2=0.36) positively correlated to particle density (Appendix 3). Although compaction effects on forest soil productivity are more directly related to changes in porosity - f (i.e., volume of pores / total soil volume) than bulk density, it is bulk density (and MBD) that is most commonly used to describe forest soil condition (Smith et al., 1997). For groups of soils where particle density varies only slightly, bulk density and MBD can accurately reflect changes in f as they affect plant growth because the change in MBD is inversely related to the volume of pores. However, for groups of soils with a range of particle densities, changes in bulk density and MBD will reflect both differences in volume of solids and particle densities, and the compaction state as it affects growth might be better described by the change in f rather than by the bulk density. To demonstrate the substantial influence of particle density on bulk density, we plotted the Proctor test curve of the soil with the highest particle density (2.97 Mg m-3) then replaced its particle density with the lowest one (2.33 Mg m-3) and  82 plotted the five points again (Fig. 3.2). The derived MBD of the hypothetical soil was 28% lower than that of the soil with the highest particle density (Fig. 3.2a), even though f remained the same (Fig. 3.2b). In forest compaction experiments where the particle density varies widely, determining f of field soils and relating it to f determined at MBD (i.e., fMBD) could serve as an alternative index of the soil compaction state. This is analagous to the approach of Joosse and McBride (2003), who proposed relating the void ratio of structurally intact agricultural soils to that of remolded soils subjected to slurry consolidation and uniaxial compression tests (preconsolidation). Use of f and fMBD could also be advantageous for evaluating compaction effects on soils with different parent materials, as described by Smith et al. (1997). 3.3.1.2 Soil organic matter The MBD was negatively and WMBD positively related to total C and oxidisable organic matter (Table 3.3). The MBD is commonly considered to be linearly related to organic matter (Soane, 1990; Zhang et al., 1997; Krzic et al., 2004). In this study; however, for both total C and oxidisable organic matter, exponential models gave a better description of the relationship (R2=0.70 and 0.64, respectively – Appendix 4) than linear models (R2=0.65 and 0.61, respectively – Appendix 4). Linear models described well the relationships between WMBD and total C and oxidisable organic matter (R2=0.65 and 0.54, respectively – Appendix 5). Greacen and Sands (1980) reported that increased organic matter in sandy soils under radiata pine (Pinus radiata D. Don) forests in South Australia, was associated with reduced compaction (measured as bulk density) under a given load. In a study carried out by Ball et al. (1989) on Gleysol and Cambisol (FAO, 1998) or Inceptisol (Soil Survey Staff, 2006) soils from Scotland, a reduction in MBD of 0.18 Mg m-3 per increase of 10 g kg-1 organic C was observed. In our study, total C had a similar effect on MBD. Oxidisable organic matter is used as an indicator of the soil organic matter quality, since  83 hydrogen peroxide oxidizes the colloidal, humified organic matter, but not the fibrous residues (Day, 1965). The relationship between MBD and oxidisable organic matter has been reported in several studies. For example, Soane (1990) indicated that highly humified material increased soil aggregate stability and soils with high oxidisable organic matter tended to be less compacted. Ball et al. (2000) have shown that oxidisable organic matter explained 63% of variation in MBD of British agricultural soils. In our study, MBD and WMBD were better correlated to total C than oxidisable organic matter (Table 3.3). It has been reported that oxidation of the soil organic matter by hydrogen peroxide is restricted in the presence of Fe, which tends to stabilize soil organic matter during oxidative degradation (Oades and Townsend, 1963). The weaker relationship of MBD and WMBD with oxidisable organic matter than with total C could be attributed to the presence of Fe- oxides, indicating the importance of including several soil properties into a model to increase compactability prediction. On the other hand, Soane et al. (1972) reported lower correlations between MBD and WMBD and organic C (R=-0.72 and -0.61, respectively) compared to oxidisable organic matter (R=-0.81 and -0.74, respectively) for 58 British agricultural topsoils. The authors used sodium dichromate mixtures to test organic C, which did not completely oxidize the organic compounds. In addition, the reaction between soil Fe- and Mn-oxides with dichromate may have further lowered the organic C (Nelson and Sommers, 1996). Organic matter affects the compaction process in at least two ways (1) it increases soil resistance to compaction by enhancing the contact between soil particles (Soane 1990) and (2) its low particle density (Redding and Devito, 2006) compared to mineral particles reduces soil particle density and therefore bulk density, especially when organic matter content is high. For our soils, total C accounted for 30.2% of the variation in particle density. The coefficient of determination for the relationship between fMBD with total C, where the density effect of total C is removed, was 5% lower than that for MBD with total C (Table 3.3). Therefore, we conclude  84 that the organic matter had the strongest effect in improving soil resistance to the compactive force for our group of soils. 3.3.1.3 Soil oxides The MBD was negatively and WMBD positively related to Al- and Fe-oxides (Table 3.3). Linear relationships for Al-oxide with MBD and WMBD were both stronger than those for Fe- oxide (Table 3.3), while adding an exponential component further improved the relationships with Al-oxide (R2=0.53 and 0.44 – Appendices 6 and 7).  In these relatively young soils of British Columbia that have developed since the most recent glaciation, oxides of Fe and Al are the main cementing agents that enhance aggregate stability (McKeague and Sprout, 1975). Addition of Fe-oxides along with organic C, liquid limit, and sand into the prediction model used in a study by Howard et al. (1981) improved the predictability of MBD by 1% (R2=0.99). Contrary to our findings, Fe-oxide was positively related to MBD in the study mentioned above. The authors used the citrate-bicarbonate-dithionite extraction method, which removed the total Fe-oxide. The positive relationship observed in their study appeared to reflect the effect of Fe on soil mass rather than on soil strength, as soil particle density increases with Fe content. The negative relationship between MBD and active Fe-oxide (extracted by ammonium-oxalate) in our study reflected the enhanced soil strength due to the presence of soil oxides. Consequently, testing of active oxides along with total oxides could provide important information to determine the mechanisms by which these materials affect compaction. 3.3.1.4 Particle size distribution Our results show that PSD was not the major factor related to variations in MBD or WMBD over the entire range of soils we studied, even though correlation coefficients were significant between MBD and medium silt, and between WMBD and clay, fine silt, medium silt, fine sand, medium sand, and coarse sand, respectively (Table 3.3). In previous studies, increasing clay content has either been associated with lower MBD (Smith et al., 1997;  85 Nhantumbo and Cambule, 2006) or had little effect on MBD (Ball et al., 2000; Aragon et al., 2000). Kurtosis and skewness were previously considered to be useful parameters in predicting soil MBD (Moolman, 1981). A low coefficient of kurtosis indicates a well-graded PSD and is expected to lead to higher MBD, while a low absolute value of the skewness coefficient indicates high symmetry of the distribution curve and higher MBD. In our study, kurtosis ranged from –1.67 to 3.46 and skewness varied from –0.26 to 2.26, but we found no relationship between MBD or WMBD and kurtosis or skewness (Table 3.4). Smith et al. (1997) suggested that the relationship between MBD and kurtosis would be confounded by the significant relationship between MBD and organic matter, which may also have occurred in our study. 3.3.1.5 Plastic and liquid limits The MBD was negatively and WMBD positively related to liquid and plastic limits (Table 3.3). Liquid and plastic limits have strong linear relationships with MBD (R2=0.72 and 0.87, respectively – Appendix 8) and WMBD (R2=0.78 and 0.89, respectively – Appendix 9). Exponential model provided a better relationship between MBD and plastic limits (R2=0.93 – Appendix 8). Our results also showed that if plastic limit can be determined on a sample, it is more closely related than liquid limit to MBD and WMBD. Plastic and liquid limits integrate several soil properties such as PSD, organic matter content, and clay mineralogy. Our findings are similar to those of Soane et al. (1972) who tested 13 properties of 58 Scottish topsoils and found that MBD and WMBD were highly related to plastic and liquid limits (R=-0.80 and -0.68 for MBD, 0.74 and 0.69 for WMBD). Howard et al. (1981) found that MBD of California forest and rangeland soils was significantly correlated to liquid limit (R=-0.96) but they did not report plastic limit. Ball et al. (2000) reported that liquid limit accounted for 43 and 48% of the variation in MBD and WMBD of British soils, while the relationship between MBD or WMBD and plastic limit was lower (R2=0.29 and 0.36). Relative to  86 our data, the correlation coefficient between MBD and liquid limit was higher in a study by Howard et al. (1981) and lower in Ball et al. (2000). The former study tested 14 Californian soils predominated by loam texture, while the latter evaluated 146 British agricultural soil samples with a wide variation in texture. Differences in the soil textures might have accounted for the difference in correlation coefficients among the above studies. In the study by Ball et al (2000), lower correlation coefficients may have been observed because some non-plastic soils had missing values filled by a statistical tool prior to running the correlation analysis. 3.3.2 Predicting MBD by a set of soil properties Principal component analysis (a multivariate analysis tool to examine relationships among several quantitative variables in a data set) showed that the first three components accounted for 67% of the variation in the data set (Table 3.5). The first component mainly explained MBD, fMBD, and WMBD and soil properties like liquid and plastic limits, total C, and oxidisable organic matter. The second and third components mainly explained soil texture and Al-oxides. This indicated that soil organic matter and liquid and plastic limits had the greatest impact on MBD, fMBD, and WMBD while particle density, Al- and Fe-oxides, and some of the particle size classes were of secondary importance. Principal component analysis allows for a reduction in the number of variables used in regression analysis because it identifies factors whose effects are independent of one another. Multiple regression analysis was carried out to find the best combination of soil properties that would explain variation in MBD. Because it was not possible to obtain the plastic limit for more than 20% of the samples, and considering that the plastic limit was previously shown (Ball et al., 2000) to be an important factor affecting MBD, we carried out separate multiple regression analyses on soil groups based on their plasticity as shown in Fig. 3.3. Soils with high plasticity were characterized by either high clay content (up to 700 g kg-1) or high total C (up to 77 g kg- 1). Moderately plastic soils have lower contents of clay (up to 560 g kg-1) and total C (up to 57 g  87 kg-1), and make up the largest group of soils in our study. Non-plastic soils had the lowest clay content (up to 17 g 100g-1) and variable total C content (from 4 to 63 g kg-1). Generally, it is difficult to determine the plastic limit on the very coarse textured soils that cover some areas of British Columbia. The hypothesis that soil MBD and WMBD could be predicted by soil physical and chemical properties was accepted in the study. We were able to predict MBD of British Columbia forest soils by combining several soil properties (Table 3.6). When non-plastic samples were included, in the overall regression analysis, liquid limit was the most highly correlated property in explaining MBD among all soil properties included in this study. Liquid limit, in combination with clay, explained more than 80% of the variation in MBD. When oxidisable organic matter and Al-oxide were added to the liquid limit and clay, predictability of MBD improved by 8% (Table 3.6). When samples were grouped according to their plasticity, fewer variables were needed in the multiple regressions to explain comparable amounts of variation in MBD, compared to the entire sample set. Multiple regressions for the non-plastic soils explained the most variation, while those for the highly plastic soils explained the least, and soils with low and moderate plasticity were intermediate (Table 3.6). For the non-plastic soils (i.e., those with low clay content), liquid limit and Al-oxide were the two most important properties in predicting MBD (R2=0.96). In the moderately plastic group, plastic limit and oxidisable organic matter were the first two properties entered into the regression to predict MBD (R2=0.89). For highly plastic soils (i.e., those with high clay and organic matter contents), total C and plastic limit explained 87% of the variation in MBD. For the majority of our soils, oxidisable organic matter was preferred to total C when predicting MBD in the multiple regression analysis, illustrating the importance of quality-related (i.e., oxidisable organic matter) rather than quantity-related (i.e., total C) soil organic matter in  88 compaction studies. In a study by Howard et al. (1981), organic C was the most important variable in predicting MBD, but the authors did not determine the active oxides and sub-groups of sand and silt size fractions, hence the importance of oxidisable organic matter was not known. Oxidisable organic matter was the second most important variable in predicting MBD in a study by Ball et al. (2000). In our study, both organic matter (either total C or oxidisable organic matter) and oxides showed importance in the prediction of MBD (Table 3.6). In all groups, organic matter (i.e., total C) showed a strong relationship in decreasing MBD (R2=0.59 to 0.72), which is similar to the results of Smith et al. (1997) who found a strong negative relationship between MBD and total C (R2=0.88), and also to Aragon et al. (2000) who showed a high dependence of MBD on organic C. Including PSD further improved the prediction. Even though clay came second in predicting MBD for the “overall” group, particle size components usually ranked third or lower in the prediction. Unlike Smith et al. (1997) who found a strong relationship between MBD and clay+silt (R2=0.63), there was no relationship between MBD and clay+silt in our non-plastic and moderately plastic groups (Fig. 3.4a, b, c). Only in the highly plastic group did clay+silt show a high correlation with MBD (Fig. 3.4d), but the positive effect we observed was opposite to that reported by Smith et al. (1997). Smith et al. (1997) found that in the lower range of clay+silt (0-400 g kg-1) the effectiveness of clay+silt and total C appear to offset each other in the compaction test; only in the higher range of clay+silt (400-1000 g kg-1) did clay+silt enhance the effect of total C in reducing MBD. Hence, the importance of these two properties cannot be compared directly in their study. As total C was positively correlated to clay+silt (R2=0.33) in their study, the strong effect of clay on MBD may be just a co-varying result of organic matter on MBD. In our study, oilfield rehabilitation trials experienced scalping of surface layers, which resulted in the surface deposition of the subsoil, and samples from these sites were normally low in soil organic matter content. When these samples were coupled with samples from roads and  89 landings with total C < 30 g kg-1, the effectiveness of clay+silt influencing MBD was substantially enhanced (Fig. 3.4e). In surface soils, the effects of pedogenic development generally over-ride the effect of PSD on maximum bulk density. Pedogenic processes in these soils include accumulation of organic matter and development of soil structure through organic or organomineral binding within aggregates following the release of Al and Fe during weathering (Schmidt and Kogel-Knabner 2002). These processes tend to alter the soil material so it has a lower MBD from the Proctor test, compared to subsoils that have been less affected by pedogenesis. For the highly plastic group in our study, there was also a close correlation between clay and total C, which was not apparent in the other groupings (Fig. 3.5), but the relationship we observed showed total C content declining with increasing clay content, opposite to what is commonly expected for a range of soils. Our highly plastic soils group appeared to contain a mix of two subgroups of soils: (1) those with very high clay content but low or intermediate organic matter and (2) those with very high organic matter content but low clay content. This may clarify why multiple regressions explained the smallest amount of variation for the highly plastic group. Because compaction is a dynamic process, the surface area, contacting points, and surface charge tend to be more important than single particles. Soil properties that more directly represent the above mechanisms should give a good description of the compaction process. Plastic and liquid limits have been proven powerful (R2>0.90) in estimating external surface area (Hammel et al., 1983); on the other hand, oxalate extractable oxides reflected the charge condition of particle surfaces. Organic matter may also be more important than PSD where living and dead roots provide a filamentous network, which resist compactive loads, and highly humified material increases the stability of aggregates (Soane, 1990).  90 3.3.3 Proposed method for using MBD as a reference in forest soil compaction studies To use MBD as a reference value in soil compaction studies, a method for obtaining the best estimate of MBD across a variable site is required. The standard approach to determining MBD using the Proctor test relies on collecting a 10 L sample from the site, and carrying out the laboratory test. On typical forestry sites, such a method may be impractical because site variability makes it difficult to identify the “typical” condition which will best represent conditions throughout the site. A better approach would be to collect samples from each variant of the soil conditions, but this too may become unwieldy because of the large number of samples required. Generally, bulk density sampling requires high numbers of samples to account for the natural variation on forestry field sites (Courtin et al., 1983; Page-Dumroese et al., 1999). The method we propose takes advantage of the strong relationships we have observed between MBD and soil properties that are relatively easy to measure. The method involves four steps. 1. Determine the relationships between MBD and properties for soils typical of the study area. As we and others have shown, MBD can be predicted with reasonable accuracy from a relatively small number of properties, but the best properties for prediction may be different for different groups of soils. The properties to use in the prediction of MBD can be selected by stratification of samples from a larger data set, as we have described. We stratified our sample set based on plasticity, but other approaches such as stratification based on pedogenesis or parent material could be applicable in a particular study. 2. Collect bulk density samples from the field sites. 3. Carry out laboratory analyses on the bulk density samples to provide data to be used in multiple regression analysis as we have described here. The analysis will produce a “predicted MBD” for each soil sample that will account for the fine scale variation in soil properties typical of forestry sites. It may be possible to carry out the analysis for different variables than we have described, depending on the needs of the study and the resources  91 available. For example, in the non-plastic soils of our study, the use of total C and Al-oxide explained a large amount of variation in MBD (R2=0.88), though not as much as liquid limit and Al-oxide (R2= 0.96). 4. Develop empirical relationships between field bulk density, MBD, and tree growth. We are conducting further investigations to test the applicability of relative measure of bulk density (i.e., field bulk density/MBD) for compaction studies on forest soils in British Columbia that have been described previously (Carter, 1990; da Silva et al.1997). 3.4 Conclusions The significance levels of single soil properties in predicting MBD were in the order of liquid and plastic limits, organic matter, and oxalate extractable oxides; while PSD alone accounted for very little variation. In the multiple regression analysis for the entire sample set, liquid limit and clay content were related to MBD. Inclusion of organic matter, Al-oxides, and other components of the PSD (e.g., very coarse sand) further improved the prediction of MBD. Stratification of the sample set by plasticity allowed for substantially improved predictions of MBD using multiple regression analysis. The best predictions were obtained for non-plastic soils, while multiple regression explained the least amount of variation for highly plastic samples. Porosity at MBD may be useful for studies relating plant growth to soil physical condition. On the other hand, use of MBD may be preferred over fMBD for evaluating soil conditions where a reference value for soil bulk density is required. Currently, only bulk density is used widely as a parameter to assess the compaction state of a soil. We have described a method to predict MBD from readily measured soil properties that could enable more effective means of providing reference values for compaction studies. This would be particularly beneficial where these attributes exhibit high point to point variation, such as in British Columbia’s forest soils. Prediction would involve first determining the  92 plasticity for a soil sample, then using the appropriate equation to determine MBD.  93 Table 3.1 Site description, biogeoclimatic (BEC) zones, and annual precipitation for 33 study sites throughout British Columbia. Study site BEC† Precipitation (mm) Soil suborder Parent material Black Pines IDF 279 Cryalf Eolian veneer over glacial till Dairy Creek IDF 279 Cryalf Eolian veneer over glacial till Emily Creek IDF 424 Cryepts Glacial till Kiskatinaw BWBS 482 Cryalf Glaciofluvial veneer over Kootenay East IDF 424 Cryepts Glacial till Log Lake SBS 615 Udepts Glacial till McPhee Creek ICH 755 Udepts Colluvium Mud Creek IDF 424 Cryepts Glacial till O’Connor Lake IDF 279 Cryalf Eolian veneer over glacial till Rover Creek ICH 755 Udepts Colluvium Skulow Lake SBS 425 Cryalf Glacial till Topley SBS 530 Cryalf Glacial till Aleza Lake SBS 930 Cryalf Lacustrine Miriam Creek ICH 420 Cryalf Glacial till Vama Vama ICH 601 Cryalf Lacustrine Gates Creek ICH 410 Cryalf Glacial till Phoenix ICH 450 Cryolls Glacial till Aitken BWBS 464 Cryalf Glacial till Bernadet BWBS 498 Cryalf Glacial till Blackhawk BWBS 619 Cryalf Glacial till Blueberry BWBS 489 Cryalf Glacial till Boot Lake BWBS 581 Cryalf Glacial till Apollo SBS 497 Cryalf Glacial till John Prince SBS 565 Udepts Glacial till Weedon SBS 606 Udepts Glacial till Younges SBS 615 Cryalf Glacial till Port Alberni CWH 2116 Udepts Glaciomarine Duncan Eagle CDF 1039 Udepts Glaciomarine  94 Study site BEC† Precipitation (mm) Soil suborder Parent material Duncan Keating CDF 1039 Aquepts Glaciomarine Duncan Somenos CDF 1039 Ustepts Marine or Lacustrine Kennedy Lake CWH 3295 Udepts Glaciomarine Saanich Cowichan CDF 906 Aquepts Glaciomarine Saanich Fairbridge CDF 906 Udepts Glaciomarine † BEC, biogeoclimatic ecosystem classification; IDF, Interior Douglas-fir; BWBS, Boreal White and Black Spruce; ICH, Interior Cedar-Hemlock; SBS, Sub-Boreal Spruce; CWH, Coastal Western Hemlock; CDF, Coastal Douglas-fir.  95 Table 3.2 Soil properties for 33 study sites in British Columbia (n=147). Soil property Min. Max. Mean SD Maximum Bulk Density, Mg m-3 0.91 1.98 1.51 0.23 WMBD†, kg kg-1 0.09 0.50 0.22 0.08 Particle density, Mg m-3 2.33 2.97 2.66 0.10 Total C, g kg-1 1.8 77.4 23.4 16.5 Oxidisable organic matter, g kg-1 2.2 76.7 28.6 15.3 Clay, g 100g-1 1.9 70.3 20.1 13.6 Silt, g 100g-1 9.4 72.8 45.0 11.4 Sand, g 100g-1 2.9 85.1 34.9 17.2 Fe-oxide, % 0.10 1.19 0.51 0.25 Mn-oxide, % 0.002 0.31 0.05 0.04 Al-oxide, % 0.08 1.56 0.32 0.28 Si-oxide, % 0.03 0.66 0.11 0.11 Liquid limit, kg kg-1 0.15 0.61 0.33 0.11 Plastic limit, kg kg-1 0.14 0.57 0.26 0.08 †Water content at which MBD was achieved.  96 Table 3.3 Correlation coefficients for comparisons among 17 selected variables.  Variable† MBD fMBD WMBD PD TC oxOM Clay FSI MSI FS MS CS VCS Al-O Fe-O LL PL MBD 1.00  fMBD -0.98 *** 1.00 WMBD -0.94 *** 0.93 *** 1.00 PD 0.60 *** 0-.42 *** -0.56 *** 1.000 TC -0.81 *** 0.78 *** 0.81 *** -0.55 *** 1.00 oxOM -0.78 *** 0.75 *** 0.74 *** -0.52 *** 0.73 *** 1.000 Clay -0.05 0.11 0.20 * 0.20 * 0.14 0.20 ** 1.00 FSI -0.12 0.13 0.22 ** -0.01 0.14 * 0.25 ** 0.70 *** 1.00 MSI -0.30 *** 0.24 ** 0.26 ** -0.40 *** 0.16 * 0.23 ** -0.26 * 0.13 * 1.00 FS 0.20 * -0.20 * -0.25 *** 0.10 -0.09 -0.24 ** -0.49 *** -0.62 *** -0.49 *** 1.00  97  Variable† MBD fMBD WMBD PD TC oxOM Clay FSI MSI FS MS CS VCS Al-O Fe-O LL PL MS 0.21 ** -0.21 * -0.31 *** 0.13 -0.18 * -0.21 ** -0.50 *** -0.53 *** -0.37 *** 0.77 *** 1.00 CS 0.08 -0.06 -0.20 * 0.10 -0.13 -0.16 * -0.57  -0.56 *** -0.34 *** 0.68 *** 0.77 *** 1.00 VCS -0.02 0.01 * -0.15 -0.04 -0.06 -0.03 -0.58 *** -0.42 *** -0.03 0.31 ** 0.58 *** 0.70 *** 1.00 Al oxide -0.61 *** 0.63 *** 0.59 *** -0.25 ** 0.33 *** 0.26 ** -0.22 ** -0.16 * 0.11 0.13 *** 0.08 0.20 ** 0.14 1.00 Fe oxide -0.40 *** 0.45 *** 0.48 *** -0.05 0.37 *** 0.46 *** 0.26 *** 0.24 ** -0.02 -0.29 *** -0.23 ** -0.19 ** -0.22 * 0.39 *** 1.00 LL -0.85 *** 0.86 *** 0.89 *** -0.42 *** 0.78 *** 0.69 *** 0.42 *** 0.29 *** 0.12 -0.32 *** -0.34 *** -0.24 ** -0.12 0.43 *** 0.36 *** 1.00 PL -0.93 *** 0.92 *** 0.94 *** -0.52 *** 0.85 *** 0.73 *** 0.21 0.18 * 0.14 -0.16 -0.16 * -0.04 -0.08 0.69 *** 0.42 *** 0.90 *** 1.00 *Significant at P < 0.05. ** Significant at P < 0.01. *** Significant at P < 0.001. †MBD, maximum bulk density; fMBD, porosity at MBD; WMBD, water content at which MBD was achieved; PD, particle density; oxOM, oxidizable organic matter; FSI, fine silt; MSI, medium silt; FS, fine silt; MS, medium sand; CS, coarse sand; VCS, very coarse sand; Al-O, Al-oxide; Fe-O, Fe-oxide; LL, liquid limit; PL, plastic limit. Other variables that were not significantly correlated to MBD, fMBD, or WMBD are not shown.  98 Table 3.4 Relationships among maximum bulk density (MBD) and particle size properties obtained at 33 study sites in British Columbia (n=147). Model† R2 p MBD = 1.72 - 0.001 (medium silt) 0.09 0.000 MBD = 1.45 + 0.0006 (fine sand) 0.04 0.021 MBD = 1.45 + 0.001 (medium sand) 0.05 0.009 MBD = 1.54 - 0.03 skewness 0.01 0.335 MBD = 1.51 - 0.005 kurtosis 0.00 0.732 MBD = 1.27 + 0.002 (fine silt) - 0.001 (medium silt) + 0.001 (coarse silt) + 0.002 (medium sand) 0.18 0.000 †fine silt, medium silt, coarse silt, fine sand, and medium sand (g kg-1).  99 Table 3.5 Principal component analysis loadings for the first three components of individual variables (n=147). Loadings‡ Variable† Component 1§ Component 2¶ Component 3# MBD −0.33 −0.15 0.00 fMBD 0.32 0.14 0.05 WMBD 0.34 0.09 0.01 Particle density −0.19 −0.14 0.17 Total C 0.30 0.11 0.00 Oxidisable organic matter 0.29 0.04 0.01 Clay 0.12 −0.38 0.16 Fine silt 0.14 −0.35 0.06 Medium silt 0.13 −0.04 −0.28 Coarse silt −0.03 −0.01 −0.48 Very fine sand −0.08 0.20 −0.41 Fine sand −0.16 0.30 0.19 Medium sand −0.16 0.29 0.29 Coarse sand −0.13 0.33 0.26 Very coarse sand −0.05 0.27 0.15 Kurtosis 0.07 −0.20 0.36 Skewness 0.08 −0.20 0.32 Al-oxide 0.17 0.29 0.05 Fe-oxide 0.20 −0.04 0.01 Mn-oxide 0.16 0.14 0.13 Si-oxide 0.08 0.24 0.01 Plastic limit 0.33 0.08 −0.04 Liquid limit 0.33 −0.01 0.08 †MBD, maximum bulk density; fMBD, porosity at MBD; WMBD, critical water content. ‡Values > 0.25 were italicized. §Accounted for 34% of variation. ¶Accounted for 19% of variation. #Accounted for 14% of variation.  100 Table 3.6 Regression constants and correlation coefficients for relationships between maximum bulk density (MBD) as the dependent variable and selected soil properties as the independent variable. Dependant variable Independent variable † R2‡ Overall (n=144) Intercept Liquid limit Clay Oxidisable organic matter Al-oxide Very coarse sand 2.07 -2.11 0.0006     0.83 2.06 -1.61 0.0006 -0.005    0.88 2.06 -1.29 0.0004 -0.005 -0.17   0.91 MBD, Mg m-3 2.02 -1.35 0.0005 -0.005 -0.16 0.0005  0.92 Non plastic (n=29) Intercept Liquid limit Al-oxide Very coarse sand Total C Clay 2.09 -1.79 -0.14    0.96 2.07 -1.89 -0.12 0.0005   0.97 MBD, Mg m-3 1.98 -1.61 -0.11 0.0006 -0.003 0.0006 0.98 Moderately plastic (n=99) MBD, Mg m-3 Intercept Plastic limit Oxidisable organic matter Medium silt Total C Fine silt Al-oxide  101 Dependant variable Independent variable † R2‡ 2.21 -2.24 -0.004     0.89 2.26 -2.16 -0.004 -0.0003    0.90 2.24 -1.83 -0.003 -0.0004 -0.002   0.91 2.28 -1.79 -0.003 -0.0004 -0.003 -0.0005  0.92 2.27 -1.62 -0.003 -0.0005 -0.003 -0.0005 -0.18 0.92 Highly plastic (n=16) Intercept Total C  Plastic limit MBD, Mg m-3 1.72 -0.004 -0.82    0.87  †Liquid and plastic limits (kg kg-1); Al- and Fe-oxide (%); Oxidisable organic matter and total C (g kg-1); Clay, medium silt, fine silt, fine sand, and very coarse sand (g kg-1). ‡ Significant at p<0.001.  102  Figure 3.1 Location of 33 study sites in British Columbia. Sites 1-3, 5, and 6 are oil- exploration-disturbed sites; sites 4, 7, 11, 15-18, and 22-26 are LTSP installations; sites 13, 14, and 19 are long-term landing rehabilitation sites; sites 20 and 21 are stumping- disturbed sites; Sites 8-10 and 12 are road rehabilitation sites; and Sites 27-33 are provincial park sites.  103 W  (kg  kg -1) 0 .32 0.34 0.36 0.38 0.40 B D  (M g m -3 ) 1 .19 1.24 1.29 1.34 1.39 1.44 1.49 1.54 W  (kg  kg -1) 0 .32 0.34 0.36 0.38 0.40 po ro si ty  (m m -3 ) 0 .474 0.476 0.478 0.480 0.482 0.484 0.486 0.488 partic le  dens ity=2.33 M g m -3 partic le  dens ity=2.97 M g m -3 a b partic le  dens ity=2 .33 M g m -3 P artic le  dens ity=2.97 M g m -3  Figure 3.2 Change of (a) soil bulk density (BD), and (b) porosity with water content (W) in the standard Proctor test. The bottom curve in (a) is hypothetical, derived by replacing the particle density of 2.97 with 2.33 Mg m-3. fMBD  104 Liquid Limit (kg kg-1) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 Pl as tic ity  In de x (% ) 0 10 20 30 40 50 60 moderate & low plastic silts and clays highly plastic silts and clays A-line U-line  Figure 3.3 Plasticity of soils from the study areas, showing highly plastic soils with liquid limit > 0.50 and soils with moderate and low plasticity (liquid limit < 0.50) as plotted on the Casagrande chart. The A-line represents the division between clays (plot on/above the A-line) and silts (plot below the A-line); the U-line refers to the upper limit.  105 a 0 200 400 600 800 1000 M B D  (M g m -3 ) 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 ***03.0,147 )(0003.064.1 2 == ++×−= Rn msifsiclayMBD b 0 200 400 600 800 1000 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 ***22.0,32 )(001.077.1 2 == ++×−= Rn msifsiclayMBD c clay+fsi+msi (g kg-1) 0 200 400 600 800 1000 M B D  (M g m -3 ) 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 ***07.0,99 )(0004.080.1 2 == ++×−= Rn msifsiclayMBD clay+fsi+msi (g kg-1) 0 200 400 600 800 1000 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 d ***36.0,16 )(0005.083.0 2 == ++×+= Rn msifsiclayMBD          Figure 3.4 Relationships among maximum bulk density (MBD) and clay+silt (fsi = fine silt; msi = medium silt) for (a) all samples, (b) nonplastic samples, (c) moderate and low plastic samples, (d) highly plastic samples, and (e) samples from oilfield rehabilitation sites, and roads and landings with total C < 30 g kg-1. ***Significant at P < 0.001. MBD = -0.0003×(clay+fsi+msi) 2  + 0.0297×(clay+fsi+msi) + 1.1691 n=47, R 2  = 0.47*** 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 0 200 400 600 800 1000 clay+fsi+msi (g kg-1) M B D  (M g m -3 ) e  106 total C (g kg-1) 0 20 40 60 80 cl ay  (g  1 00 g- 1 ) 0 20 40 60 80 a ***02.0,147 11.082.17 2 == ×+= Rn totalCclay 0 20 40 60 80 0 20 40 60 80 total C ***05.0,32 05.016.6 2 == ×+= Rn totalCclay b 0 20 40 60 80 cl ay  (g  1 00 g- 1 ) 0 20 40 60 80 ***02.0,99 10.015.20 2 == ×+= Rn totalCclay c total C (g kg-1) 0 20 40 60 80 0 10 20 30 40 50 60 70 80 ***60.0,16 78.052.74 2 == ×−= Rn totalCclay d  Figure 3.5 Relationships between clay and total C for (a) all samples, (b) nonplastic samples, (c) moderate and low plastic samples, and (d) highly plastic samples. ***Significant at P < 0.001.  107 3.5 References Aragon, A., M.G. Garcia, R.R. Filgueira and Y.A. Pachapsky. 2000. Maximum compactibility of Argentina soils from the proctor test: The relationship with organic matter and water content. Soil Tillage Res. 56:197-204. ASTM. 2000. Design D698-00a. Standard test methods for laboratory compaction characteristics of soil using standard effort (12,400 ft-lbf/ft3 (600 kN-m/m3)). ASTM, West Conshohocken, PA. Ball, B.C., D.J. Campbell and E.A. Hunter. 2000. Soil compactibility in relation to physical and organic properties at 156 sites in UK. Soil Tillage Res. 57:83-91. Ball, B.C., R.W. Lang, M.F. O’Sullivan and M.F. Franklin. 1989. Cultivation and nitrogen requirements for continuous winter barley on a Gleysol and a Cambisol. Soil Tillage Res. 13:333-352. Carter, M.R.  1990. Relative measures of soil bulk density to characterize compaction in tillage studies on fine sandy loams. Can. J. Soil Sci. 70:425-433. Courtin, P., M.C. Feller and K. Klinka. 1983. Lateral variability in some properties of disturbed forest soils in southwestern British Columbia. Can. J. Soil Sci. 63:529-539. da Silva, A.P. and B.D. Kay. 1997. Estimating the least limiting water range of soils from properties and management. Soil Sci. Soc. Am. J. 61:877-883. da Silva, A.P., B.D. Kay and E. Perfect.  1994. Characterization of the least limiting water range of soils. Soil Sci. Soc. Am. J. 58:1775-1781. Day, P.R. 1965. Particle fractionation and particle size analysis. In C.A. Black (ed.) Methods of soil analysis. Part 1. Physical and mineralogical properties, including statistics of measurement and sampling. Agron. Monogr. 9. ASA and SSSA, Madison, WI. pp. 545-567. Donkin, M.J. 1991. Loss-on-ignition as an estimator of soil organic carbon in A-horizon forestry  108 soils. Commun. Soil Sci. Plant Anal. 22:233-241. Dorel, M., J., Roger-Estrade, H. Manichon and B. Delvaux. 2000. Porosity and soil water properties of Caribbean volcanic ash soils. Soil Use Manage. 16:133-140. FAO, 1998. World Reference Base for Soil Resources. World Soil Resources Report 84. FAO, Rome. Flint, A.L. and L.E. Flint. 2002. Particle density. p. 229-240. In J.H. Dane and G.C. Topp (eds.) Methods of soil analysis. Part 4. Physical methods. SSSA Book Series no. 5. SSSA, Madison, WI. Froehlich, H.A., D.W.R. Miles and R.W. Robbins. 1985. Soil bulk-density recovery on compacted skid trails in central Idaho. Soil Sci. Soc. Am. J. 49:1015-1017. Gee, G. and D. Or. 2002. Particle-size analysis. p. 255-293. In J.H. Dane and G.C. Topp (eds.) Methods of soil analysis. Part 4. Physical methods. SSSA Book Series no. 5. SSSA, Madison, WI. Greacen, E.L. and R. Sands. 1980. Compaction of forest soils: A review. Aust. J. Soil Res. 18:163-189.  Hakansson, I. and J. Lipiec. 2000. A review of the usefulness of relative bulk density values in studies of soil structure and compaction. Soil Tillage Res. 53:71-85. Hammel , J.E., M.E. Sumner and J. Burema. 1983. Atterberg limits as indices of external surface areas of soil. Soil Sci. Soc. Am. J. 47:1054-1056. Howard, R.F., M.J. Singer and G.A. Frantz. 1981. Effects of soil properties, water content, and compactive effort on the compaction of selected California forest and range soils. Soil Sci. Soc. Am. J. 45:231-236. Joosse, P.J. and R.A. McBride. 2003. Assessing physical quality of plastic soils of differing mineralogy and pre-stress history using mechanical parameters. I. Saturated compression tests. Can. J. Soil Sci. 83:45-63.  109 Krzic, M., C.E. Bulmer, F. Teste, L. Dampier and S. Rahman. 2004. Soil properties influencing compactability of forest soils in British Columbia. Can. J. Soil Sci. 84:219-226. Larson, W.E., S.C. Gupta and R.A. Useche. 1980. Compression of agricultural soils from eight soil orders. Soil Sci. Soc. Am. J. 44:450-457. Lipiec, J., I. Hakansson, S. Tarkiewicz and J. Kossowski. 1991. Soil physical properties and growth of spring barley as related to the degree of compactness of two soils. Soil Till. Res. 19:307-317. Lipiec, J. and R. Hatano. 2003. Quantification of compaction effects on soil physical properties and crop growth. Geoderma. 116:107-136. Loeppert, R.L. and W.P. Inskeep. 1996. Iron. p. 639-664. In D.L. Sparks et al. (eds.) Methods of soil analysis. Part 3. Chemical methods. SSSA, Madison, WI. McBride, R.A. 2002. Atterberg limits. p. 519-527. In J.H. Dane and G.C. Topp (eds.) Methods of soil analysis. Part 4. Physical methods. SSSA Book Series no. 5. SSSA, Madison, WI. McKeague, J.A. and P.N. Sprout. 1975. Cemented subsoils (duric horizons) in some soils of British Columbia. Can. J. Soil Sci. 55:189-203. Meidinger, D.V. and J. Pojar. 1991. Ecosystems of British Columbia. Special Report Series 6. BC Min. of Forests, Victoria, BC. Miller, R.E., W. Scott and J.W. Hazard. 1996. Soil compaction and conifer growth after tractor yarding at three coastal Washington locations. Can. J. For. Res. 26:225-236. Moolman, J.H. 1981. Soil textural properties influencing compactibility. Agrochemophysica 13:13-19. Nelson, D.S. and L.E. Sommers. 1996. Total carbon, organic carbon, and organic matter. p. 961- 1010. In D.L. Sparks et al. (eds.) Methods of soil analysis. Part 3. Chemical methods. SSSA, Madison, WI. Nhantumbo, A.B.J.C. and A.H. Cambule. 2006. Bulk density test as a function of texture for  110 agricultural soils in Maputo province of Mozambique. Soil Tillage Res. 87:231-239. Oades, J.M. and W.N. Townsend. 1963. The influence of iron on the stability of soil organic matter during peroxidation. J. Soil Sci. 14:134-143. Padmanabhan, E. and A.R. Mermut. 1995. The problem of expressing the specific surface-areas of clay fractions. Clays Clay Miner. 43:237-245. Page-Dumroese, D.S., A.E. Harvey, M.F. Jurgensen and M.P. Amaranthus. 1998. Impacts of soil compaction and tree stump removal on soil properties and outplanted seedlings in the northern Idaho, USA. Can. J. Soil Sci. 78:29-34. Page-Dumroese, D.S., M.F. Jurgensen, R.E. Brown and G.D. Mrosz. 1999. Comparison of methods for determining bulk densities of rocky forest soils. Soil Sci. Soc. Am. J. 63:379- 383. Powers, R.F. 2006. Long-term soil productivity: genesis of the concept and principles behind the program. Can. J. For. Res. 36:519-528. Proctor, R.R. 1933. Fundamental principles of soil compaction. Description of field and laboratory methods. Eng. News Record 111:286-289. Redding, T.E. and K.J. Devito. 2006. Particle densities of wetland soils in northern Alberta, Canada. Can. J. Soil Sci. 86:57-60. Ruhlmann, J., M. Korschens and J. Graefe. 2006. A new approach to calculate the particle density of soils considering properties of the soil organic matter and the mineral matrix. Geoderma 130:272-283. SAS Institute. 1990. SAS/STAT user’s guide. Version 6. SAS Inst. Inc., Cary, NC. Schmidt, M.W. and I. Kogel-Knabner. 2002. Organic matter in particle-size fraction from A and B horizons of Haplic Alisol. Eur. J. Soil Sci. 53: 383-391. Sheldrick, B.H. and C. Wang. 1993. Particle size distribution. p. 499-512. In M.R. Carter (ed.) Soil sampling and methods of analysis. CSSS, Lewis Publishers, Boca Raton, FL.  111 Smith, C.W., M.A. Johnston and S. Lorentz. 1997. Assessing the compaction susceptibility of South African forestry soils. II. Soil properties affecting compactibility and compressibility. Soil Tillage Res. 43:335-354. Soane, B.D. 1990. The role of organic matter in soil compactibility: A review of some practical aspects. Soil Tillage Res.16:179-201. Soane, B.D., D.J. Campbell and S.M. Herkes. 1972. The characterization of some Scottish arable topsoils by agricultural and engineering methods. J. Soil Sci. 23:93-104. Soil Classification Working Group. 1998. The Canadian system of soil classification. 3rd ed. Agric. and Agri-Food Canada. Publ. 1646. National Res. Council of Canada, Ottawa, Ontario. 187 p. Soil Survey Staff. 2006. Keys to Soil Taxonomy. Handbook, 10th edition. USDA and NRCS, Washington, 332 pp. Topp, G.C., W.D. Reynolds, F.J. Cook, J.M. Cook and M.R. Carter. 1997. Physical attributes of soil quality. pp. 21-58. In Gregorich E.G. and M.R. Carter (Eds.), Soil quality for crop production and ecosystem health. Elsevier, Amsterdam. The Netherlands. Webster, R. and M.A. Oliver. 1990. Statistical methods in soil and land resources survey. Oxford University Press, Oxford, UK. Wert, S. and B.R. Thomas. 1981. Effects of skid roads on diameter, height, and volume growth in Douglas-fir. Soil Sci. Soc. Am. J. 45:629-632. Yee, C.S. and R.D. Harr. 1977. Effect of wetting mode on shear strength of two aggregated soils in the Oregon coast ranges. USDA Forest Service Research Note, Pacific Northwest Forest and Range Experiment Station. No. PNW-303. Zhang, H., K.H. Hartge and H. Ringe. 1997. Effectiveness of organic matter incorporation in reducing soil compactibility. Soil Sci. Soc. Am. J. 61:239-245. Zou, C., C. Penfold, R. Sands, R.K. Misra and I. Hudson. 2001. Effects of air-filled porosity,  112 soil matric potential and soil strength on primary root growth of radiate pine seedlings. Plant and Soil 236:105-115. 113 4 CHARACTERIZATION OF SOIL COMPACTION AND TREE HEIGHT GROWTH BY RELATIVE BULK DENSITY3 4.1 Introduction The use of heavy machinery in forest management often leads to soil disturbance and compaction, which in turn strongly affect ecosystem stability and site productivity (Froehlich 1979; Wronski and Murphy 1994; Kuan et al. 2007). Soil disturbance and compaction are particularly severe on temporary access areas such as forest landings (areas of cutblocks where harvested trees are processed and loaded onto trucks) and skid trails. These areas may be unproductive unless soil rehabilitation is carried out. Tree species growing on compacted soil are characterized by reduced root elongation rate (Whalley et al. 1995), and sometimes by reduced height growth (Greacen and Sands 1980; Ares et al. 2007; Bulmer et al. 2007). Two possible reasons for variable tree growth response to soil compaction are that compaction did not reach growth-limiting levels in some studies or that compaction indicators were not always successful in describing the relationship between soil compaction and tree growth. Since soil rehabilitation practices are expensive to apply, a better understanding of soil compaction effects on tree growth is needed. Bulk density (BD) has been traditionally used as the most common measure of soil compaction, but establishment of growth-limiting BD thresholds is not straightforward. Any threshold value of BD depends on soil properties (e.g., texture, quantity and quality of organic matter, particle density), site characteristics (e.g., climate), and the criteria used to evaluate when growth is affected. Daddow and Warrington (1983) summarized several studies and reported that growth-limiting BD for sandy loams and loamy sands were near 1.75 Mg m-3,  3A version of this chapter has been submitted for publication. Zhao, Y.H., Krzic, M., Bulmer, C.E., Schmidt, M.G., and Simard, S.W. Characterization of Soil Compaction and Tree Height Growth by Relative Bulk Density. 114 while clay, silty clay loam, silty clay, and silt soils had growth-limiting BD around 1.40 Mg m-3. The root growth-limiting BD values varied from 1.70 - 1.80 Mg m-3 for Douglas-fir (Pseudotsuga menziesii (Mirb) Franco) seedlings grown on sandy loam to loam soils (Heilman 1981), while an artificially created BD of 1.59 Mg m-3 had stopped root penetration of two-year- old Douglas-fir seedlings grown on sandy loam soil in pots (Heninger et al. 2002). Jones (1983), on the other hand, considered the threshold to be a BD level where root growth was reduced to 20% of optimum. These variable results illustrate why a single growth-limiting BD threshold is unrealistic for all situations on all sites. Efforts have been made to identify high-level soil parameters, other than BD, that can integrate several soil properties and relate them to plant growth. One of these parameters was the least limiting water range (LLWR) introduced by da Silva et al. (1994), based on earlier work by Letey (1985). The LLWR describes the range of soil water contents where water availability, soil mechanical resistance, and air-filled porosity do not exceed assigned values associated with growth limitation. Changes in LLWR for a particular soil type are mainly driven by changes in soil compaction (da Silva et al. 1994), hence, determination of the LLWR may not always be necessary to derive valuable information about conditions affecting plant growth in compacted soils. Other high-level integrating soil parameters that were found to correlate well with plant growth include relative bulk density (RBD) and degree of compactness (D). Both parameters represent the proportion of field BD to a reference BD and they only vary in the method used to obtain the reference BD (Eriksson et al. 1974; Pidgeon and Soane 1977; Carter 1990; Hakansson and Lipiec 2000). Relative bulk density was strongly correlated (R2=0.69) to the relative grain yield of spring barley (Hordeum vulgare L.) and spring wheat (Triticum aestivum L.) in a study by Carter (1990) carried out on a fine sandy loam Orthic Humo-Ferric Podzol at Prince Edward Island. An RBD range of 0.77 - 0.84 was associated with a relative grain yield ≥ 115 95%, while RBD > 0.89 corresponded to relative yield < 80%, and at that point aeration porosity was at the critical point to impede growth. Degree of compactness was correlated to spring barley yield on a wide range of soil types in Sweden with clay content between 20 – 600 g kg-1 and organic matter content from 10 – 110 g kg-1 (Hakansson 1990). The author found that the optimal degree of compactness (Dopt) was independent of soil PSD and organic matter content (R2=0.00) and that it was consistently at 0.87. Since the reference BD obtained by the uniaxial test in the study of Hakansson (1990) was 7 - 17% lower than that obtained by the Proctor test in the study of Carter (1990), the Dopt of 0.87 corresponded to an optimal RBD of 0.74 - 0.81. Although RBD has been used successfully to relate soil compaction to growth of annual plant species, its usefulness has not yet been tested for assessment of tree growth in forest ecosystems. Development of such a high-level integrating parameter of soil compaction that can also be successfully related to tree growth will be helpful to guide operational practices, and to assess the viability of rehabilitation to restore productivity to degraded areas (Richardson et al. 1999). The objectives of this study were to: (1) determine RBD for soils on heavily disturbed timber-growing sites such as landings and roads, and (2) assess the relationship between RBD and tree height growth. We also evaluated the influence of thickness of surface organic materials (i.e., wood waste mulches applied to disturbed sites or natural forest floors) on tree height growth. Hypotheses tested were (1) presence of organic matter could mitigate the severity of soil compaction on tree growth; (2) optimal tree height growth was expected within a certain RBD range; and (3) there existed an RBD threshold limiting tree height growth, which warranted the necessity of soil decompaction or rehabilitation. 4.2 Materials and methods 4.2.1 Site description Five experimental sites, including forest landings and roads, were selected throughout 116 interior British Columbia (BC) (Table 4.1). When selecting sites for this study, we focused on those with a broad range of soil disturbance and rehabilitation treatments, and hence, a range of compaction levels (Table 4.2). Experiments 1, 2, 4, and 5 were established to evaluate the effectiveness of tillage and biological inoculation on conifer seedlings (Campbell et al. 2008; Teste et al. 2004), while ectomycorrhizal inoculation did not result in any significant differences in seedling height growth at the time of this study; and experiment 3 was established to determine the effect of tillage and woodwaste amendment on soil rehabilitation (Bulmer et al. 2007). Each experiment was laid out as a randomized complete block design with three blocks. One-year-old nursery-grown seedlings of interior Douglas-fir (Pseudotsuga menziesii var. glauca [Bessin] Franco), lodgepole pine (Pinus contorta Dougl. ex Loud. var. latifolia Engelm.), and hybrid white spruce (Picea glauca [Moench] Voss × engelmannii Parry ex Engelm.) were planted. The species distribution on the experimental sites selected for this study, time for the site establishment, and tree growth measurement are shown in Table 4.2. Planting densities ranged from 2000 - 5000 stems per hectare in experiments 2, 3, and 5. Seedlings were planted in rows on the roads at experiments 1 and 4 and row spacing was typically 2 m, with an intra-seedling distance of 0.5 m. At the time of measurement, interaction between neighbouring trees (i.e., competition) was not considered to be a factor affecting the results. 4.2.2 Field and laboratory methods Three soil BD samples per plot were collected immediately below the interface between the surface organic material (if present) and mineral soil to a depth of 20 cm in the underlying mineral material. On sites with < 25% coarse fragment content (diameter > 2 mm), mineral soil samples for BD determination were collected in 518 cm3 cores. On sites with coarse fragment content > 25%, BD samples were collected by the excavation method (Grossman and Reinsch 2002); water was used to determine the sample volume. Coarse fragments within the sample were screened out and weighed. Volume of coarse mineral fragments was determined from dry 117 mass, assuming a particle density of 2.65 Mg m-3. Bulk densities of mineral soil samples were calculated as the mass of dry, coarse fragment-free soil per volume of field-moist soil, where volume was also calculated on a coarse fragment-free basis. Maximum bulk density was derived using four models developed by Zhao et al. (2008). These models related MBD obtained by the Proctor method (ASTM 2000) to soil physical and chemical properties using a subset of samples from a wide range of sites in BC (n=144). Methods used to determine total C, oxidisable organic matter, oxides of Al and Fe, PSD, and plastic and liquid limits were described by Zhao et al. (2008). For each of the soil samples collected in the current study, plasticity was determined and the appropriate model was used to estimate MBD (Table 4.3). Relative bulk density was determined as the ratio of field BD to the predicted MBD, while three non-cohesive samples with very high liquid limit were excluded in the MBD estimation where the overall model and the specified model (Zhao et al. 2008) provided very different estimates of MBD. With few exceptions, three RBD values at each plot were averaged to represent the RBD of the plot. The highest RBD value (0.92) from an untreated landing of experiment 5 was not included in the analysis because of incomplete tree growth data at that plot. The thickness of surface organic material was measured at each BD sampling location. Tree height growth was measured at the end of the growing season (late September - early October), which corresponded with the time of BD sampling. Tree heights were measured from ground level to the terminal bud for all the live trees present at the plot (17,111 height measurements were used for this study). Height growth was not measured annually on all experiments since seedlings varied in age among experiments (Table 4.2). Varying tree age precluded the use of height increments in our study. We developed a relative height growth index (RHGI) by relating height growth for undamaged trees on our disturbed sites to height growth expected on undisturbed sites with similar ecological characteristics using the SiteTools 118 software (BC Ministry of Forests and Range 2004) as follows: )( )( ),( cmagesametheatSiteToolsbygrowthheight cmmeasuredgrowthheightagespeciesRHGI = where height growth by SiteTools at the measurement age was derived based on the average site index for the measured species and ecosystem. SiteTools provides access to site index and height growth equations, as well as site index conversion equations and growth intercept equations for dozens of British Columbia’s commercial tree species (BC Ministry of Forests and Range 2004). This study took SiteTools’ function of calculating height growth under natural condition with species, age, and site index used as input. The use of RHGI eliminated variation caused by differences in climate and other site conditions among study sites, thus isolating soil compaction effects on tree growth. Where RHGI = 1.0, soil conditions of the rehabilitated site are considered similar to comparable undisturbed sites supporting expected tree growth. Where the index > 1.0, tree growth on the treated areas is better than that in undisturbed conditions. Rehabilitation of disturbed areas may be needed when RHGI < 1.0, because it suggests that trees are growing more slowly than trees on undisturbed sites with comparable ecological conditions. In practice, the effectiveness of rehabilitation is based on the degree to which pre-disturbed conditions are reached; in other words, rehabilitation will be considered unsuccessful if trees on the rehabilitated site are smaller than on undisturbed sites. In this study, we considered a RHGI of 1.0 to represent the growth-limiting RBD threshold. 4.2.3 Data analysis We used SAS REG procedure (SAS Institute 1990) to carry out multiple regression analysis by experiment, with RBD, surface organic material, and derived variables from surface organic material and RBD (e.g., RBD-1, RBD2, surface organic material × RBD) as independent variables, and RHGI as the dependent variable. Thickness of the surface organic material was 119 used as a dummy variable and its value was set at 1 when thickness was ≥ 3 cm and 0 when thickness was < 3 cm. A stepwise method was used to exclude any independent variables that may have overlapping effects on the dependent variable. The χ2 significance level was set at 0.25 for entry of variables and 0.10 for retention of variables, respectively. To derive the relationship between RBD and RHGI, we used the curve fitting functions in SigmaPlot (Systat Software 2000). For the experiment with a poor relationship between RHGI and RBD (i.e., spruce growth on experiment 4), studentized residuals were calculated to remove outliers (SAS Institute 1990) and the new data were fitted with both linear and non-linear equations. Data from the experiments covering the widest RBD range for each tree species were analyzed separately. These included experiment 1 (planted with lodgepole pine) and experiment 2 (planted with Douglas-fir). Experiment 4 was located in north-central BC, which is near the northern limit for Douglas-fir and lodgepole pine (BC Ministry of Forests and Range 2006); data for these two species in this experiment were analyzed separately. 4.3 Results 4.3.1 Relative bulk density and surface organic material The RBD values ranged from 0.48 - 1.01 (Table 4.4), with the highest values in disturbed plots without rehabilitation (e.g., experiments 1 and 4) and the lowest values usually observed on rehabilitated sites (e.g., experiments 2, 3, and 5). The RBD values did not always match treatments as expected. For example, the un-rehabilitated roads at experiment 4 had very low RBD values (0.54 for experiment 4B and 0.67 for experiment 4C), while at experiment 2, a shallow tillage treatment yielded an RBD of 0.94 which was even higher than RBD of some un- rehabilitated plots. The thickness of surface organic material varied from 0 - 16 cm (Table 4.4). Surface organic material was not present on un-rehabilitated plots, with the exception of several un- 120 rehabilitated roads and landings that had developed a thin (1 - 2 cm depth) layer of fan moss (Rhizomnium glabrescens (Kindb) T. Kop.) and Juniper moss (Polytrichium juniperinum Hedw.). In experimental plantations and organic amendment treatments, the surface organic material ranged from 7 - 16 cm in thickness. The relationship between RBD and BD for all soils was best described by a linear regression (Fig. 4.1a), with 64% of variation in RBD explained by BD. A stronger relationship (R2= 0.72) was obtained for cohesive (plastic) soils only, which also tended to have a higher RBD (P < 0.003) than non-cohesive (non-plastic) soils for a given BD. Substantial variation in RBD was observed for the  soils with intermediate BD values (i.e., between 0.80 - 1.40 Mg m-3 of BD, Fig. 4.1a). The very loose soils with both low RBD and low BD values tended to be non- cohesive soils with no surface organic material, while the very compacted soils with both high BD and high RBD values were cohesive soils with no surface organic material (Fig. 4.1a; soil organic matter data not shown). Among the cohesive soils, a wider range of BD was observed for soils that had < 3 cm of surface organic material, compared to those with thicker surface organic material (Fig. 4.1b). Relationships between RBD and BD were significant for both soil groups, while the group with surface organic material < 3 cm had significantly higher RBD values (P < 0.001) than the group with thicker surface organic material (Fig. 4.1b). Surface organic material was the first co-variable used in the multiple regression analysis in the years immediately after planting for several experiments (Table 4.5). For the experiments where surface organic material was strongly correlated to tree height growth, the amount of variation in height growth explained by surface organic material was generally found to decrease over successive growing seasons (Table 4.5). Experiment 3 had a narrow range of RBD, and surface organic material in combination with RBD or RBD2 was positively related to tree growth. For experiments with a wider range of RBD values (i.e., experiments 1, 2, 4, and 5), surface organic material was usually the second variable or was excluded from the regression 121 analysis in the late growing seasons: RBD2 was the main variable negatively related to RHGI (Table 4.5). Strength of surface organic material and RBD in explaining RHGI was not improved when soils were grouped according to the presence of surface organic material (Appendix 10), because grouping narrowed the RBD range and decreased the number of observations. 4.3.2 Relative bulk density and relative height growth index: north-central BC For both interior Douglas-fir and lodgepole pine, RHGI in the seventh growing season was always < 1.0 throughout all the three RBD ranges (RBD < 0.70, 0.70 – 0.83, and > 0.83). This indicates that environmental conditions on the experimental sites were not suitable for optimal growth of these two species (Fig. 4.2). The presence of surface organic material did not affect RHGI of interior Douglas-fir or hybrid white spruce, while it significantly improved RHGI of lodgepole pine at RBD of 0.70 - 0.83 and RBD > 0.83 (Fig. 4.2b). Although higher RHGI of hybrid white spruce was observed for soils with thick versus thin surface organic material throughout the three RBD ranges, the differences were not significant (P > 0.32) (Fig. 4.2c). The RHGI of interior Douglas-fir was higher at low RBD (< 0.70) than at high RBD ranges (0.70 – 0.83, and > 0.83), while no growth differentiation was observed between the two high RBD ranges (P > 0.12; Fig. 4.2a). Height growth of lodgepole pine decreased with increasing of RBD for trees growing in the soils with surface organic material < 3 cm (P < 0.03; Fig. 4.2b) , while no significant growth differentiation was observed in trees growing in soils with thick surface organic material (P > 0.08; Fig. 4.2b). Variation in the RHGI of hybrid white spruce was wider than of interior Douglas-fir and lodgepole pine, and RHGI was not significantly different among the three RBD groups (P > 0.13; Fig. 4.2c). 4.3.3 Relative bulk density and relative height growth index: hybrid white spruce The RHGI of hybrid white spruce decreased linearly (P < 0.001) with increasing RBD in 122 the fourth and seventh growing seasons, but the linear relationships only explained 25 - 27% of the variation (Fig. 4.3a). For both growing seasons, there were non-linear relationships between RHGI and RBD at RBD < 0.75, while no substantial change in RHGI was observed at RBD > 0.75. The RHGI was generally > 1.0 at RBD < 0.75, with two exceptions where RHGI was around 0.5 when RBDs were at 0.67 and 0.68. These two RHGI-RBD data points gave the highest studentized residuals (> 1.9), and they were from a road section with large amounts of buried wood and poor tree growth. Removing these two data points strengthened the linear relationship between RHGI of the seventh growing season and RBD by 15%, and a peak model (Systat Software 2000) best fit the data distribution (R2= 0.66) (Fig. 4.3b). The peak model also indicated that RBD threshold associated with RHGI < 1.0 was at 0.78 for the seventh growing seasons, and height growth peaked at RBD of 0.63 (Fig. 4.3b). 4.3.4 Relative bulk density and relative height growth index: interior Douglas-fir Height growth of interior Douglas-fir at experiment 2 did not vary with soil compaction during the first growing season, with RHGI values above 2.0, but intercepts of the trend lines dropped over subsequent growing seasons (Fig. 4.4a). After two growing seasons, 54% of variation in the height growth was explained by RBD. For the fifth and seventh growing seasons, about 70% of variation in height growth was associated with changes in RBD, implying there was an increasing influence of compaction on height growth. The best-fit (exponential) regression lines showed that RBD values > 0.72 were associated with height growth lower than expected (i.e., RHGI < 1.0) from the fifth growing season onward. On the other hand, RHGI values above 1.0 over the whole range of RBD during the first and second growing seasons reflected the superior height of planting stock (24 cm) relative to height of natural regeneration (20 cm) predicted by SiteTools for the undisturbed sites. Height growth of Douglas-fir at experiment 4 was weakly related to RBD, and the RHGI kept dropping through the two growing seasons (Fig. 4.4b). There was a sharp drop in RHGI at 123 RBD > 0.70 for the two growing seasons, while further change in RHGI was not observed when soils were additionally compacted. The data showed that interior Douglas-fir did not grow well in both normal (RBD of 0.70 - 0.80) and compacted (RBD > 0.80) soils in north-central BC. 4.3.5 Relative bulk density and relative height growth index: lodgepole pine The RBD values from experiment 1 were distributed over a wide range and the relationships between lodgepole pine height growth and RBD were relatively strong (Fig. 4.5a). During the first growing season, height growth was not related to RBD. From the second growing season onward, height growth varied with RBD, indicating that compaction affected tree growth. During the second growing season, a linear regression best described the relationship between height growth and RBD, and better height growth than in undisturbed conditions was obtained at RBD < 0.78. From the fifth to the eighth growing season, a second order regression best described the relationship, and better growth occurred when RBD was < 0.80 (season 5) and < 0.87 (season 8). Pooling lodgepole pine height growth data over experiments 1, 3, and 5 (Fig. 4.5b) showed the same trends as described for experiment 2 (Fig. 4.5a), where reduced height growth is associated with an RBD limit of 0.78 and 0.84 for growing seasons 5 and 8, respectively. Height growth peaked at RBD of 0.60 - 0.63 in the fifth and eighth growing seasons. 4.4 Discussion Soils with surface organic material had a narrower range of BD and very few of them were severely compacted. The RBD values for these soils were consistently lower than for soils without surface organic material, partly reflecting that plantation soils with undisturbed forest floor layers received less machine traffic than roads and landings, but also potentially illustrating the importance of surface organic material in mitigating the effect of machine traffic (Soane 1990). It is also possible that the natural aggregates in undisturbed soils under the 124 surface organic material made them harder to compact (Dexter 1988). Although the presence of surface organic material mitigated the negative influence of compaction on height growth, our study showed that tree height growth was better related to RBD than to the thickness of surface organic material. The presence of organic material can reduce daytime mineral soil temperature during the growing season, reduce evaporation, and has been associated with greater soil water content in moderately compacted soils (Bulmer et al. 2007). Despite this, when soils were severely compacted, other factors such as poor aeration and high mechanical resistance associated with high RBD more likely limited plant growth (da Silva et al. 1994; da Silva and Kay 1996). Therefore, it may be important to reduce soil compaction below a limiting level so that presence of surface organic material can enhance tree height growth. The RHGI of hybrid spruce was consistently near its minimum values when RBD was > 0.75, with only several instances where it exceeded 1.0. Hybrid spruce is a shallow rooted species, forming over 87% of its root mass in the top 15 cm of soil (including both forest floor and an A horizon) (Safford and Bell 1972; Kimmins and Hawkes 1978; Strong and La Roi 1983), suggesting that it is sensitive to soil compaction. The fact that hybrid spruce attained height growth equivalent to the undisturbed soils in some cases may be attributed to the presence of cracks and fissures along which roots could grow, and the presence of lateral roots close to the surface of the roads. The height growth models used by SiteTools to calculate RHGI for spruce may also slightly underestimate the ‘expected’ height at these young ages, in which case our proposed RBD limiting values would need to be revised down. In north-central BC, the range of RHGI for seven year old spruce trees growing on the worst performing plots versus the best performing plots (0.55 - 2.1 or about 3.8 times greater for the best performers) is greater than that for lodgepole pine (0.39 - 1.04 or about 2.6 times greater) and Douglas fir (0.21 - 1.3 or about 6.2 times greater). The relative magnitude of the 125 growth effects for spruce, pine, and Douglas-fir on these sites may provide insight into their relative susceptibility to compaction, particularly during the seedling establishment phase when the majority of roots are confined to surface soil layers. Considering that RHGI did not respond to increasing RBD in very compact soils, it could be more informative to use response measures (e.g., mortality, transpiration rate, or chlorophyll fluorescence) rather than height growth to reveal compaction effects on spruce seedling performance at extremely high RBD values (i.e., RBD > 0.80). Height growth of Douglas-fir and lodgepole pine in the first growing season was not related to RBD. Root development in the nursery is a major factor controlling survival and first- year growth of planted seedlings (Ritchie and Dunlap 1980). When seedlings are planted, a hole is created that is at least large enough to fit the root plug, resulting in a looser rooting area relative to the surrounding soil. This planting technique results in similar rooting conditions during the first growing season in the field to those in the nursery. In addition, one-year-old nursery-grown seedlings have a reasonable nutrient reserve, contributing to good growth during the first year following planting. When the roots extend out of the planting hole, properties of the neighboring soil, such as pore size, soil water availability, soil strength, aeration, and temperature, start to affect seedling growth (Letey 1985; Lipiec et al. 1991). The high RHGI of Douglas-fir (> 2.0) observed in the first growing season reflected the robust initial growth that resulted from the large planting stock produced by good nursery cultural practices. While the initial seedling conditions affected relative height growth in the following growing seasons, the influence of these conditions was not as strong. We found that RBD influenced height growth through the seventh (Douglas-fir) or eighth (lodgepole pine) growing season, and some research suggests that such effects may persist for many years or even decades. Froehlich et al. (1985) reported that trees planted on compacted skid trails in west-central Idaho still had not recovered to the undisturbed conditions 23 years after logging. 126 Interior Douglas-fir did not overcome transplant shock in the north-central interior sites in our study, which was evidenced by the continuous decrease in its relative height growth from the first through the seventh growing seasons, even when RBD was < 0.80. Our results imply that climate or other site conditions may have reinforced the effect of compaction on early height growth for interior Douglas-fir on these sites, which are often characterized by forest floor removal and site disturbance. At experiment 2, which is located in the middle of Douglas fir’s geographic range in BC, RHGI of seedlings growing in compacted soils (RBD > 0.72) appeared to decline continuously from planting through seven years. In contrast, RHGI for seedlings growing in un-compacted soils appeared to converge near a value of 1 after five to seven years, indicating that the trees were meeting expectations where soil conditions were favourable. Successful conifer establishment on harvested sites depends on ectomycorrhizal development to capture scarce site resources (Danielson 1985; Perry et al. 1987), and forest floor removal has been shown to dramatically reduce ectomycorrhizal colonization (Simard et al. 2003) while compaction can reduce colonization even further. For example, in well-aerated soils, un-inoculated interior Douglas-fir readily formed ectomycorrhizae within several months (Teste et al. 2004), but in compacted and disturbed soils, colonization of Douglas-fir seedlings was restricted (Skinner and Bowen 1974; Wert and Thomas 1981). Early mycorrhization in the field is especially important for Douglas-fir because this species does not generally form mycorrhizae under nursery conditions (Berch et al. 1999). It is possible that poor soil structure and poor mycorrhizal colonization resulting from site preparation may have affected interior Douglas-fir growth at this early stage in our study. Still, RBD could be used to establish the relationship between height growth and soil conditions; this is because compaction changes soil structure, which in turn affects seedling growth and mycorrhization through its effects on penetration resistance, aeration, and water supply. 127 During early growth of lodgepole pine, our observation that RBD < 0.80 supported similar tree height growth to undisturbed conditions agrees with other studies (da Silva et al. 1994; Carter 1990; Hakansson 1990). In a loamy sand soil grown with alfalfa (Medicago sativa L.), an RBD of 0.80 was associated with high LLWR (corresponding with aeration ≥ 10%, maximum available water, and penetration resistance < 2,500 kPa). Where RBD > 0.80, however, there was a sharp drop in LLWR (da Silva et al. 1994). An RBD of 0.90 was associated with minimal LLWR in the study of da Silva et al (1994) and they found that annual plant growth was impaired under such a compacted condition. Our findings that RHGI < 1.0 occurs at very high RBD agrees with the study of da Silva et al. (1994).  In particular, we found that tree height growth was substantially impeded (RHGI < 0.5) at the uppermost RBD value of 1.01. In our study, the RBD threshold at which compaction limited the height growth of both lodgepole pine and hybrid spruce was between 0.78 – 0.84. Maximum height growth of these two species occurred at RBD 0.60 - 0.63.  In studies that used annual plant species such as spring barley, an RBD of around 0.80 corresponded to the maximum yield (Carter 1990; Hakansson 1990; Hakansson and Lipiec 2000), while an RBD of 0.89 started to limit yield (Carter 1990). Interestingly, the most common RBD values reported for continuously tilled soils were around 0.66 (Arvidsson and Hakansson 1991) and values as low as 0.63 were seldom reported in studies with annual plant species. Lodgepole pine showed more resistance to compaction than spruce. Not only did lodgepole pine have increasing threshold RBD values over successive growing seasons (e.g. 0.87 in the eighth growing season), but its relative height growth also declined more slowly than for hybrid spruce when RBD was > 0.63.  In compacted soils, lodgepole pine develops a dual root system with the lateral roots growing largely within the top 30-60 cm and vertical roots extending to the rock layer at a depth of 100-120 cm (Berndt and Gibbons 1958; Bishop 1962). 128 On blade-scarified sites with forest floor incorporation into the mineral soil, lodgepole pine developed a root system larger than spruce by the second growing season, and this trend continued during the rest of the five-year experiment (McMinn 1978). After seven growing seasons, lodgepole pine trees have likely developed a root system that extends deeper than the level of our soil measurements. This would reduce the effect of compaction on growth, as the trees could be deriving a greater proportion of their growth resources from relatively undisturbed soils at depth. Calculated as a ratio of field BD to the reference BD of the same soil, RBD removes influences of intrinsic soil properties (i.e., particle density, texture) on BD that are not directly affected by compaction. Based on our study, an RBD range of 0.75 - 0.80 appears to represent a growth-limiting threshold for Douglas-fir, lodgepole pine, and hybrid spruce at their early growth stage (i.e., within four to five growing seasons), regardless of soil texture and particle density; while BD threshold for the three species varied from 1.20 Mg m-3 on experiment 2, to 1.35 Mg m-3 on experiment 1, and to 1.40 Mg m-3 on experiment 4 (Appendix 11), which varied by 17%. On the other hand, soil particle density of experiment 2 was significantly lower than that of the other two experiments (Table 2.5). Bulk density thresholds reported in other studies varied substantially with soil texture. For example, Daddow and Warington (1983) reported BD thresholds of 1.60 - 1.80 Mg m-3 for sandy loam and 1.40 Mg m-3 for silt loam. Uncertainty of BD in characterizing soil compaction and plant growth was also reflected in its correlation with RBD, as these two parameters are not entirely interchangeable (R2=0.64) in describing soil compaction, especially when BD was not high. For example, at experiment 1 of our study, the burn and deep-till treatments at one plot had the same low BD (0.74 Mg m-3), but interestingly, the RBD (0.68) of the deep till treatment differed substantially from that of the burn treatment (0.42). Similarly, a deep-till treatment and a burn treatment from another plot had quite different BD values (1.03 and 1.26 Mg m-3, respectively) while the RBD did not differ 129 (0.73). In studying productivity and foliar N response of lodgepole pine and hybrid spruce to soil disturbance at the Long-Term Soil Productivity (LTSP) sites in the Sub-Boreal Spruce biogeoclimatic zone, Kranabetter et al. (2006) did not find universal criteria defining detrimental soil disturbance. Where machine traffic and soil disturbance lead to subtle differences in BD, expected compaction levels would not be reached based on BD since BD does not necessarily indicate level of compaction, therefore determination of the RBD may provide additional insight into the factors affecting forest productivity on compacted soils, compared to BD alone. Heterogeneity of forest soil and complexity of site conditions in forest ecosystems often make it difficult to achieve the expected levels of compaction in field experiments. Our findings suggest that BD may not always be a good indicator of forest soil compaction, and it is more beneficial to characterize soil compaction by RBD. Bulmer et al. (2007) studied the effects of tillage and wood waste amendment on lodgepole pine seedling growth on the same site as our experiment 3, and they found that rehabilitation methods did not result in an expected increase in height growth. On this experiment, we found that the untreated plots already had a very low RBD (0.70), and height growth was not reduced. This low RBD value implied that rehabilitation using tillage was not necessary, while such a finding could not be made based on BD values alone (Bulmer et al. 2007). In studies focusing on compaction impacts, it would be more informative to quantify (i.e., determine RBD) rather than qualify the level (i.e., simply state a generic level of soil compaction, such as light, medium, and heavy) of compaction. By stratifying soils into plasticity groups, as we have done, such interpretations could be further refined. Our findings suggest that rehabilitation practices are needed at sites where RBD exceeds 0.80, and that compaction is not detrimental at lower RBD values. 130 4.5 Conclusions Relative bulk density should be considered as an indicator of forest soil compaction with consequences on tree height growth and hence site productivity. Relative bulk density values observed in this study ranged from 0.48 - 1.01, and rehabilitated roads, landings, and undisturbed soils were often associated with low RBD values. While soils with thin surface organic material had high BD and RBD values, un-rehabilitated soils did not always have high RBD values and thus did not require rehabilitation. Presence of surface organic material mitigated the severity of compaction and was associated with lower RBD values. When interior Douglas-fir was planted close to the northern limits of its geographic range in BC, and where lodgepole pine was planted on low-elevating areas and clay-rich soils, these species did not grow well and RBD was weakly related to the height growth. Height growth of interior Douglas-fir was limited when RBD was > 0.72. Threshold RBD values obtained for the 0 - 20 cm soil layer that limited height growth of lodgepole pine increased from 0.78 to 0.84 as the trees got older. An RBD of 0.60 - 0.63 corresponded to the maximum height growth of lodgepole pine and hybrid white spruce. To obtain good seedling establishment, rehabilitation involving soil decompaction must be considered when RBD exceeds 0.80. The relationships found in our study have implications in assessing forest soil compaction and its effect on site productivity. The results will also help predict and monitor soil behaviour and associated tree growth in response to timber harvesting and site rehabilitation. 131 Table 4.1 Location, elevation, annual precipitation†, temperature, and soil texture for the experiments. Experiment No. Location Latitude, longitude Elevation (m asl) Annual precipitation (mm) Annual temperature (oC) Soil texture‡ 1 Bear Lake 54°37' N, 123°13' W 820 805 3.0 L/SL 2 Miriam Creek 50°24' N, 118°57' W 790–1050 601 4.7 SL 3 OK Falls 49°18' N, 119°26' W 1100 517 4.8 SL 4A Apollo Lake 54°32' N, 124°15' W 818 514 2.6 SiL/SiCL 4B John Prince 54°41' N, 124°28' W 900 561 2.2 L 4C Younges Road 54°21' N, 123°27' W 880 586 2.3 L/SL 5 Will Lake 50°27' N, 119°38' W 1260 511 3.4 SL †mean annual temperature and precipitation were estimated for all sites based on the Climate BC model and normal conditions from 1971- 2000 (Wang et al. 2006). ‡L = loam; S = sand; Si = silt; C = clay 132 Table 4.2 Establishment time, treatments, number of soil samples, tree species, and year of tree height measurements for the experiments included into this study. Experiment Year established Treatments† No. sample Species and year when tree growth was measured ‡ 1 2000 B, DM, P, S, U 45 Pl: 2000, 2002, 2004§, 2007 2 2000 P, DD, SD 27 Fd: 2000, 2001, 2004, 2006 3 1998 D, DT, DSS, U 35 Pl: 2001, 2002, 2005 4 2001 B, P, U 81 Fd: 2004, 2007; Pl: 2004, 2007; Sx: 2004, 2007 5 2000 P, D, DTBP, U 35 Pl: 2007¶ †B, burn; D, decompact; DD, deep till; DM, deep till and mulch; DSS, decompact and sortyard waste on surface; DT, decompact and topsoil; DTBP, till and burnpile/topsoil; P, plantation; S, scratch; SD, shallow till; U, untreated. ‡Fd, interior Douglas-fir; Pl, lodgepole pine; Sx, hybrid white spruce §no height data for plantation ¶only have tree data for treatments D and DTBP. 133 Table 4.3 Four regression models used to derive maximum bulk density (MBD). Name† Model‡ R2 n Overall MBD=2.02-1.35LL+0.0005CL-0.005oxOM-0.16AlO+0.0005VCS 0.92 144 Non-plastic MBD=1.98-1.61LL-0.11AlO+0.0006VCS-0.003TC+0.0006CL 0.98 29 Moderately plastic MBD=2.27-1.62PL-0.003oxOM-0.0005MSI-0.003TC-0.0005FSI- 0.18AlO 0.92 99 Highly plastic MBD=1.72-0.004TC -0.82PL 0.87 16 †non-plastic, soils with no plastic limit; moderately plastic, soils had plastic limit while liquid limit < 0.50 kg kg-1; highly plastic, soils with liquid limit > 0.50 kg kg-1. ‡LL, liquid limit and PL, plastic limit (kg kg−1); AlO, Al-oxide (%); oxOM, oxidisable organic matter; TC, total C (g kg−1); CL, Clay; MSI, medium silt; FSI, fine silt; VCS, very coarse sand (g kg−1). 134 Table 4.4 Ranges of relative bulk density (RBD) and site indices for interior Douglas-fir (Fd), lodgepole pine (Pl), and hybrid white spruce (Sx) of the experiments. Site index Experiment RBD range Thickness range of surface organic material (cm) BEC† unit Series Fd Pl Sx 1 0.54 (B‡) – 1.01 (U) 0 – 10 SBSmk1 01 —¶ 20.1 — 2 0.48 (C) – 0.94 (SD) 0 – 5 ICHmw2 03 21.0 — — 3 0.50 (DSS) – 0.71 (D) 0 – 11 IDFdm1 01 — 18.0 — 4A 0.72 (P) – 0.86 (P) 0 – 12 SBSdw3 01/06 18.0 21.0 18.4/15.0 4B 0.54 (U) – 0.96 (U) 0 – 16 SBSdw3 01 18.0 21.5 18.4 4C 0.67 (U) – 0.86 (P) 0 – 13 SBSmk1 07/08 21.0 21.0 20.8 5 0.63 (DTBP) – 0.91 (D) 0 IDFdk2 03 — 18.0 — †BEC, biogeoclimatic ecosystem classification. ‡treatment code corresponding to the RBD value. See Table 4.2 for treatment codes. ¶not applicable. 135 Table 4.5 Regression analysis of thickness of surface organic material (FF) and relative bulk density (RBD) on the relative height growth index. Species†, growing seasons Intercept Coefficient and variable R2 P Experiment 1 (n=15) Pl 1 0.99 0.08 FF × RBD2 0.20 0.096 Pl 2 1.92 -1.20 RBD  0.75 0.000 Pl 5 1.97 -1.53 RBD2 0.74 0.000 Pl 8 2.18 -1.61 RBD2 0.78 0.000 Experiment 2 (n=9) Fd 1 2.39 0.31 FF × RBD2  0.69 0.000 Fd 2 1.62 0.51 FF – 0.47 RBD 0.98 0.000 Fd 5 1.59 -1.46 RBD2 + 1.09 FF × RBD2 0.86 0.003 Fd 7 1.59 -1.65 RBD2 + 1.06 FF × RBD2 0.82 0.006 Experiment 3 (n=12) Pl 4 0.79 0.56 FF × RBD 0.62 0.002 Pl 5 0.85 0.96 FF × RBD2  0.61 0.003 Pl 8 1.15 0.72 FF × RBD2 0.33 0.052 Experiment 4  (n=27) Fd 4 -0.41 0.89 RBD-1 0.29 0.005 Fd 7 -0.90 1.06 RBD-1 0.42 0.000 Pl 4 0.84 - 0.44 RBD2 - 0.59 FF × Log (RBD) 0.37 0.004 Pl 7 0.95 0.14 FF - 0.50 RBD2 0.39 0.003 Sx 4 -1.25 2.08 RBD-1 0.29 0.004 Sx 7 0.58 -4.65 Log (RBD) 0.25 0.008 136 Species†, growing seasons Intercept Coefficient and variable R2 P Experiment 5 (n=6) Pl 8 1.83 -1.44 RBD2 0.83 0.011 †Fd, interior Douglas-fir; Pl, lodgepole pine; and Sx, hybrid white spruce. 137 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 1.10 cohesive soils: R2 = 0.72*** noncohesive soils: R2 = 0.65*** cohesive + non-cohesive soils: R2 = 0.64*** a b Bulk density (Mg m-3) 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 R el at iv e bu lk  d en si ty 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 1.10 surface organic material < 3 cm: R2 = 0.73*** surface organic material >= 3 cm: R2 = 0.73***  Figure 4.1 Relationship between relative bulk density and field bulk density for (a) cohesive and non-cohesive soils, and (b) cohesive soils with (>= 3 cm) and without (< 3 cm) surface organic material.*** Significant at P < 0.001. 138 Figure 4.2 Relationship between the relative height growth index of (a) interior Douglas-fir (Fd), (b) lodgepole pine (Pl), and (c) hybrid white spruce (Sx) and relative bulk density for the seventh growing season at experiment 4. Error bars are SEs. Relative height growth bars with the same letter are not significantly different at P = 0.05.  0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 RBD < 0.70 RBD 0.70-0.83 RBD > 0.83 R el at iv e he ig ht  g ro w th  in de x surface organic material < 3 cm surface organic material >= 3 cm a a ac b b bcb Fd  0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 RBD < 0.70 RBD 0.70-0.83 RBD > 0.83 R el at iv e he ig ht  g ro w th  in de x surface organic material < 3 cm surface organic material >= 3 cm b a ad c ab a Pl 139  0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 RBD < 0.70 RBD 0.70-0.83 RBD > 0.83 R el at iv e he ig ht  g ro w th  in de x surface organic material < 3 cm surface organic material >= 3 cm c a a aa aa Sx  140 Relative bulk density 0.50 0.60 0.70 0.80 0.90 1.00 R el at iv e he ig ht  g ro w th  in de x 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Sx7 ***390:lLinearmode ***660:Peakmodel 2 2 .R .R = = 0.50 0.60 0.70 0.80 0.90 1.00 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Sx4: R2 = 0.27** Sx7: R2 = 0.25** b a  Figure 4.3 Relationship between relative height growth index of hybrid white spruce (Sx) and relative bulk density at experiment 4 (a) with all data, and (b) with outliers being removed. Number following name ‘Sx’ indicates number of growing seasons in the field. Data distributed above the dotted y = 1.0 line indicate better height growth than in the pre- disturbed conditions. **, *** Significant at P < 0.01, 0.001, respectively. 141 0.40 0.50 0.60 0.70 0.80 0.90 1.00 R el at iv e he ig ht  g ro w th  in de x 0.0 0.5 1.0 1.5 2.0 2.5 Fd1: R2 = 0.01 Fd2: R2 = 0.54* Fd5: R2 = 0.70** Fd7: R2 = 0.70** a Relative bulk density 0.40 0.50 0.60 0.70 0.80 0.90 1.00 0.0 0.5 1.0 1.5 2.0 2.5 Fd4: R2 = 0.25** Fd7: R2 = 0.38*** b  Figure 4.4 Relationship between relative height growth index of interior Douglas-fir (Fd) and relative bulk density (a) at experiment 2, and (b) at experiment 4. Number following name ‘Fd’ indicates number of growing seasons in the field. Data distributed above the dotted y = 1.0 line indicate better height growth than in the pre-disturbed conditions. *, **, *** Significant at P < 0.05, 0.01, 0.001, respectively. 142 0.40 0.50 0.60 0.70 0.80 0.90 1.00 1.10 0.0 0.5 1.0 1.5 2.0 Pl1: R2 = 0.01 Pl2: R2 = 0.75*** Pl5: R2 = 0.75*** Pl8: R2 = 0.81*** a Relative bulk density 0.40 0.50 0.60 0.70 0.80 0.90 1.00 1.10 R el at iv e he ig ht  g ro w th  in de x 0.0 0.5 1.0 1.5 2.0 Pl5: R2 = 0.33** Pl8: R2 = 0.57*** b  Figure 4.5 Relationship between relative height growth index of lodgepole pine (Pl) and relative bulk density (a) at experiment 1, and (b) at experiments 1, 3, and 5. Number following name ‘Pl’ indicates number of growing seasons in the field. Data distributed above the dotted y = 1.0 line indicate better height growth than in the pre-disturbed conditions. **, *** Significant at P < 0.01, 0.001, respectively. 143 4.6 References American Society for Testing Materials. 2000. Design D698-00a. Standard test methods for laboratory compaction characteristics of soil using standard effort (12,400 ft-lbf/ft3 (600 kN- m/m3)). ASTM, West Conshohocken, PA. Ares, A., Terry, T., Harrington, C., Devine, W., Peter, D., and Bailey, J. 2007. Biomass removal, soil compaction, and vegetation control effects on five-year growth of Douglas-fir in coastal Washington. For. Sci. 53:600-610. Arvidsson, J., and Hakansson, I. 1991. A model for estimating crop yield losses caused by soil compaction. Soil & Tillage Res. 20:319-332. B.C. Ministry of Forests and Range. 2004. SiteTools. Version 3. Victoria, BC. (URL:http://www.for.gov.bc.ca/hre/sitetool/release.htm). B.C. Ministry of Forests and Range. 2006. Biogeoclimatic zones of British Columbia. Map 1:5,000,000. Victoria, BC. Berch, S.M., Xiao, G., and Bulmer, C. 1999. Commercial mycorrhizal inoculants: value added conifers for site rehabilitation or just another way to spend money? In Proceedings: 19th Annual Meetings of the Forest Nursery Association of British Columbia, Sept 27-30, Vancouver, B.C. Extension Services, Tree Improvement Br., Surrey, Canada. Berndt, H.W., and Gibbons, R.D. 1958. Root distribution of some native trees and understory plants growing on three sites within ponderosa pine watersheds in Colorado. Station Paper. Rocky Mountain Forest and Range Experiment Station 14. Bishop, D.M. 1962. Lodgepole pine rooting habits in the Blue Mountains of north-eastern Oregon. Ecology 43:140-142. Bulmer, C., Venner, K., and Prescott, C. 2007. Forest soil rehabilitation with tillage and wood waste enhances seedling establishment but not height after 8 years. Can. J. For. Res. 144 37:1894-1906. Carter, M.R. 1990. Relative measures of soil bulk density to characterize compaction in tillage studies on fine sandy loams. Can. J. Soil Sci. 70:425-433. Campbell, D.B., Bulmer, C.E., Jones, M.D., Philip, L.J., and Zwiazek, J.J. 2008. Incorporation of topsoil and burn-pile debris substantially increases early growth of lodgepole pine on landings. Can. J. For. Res. 38:257-267. da Silva, A.P., Kay, B.D., and Perfect, E. 1994. Characterization of the least limiting water range of soils. Soil Sci. Soc. Am. J. 58:1775-1781. da Silva, A.P., and Kay, B.D. 1996. The sensitivity of shoot growth of corn to the least limiting water range of soils. Plant and Soil 184:323-329. Daddow, R.L., and Warrington, G.E. 1983. Growth-limiting soil bulk densities as influenced by soil texture. USDA Agric. For. Serv., Rep. Rep. WSDG-TN-00005. Danielson, R.M. 1985. Mycorrhizae and reclamation of stressed terrestrial environments. In Soil reclamation processes: microbiological analyses and applications. Edited by R.L. Tate III & D.A. Klein. Marcel Dekker, Inc., New York. pp. 173-201. Dexter, A.R., 1988. Strength of soil aggregates and of aggregate beds. Catena Suppl. 11:35-52. Eriksson, J., Hakansson, I., and Danfors, B. 1974. The effect of soil compaction on soil structure and crop yields. USDA-ARS, Rep. 354. Froehlich, H.A. 1979. Soil compaction from logging equipment: effects on growth of ponderosa pine. J. Soil Water Conserv. 34:276-278. Froehlich, H.A., Miles, D.W.R., and Robbins, R.W. 1985. Soil bulk density recovery on compacted skid trails in Central Idaho. Soil Sci. Soc. Am. J. 49:1015-1017. Greacen, E.L., and Sands, R. 1980. Compaction of forest soils:a review. Aust. J. Soil Res. 18:163-189. Grossman, R.B., and Reinsch, T.G. 2002. Bulk density and linear extensibility. In Methods of 145 soil analysis. Part. 4. SSSA. Edited by J.H. Dane & G.C. Topp. Madison, WI. pp. 201-254. Hakansson, I. 1990. A method for characterizing the state of compactness of the plow layer. Soil & Tillage Res. 16:105-120. Hakansson, I., and Lipiec, J. 2000. A review of the usefulness of relative bulk density values in studies of soil structure and compaction. Soil & Tillage Res. 53:71-85. Heilman, P. 1981. Root penetration of Doulas-fir seedlings into compacted soil. For. Sci. 27:660-666. Heninger, R., Scott, W., Dobkowski, A., Miller, R., Anderson, H., and Duke, S. 2002. Soil disturbance and 10-year growth response of coast Douglas-fir on nontilled and tilled skid trails in the Oregon Cascades. Can. J. For. Res. 32:233-246. Jones, C.A. 1983. Effect of soil texture on critical bulk densities for root growth. Soil Sci. Soc. Am. J. 47:1208-1211. Kimmins, J.P., and Hawkes, B.C. 1978. Distribution and chemistry of fine roots in a white spruce-subalpine fir stand in British Columbia: implications for management. Can. J. For. Res. 8:265-279. Kranabetter, J.M., Sanborn, P., Chapman, B.K., and Dube, S. 2006. The contrasting response to soil disturbance between lodgepole pine and hybrid white spruce in subboreal forests. Soil Sci. Soc. Am. J. 70:1591-1599. Kuan, H.L., Hallett, P.D., Griffiths, B.S., Gregory, A.S., Watts, C.W., and Whitmore, A.P. 2007. The biological and physical stability and resilience of a selection of Scottish soils to stresses. Eur. J. Soil Sci. 58:811-821. Letey, J. 1985. Relationship between soil physical properties and crop production. Adv. Soil Sci. 1:276-294. Lipiec, J., Hakansson, I., Tarkiewicz, S., and Kossowski., J. 1991. Soil physical properties and growth of spring barley as related to the degree of compactness of two soils. Soil Tillage 146 Res. 19:307-317. McMinn, R.G. 1978. Root development of white spruce and lodgepole pine seedlings following outplanting. In Proceedings of the root form of planted trees symposium, May 16-19, Victoria, BC. Edited by E. Van Eerden, and J. M. Kinghorn. Victoria, BC. pp. 186-190. Perry, D.A., Molina, R., and Amaranthus, M.P. 1987. Mycorrhizae, mycorrhizospheres, and reforestation: current knowledge and research needs. Can. J. For. Res. 17:929-940. Pidgeon, J.D., and Soane, B.D. 1977. Effects of tillage and direct drilling on soil properties during growing season in a long term barley mono-culture system. J. Agric. Sci. 88:431-442. Richardson, B, Skinner, M.F., and West, G. 1999. The role of forest productivity in defining the sustainability of plantation forests in New Zealand. For. Ecol. Manage. 122:125-137.  Ritchie, G.A., and Dunlap, J.R. 1980. Root growth potential: its development and expression in forest tree seedlings. N. Z. J. For. Sci. 10:218-248. Safford, L.O., and Bell, S. 1972. Biomass of fine roots in a white spruce plantation. Can. J. For. Res. 2:169-172. SAS Institute. 1990. SAS/STAT user’s guide. Version 6. 4th ed. SAS Inst. Inc., Cary, NC. Simard, S.W., Jones, M.D., Durall D.M., Hope, G.D., Stathers, R.J., Sorensen N.S., and Zimonick B.J. 2003. Chemical and mechanical site preparation: effects on Pinus contorta growth, physiology, and microsite quality on grassy, steep forest sites in British Columbia. Can. J. For. Res. 33:1495-1515. Skinner, M.F., and Bowen, G.D. 1974. The penetration of soil by mycelial strands of ectomycorrhizal fungi. Soil Biol. Biochem. 6:57-61. Soane, B.D. 1990. The role of organic matter in soil compactibility: A review of some practical aspects. Soil Tillage Res. 16:179-201. Strong, W.L., and La Roi, G.H. 1983. Root system morphology of common boreal forest trees in Alberta, Canada. Can. J. For. Res. 13:1164-1173. 147 Systat Software 2000. SigmaPlot Software. Systat Software Inc., Chicago, IL Teste, F.P., Schmidt, M.G., Berch, S.M., Bulmer, C., and Egger, K.N. 2004. Effects of ectomycorrhizal inoculants on survival and growth of interior Douglas-fir seedlings on reforestation sites and partially rehabilitated landings. Can. J. For. Res. 34:2074-2088. Wang, T., Hamann, A., Spittlehouse, D.L., and Aitken, S.N. 2006. Development of scale-free climate data for western Canada for use in resource management. Intl. J. Climatology, 26(3):383-397. Wert, S., and Thomas, B.R. 1981. Effects of skid roads on diameter, height, and volume growth in Douglas-fir. Soil Sci. Soc. Am. J. 45:629-632. Whalley, W.R., Dumitru, E., and Dexter, A.R. 1995. Biological effects of soil compaction. Soil & Tillage Res. 35:53-68. Wronski, E.B., and Murphy, G. 1994. Responses of forest crops to soil compaction. In Soil compaction in crop production. Editied by  B.D. Soane & C. van Ouwerkerk. Elsevier, Amsterdam. pp. 317–342. Zhao, Y.H., Krzic, M., Bulmer, C.E., and Schmidt, M.G. 2008. Maximum bulk density of British Columbia forest soils from the proctor test: Relationships with selected physical and chemical properties. Soil Sci. Soc. Am. J. 72:442-452. 148 5 CONCLUDING CHAPTER 5.1 Synthesis of results The research presented in this thesis addressed the overall goal of developing a high- level integrated indicator to characterize forest soil compaction and correlating it to tree growth. The finding in Chapter 2 that soil particle density varied considerably within and among sites warranted the evaluation of RBD as a soil compaction indicator in relation to soil productivity (as addressed in Chapter 4). The strong relationship found between soil particle density and soil physical and chemical properties (i.e., soil organic matter, oxides, and PSD) also explained why particle density, an important soil property influencing the bulk density, was not included in the multiple regression analysis in the modeling of MBD in Chapter 3. I found that the gauge precision was an important factor influencing the accuracy of the volume-constant GP in testing the volume of solids. The uncertainty of the pycnometer decreased with an increase in the volume of the sample. Change in the atmospheric pressure did not alter the volume estimation and other sources of error were negligible when temperature was maintained at a constant level during each test. Soil mineralogy accounted for more variation (60%) in soil particle density than SOM in my study. Grouping the samples by the geographic location substantially improved the relationship between soil particle density and SOM. Oxalate-extractable Fe- and Al-oxides and PSD were related to mineral particle density, with more variation explained by the former. Soil physical and chemical properties can be employed to predict soil particle density and SOM and Al- and Fe-oxides were important soil properties in this regard. Particle density of mineral for forest soils from the interior BC varied substantially among geographic locations. Chapter 3 focused on developing models to predict MBD. When single soil properties were used to predict the MBD, the properties in order of highest to lowest significance were 149 liquid and plastic limits, total carbon and oxidisable organic matter, and oxalate-extractable oxides. The PSD alone accounted for very little variation in MBD. Liquid limit and plastic limit had similar functions in relation to soil compaction, as did total carbon and oxidisable organic matter. In general, liquid limit, clay, oxidisable organic matter, Al-oxides, and very coarse sand content can be used in combination to predict MBD with high accuracy (R2=0.92). When soils were stratified according to the plasticity, prediction of MBD was substantially improved; the best predictions were obtained for nonplastic soils (R2=0.98), while multiple regression explained the least amount of variation for highly plastic samples (R2=0.92). The successful realization of study objectives 4 – 6 (addressed in Chapter 3) made it possible to avoid labour-intensive and time-consuming Proctor test while using MBD as the reference bulk density in the next part of this thesis research (addressed in Chapter 4), which enabled me to pursue the study of RBD in relation to soil compaction and tree growth on landings, rehabilitated roads, and undisturbed soils. Relative bulk density ranged from 0.48 - 1.01 on the study sites, and undisturbed soils were often associated with low RBD values. Where surface organic materials were thin or absent, soils tended to have higher RBD than those with thicker forest floors, suggesting that surface organic materials may mitigate the severity of compaction. However, disturbed soils lacking surface organic materials did not always have high RBD values and thus did not always require rehabilitation. I used the relative height growth index (RHGI) to represent the status of tree height growth in situ relative to the site index of undisturbed growth conditions. Height growth of interior Douglas-fir was limited when RBD was > 0.72. Threshold RBD limiting height growth of lodgepole pine increased from 0.78 to 0.84 as the trees grew older. An RBD of 0.60 - 0.63 corresponded to the maximum height growth for lodgepole pine and hybrid white spruce. On the other hand, interior Douglas-fir planted close to the northern limits of its geographic range and lodgepole pine planted on low elevation areas and clay-rich soils did not grow well and RBD was weakly related to their height 150 growth. Findings of this study show that RBD should be considered as an indicator of forest soil compaction and site productivity, since it can be successfully correlated to tree height growth. To obtain good seedling establishment, I recommend that rehabilitation involving soil decompaction should be considered when RBD exceeds 0.80. 5.2 Strengths of the thesis My thesis provided theoretical and applied contributions to the discipline of applied soil physics on forest soils. In this thesis, I have made the following original findings and contributions: • I carried out an innovative method to analyze the accuracy of the gas pycnometer, which not only shed light on the importance of the precision of the gauge pressure transducer, but also provided insight into the influence of the sample volume on test accuracy. Findings of this study will be helpful in the selection of a suitable gauge in building a custom gas pycnometer. The evaluation of particle density of forest soils from the interior BC revealed the sources of variation in soil particle density. My findings that mineral particle density in BC forest soil varied considerably among and within the sites and that it varied away from the commonly referred value of 2.65 Mg m-3 will help the accurate determination of soil porosity and PSD. These findings support consideration of particle density when using MBD as an indicator of soil compaction susceptibility of soils varying in organic matter content and mineralogy. • I developed models to predict soil compaction susceptibility. To examine the relationship between RBD and tree growth, collection of a large number of soil samples and tree growth data is necessary, especially when heterogeneous forest soils are encountered. Under such circumstances, it is challenging (in terms of time and cost) to obtain the reference bulk density (i.e., MBD) by carrying out the Proctor test. I successfully solved this problem by 151 developing models to predict the MBD. A novel finding achieved during the modeling process was to group soils based on their plasticity and assess them in relation to soil compaction. I also found that oxalate-extractable Fe-oxide described the function of oxides in relation to MBD. The method described in predicting MBD from readily measured soil properties could enable more effective assessments of soil compaction susceptibility in forest ecosystems (Zhao et al. 2008). This would be particularly beneficial in regions characterized by high spatial variation in soil properties, such as in BC’s forest soils. Prediction would initially involve determination of soil plasticity followed by use of the appropriate model to determine MBD. • My study was the first attempt to evaluate RBD as an indicator of forest soil compaction and to correlate it to tree growth on a wide range of field study sites. I found that maximum tree height growth occurred within a narrow RBD range (0.60 – 0.63): values outside this range did not benefit height growth. The presence of surface organic material mitigated the severity of compaction and was associated with lower RBD values. The finding that RBD > 0.80 was generally associated with impeded tree height growth can help guide forest soil rehabilitation activities, especially for concentrating resources in problematic areas and avoiding unnecessary rehabilitation efforts. The relationships found in my study have implications for assessing forest soil compaction and its effect on site productivity. The results will also help predict and monitor soil behaviour and associated tree growth in response to timber harvesting and site rehabilitation. 5.3 Limitations of the thesis In this thesis project, I focused on soil and tree data from mechanically disturbed forest sites such as landings, roads, as well as other long-term forest soil rehabilitation experiments, which limited my exploration of the soil-tree relationships under natural (undisturbed) conditions. For example, the RBD range for the natural forest soils is unknown. In addition, tree 152 growth was obtained for just three species (interior Douglas-fir, lodgepole pine, and hybrid white spruce), and it is not clear if other economically important tree species grown in BC would follow the same trends. Furthermore, it is not clear if the RBD range that supported optimal tree height growth obtained in my study coincides with the RBD range under natural conditions. We did not have continuous year by year measurements of tree height growth. Therefore, growth rate could not be used to develop the relationship between height growth and soil compaction. I used the SiteTool software (BC Ministry of Forests and Range 2004) to estimate species specific site index at specific site, and then expressed the measured height growth relative to this site index (i.e., RHGI). This method was useful in determining the relationship between height growth and RBD within species, but the relationship could not be compared among species within the same RBD range, even when data were obtained on the same site. I used RHGI to describe tree response to varying RBD, but tree height growth under severely compacted conditions was not well correlated with RBD. For example, height growth of spruce did not change with RBD > 0.80. It would have been more informative to use other indicators of tree response (e.g., mortality, transpiration rate, or chlorophyll fluorescence) to reveal compaction effects on seedling performance (Kozlowski 1999). Another limitation of my study was the poor performance of the model at predicting MBD for highly plastic soils (R2=0.87). The soil samples collected from 33 forest sites throughout BC were relatively coarse textured, with only 16 out of 147 samples being highly plastic (Chapter 3). When collecting soil samples for this study, we tried to obtain a reasonable representation of forest soil types present in BC, but only a limited number of regions (southeastern part of the Vancouver Island, Peace River region) have fine-textured and highly plastic soils. It would have been better had we included an equal number of soil samples of high, 153 moderate, and low plasticity. Because of the low sample size, only two independent variables (i.e., total C and plastic limit) were selected for the highly plastic model during the multiple regression analysis, while other independent variables had low χ2 scores and were excluded from the model. In Chapter 4, nine soil samples were highly plastic, and the model applied to these samples would have resulted in a less accurate approximation for MBD and hence RBD. 5.4 Future research The contributions of this thesis highlighted several themes for future research that would serve to further contribute to forest soil science: • This study focused on forest sites that were mechanically disturbed during timber- harvesting and site preparation. It would be beneficial to expand my research achievements to old growth forest and regions under natural disturbances such as infestation by mountain-pine beetle (Dendroctonus ponderosae) or fire. In 2007, over 710 million m3 of wood and about 13.5-million hectares of forest in BC were affected by mountain-pine beetle and most of these areas were disturbed by fire (Ministry of Forests and Range 2008). To recover the greatest value from dead timber before it burns or decays following mountain-pine beetle infestation, BC’s government has increased the annual allowable cut to increase salvage of mountain-pine beetle affected wood (Ministry of Forests and Range 2007). Since the wood must be harvested within a short time frame to prevent decay, harvesting has increased in intensity and scale, which in turn has caused higher levels of site disturbance and soil compaction, especially because soil conditions are often wetter than normal on mountain-pine beetle salvage cutblocks. Applying my findings, such as modeling of MBD and relating RBD to RHGI, to the harvested and un- harvested sites in the mountain-pine beetle infested forests will help assess the post-harvest soil condition and guide site regeneration, and it could also help guide harvesting practices to avoid compaction. 154 • In this study I did not observe root morphology and elongation within the RBD range. When soils are mechanically disturbed, indigenous root habits for specific tree species are altered differently under different levels of compaction and other soil conditions. Tree height growth in the disturbed soil is affected by the root deformation due to reduced access to nutrients and water (Greacen and Sands 1980). On the other hand, root habit can indicate the degree of compaction resistance a species exerts, and therefore its ability to regenerate under the unfavourable soil conditions (Robert and Lindgren 2006). A study on the response of root morphology with varying compaction levels on different soil types would shed more light into the relationship between RBD and tree growth. • Tree height growth related non-linearly to RBD when RBD was < 0.80, and maximum height growth was reached when RBD was around 0.60 – 0.63 for lodgepole pine and hybrid white spruce. While this thesis study provided insight into impeded height growth under compacted conditions, that trees did not grow well either at RBD lower than this range could not be explained. Plant growth is governed by available water, soil mechanical resistance, and air- filled porosity (Letey 1985). A study on the relationship between RBD and the above soil characteristics could help explain this observation. • There is a possibility to refine some of the soil parameters used in this study. In the Atterberg test, soil particles and aggregates < 0.425 mm were used to obtain plastic and liquid limits, and these values were used to represent plasticity of the same soil with particles < 2.00 mm (McBride, 2002). This could cause some bias in the analysis, because two soils with the same plasticity but different content of particles < 0.425 mm would behave differently in the Proctor test. To adjust the plastic and liquid limits based on the percentage of particles < 0.425 mm would provide another soil parameter that could be used to describe soil compaction susceptibility. Another improvement for the relationship study would be from the more accurate 155 determination of MBD. In the Proctor test, soil particles and aggregates < 4.75 mm were used to determine MBD; this value was then corrected to MBD of the fine fraction (< 2 mm) (ASTM 2000). In the correction process, particle density was set at 2.65 Mg m−3 to calculate volume of coarse fragments (Zhao et al. 2008). Based on this thesis, density of mineral particles < 2 mm in interior BC varied considerably, and it is thus possible that coarse fragments would have variable particle densities as well. Determination of the actual particle density of coarse fragments with a large-sample GP would improve the accuracy of the fine-fraction MBD, which would provide further insight into the relationship between MBD and soil physical and chemical properties. • New research on finding the extent to which subsoil RBD controls rooting depth in the province would highlight the importance of the sampling depth. Although rooting depth in some parts of BC is limited by soil temperature, in other places the potential for deep root development could be limited by soil conditions (Bulmer et al. 2004). Currently, RBD in our study was determined from mineral soils with a sampling depth of 0 – 20 cm, which is part of the soil profile where the majority of roots are distributed immediately after seedling transplantation. However, with the aging of the trees, some roots found their ways in the deeper soil and plant growth would become increasingly dependent on these deeper roots. Therefore the threshold RBD we obtained from the top 0-20 cm soil may be best suited to the establishment and early growth phase, not the long-term growth. One such evidence is that the threshold RBD increased with age for the lodgepole pine at site 3. With the climate change, deeper roots may play a more important role in allowing trees to survive the expected periods of drought that will accompany warming, and such a role could be revealed by the relationship study between subsoil RBD and the tree growth. In addition to future research highlighted by my thesis, there are many important avenues for further research to make RBD a more biologically-meaningful parameter. Soil 156 conditions such as water content, pH, and nutrient availability are often considered important abiotic variables in the study of site productivity. Relative bulk density has the potential to be included as another abiotic variable in this broader domain, yet its function needs to be explored. 5.5 References American Society for Testing Materials. 2000. Design D698-00a. Standard test methods for laboratory compaction characteristics of soil using standard effort (12,400 ft-lbf/ft3 (600 kN- m/m3)). ASTM, West Conshohocken, PA. B.C. Ministry of Forests and Range. 2004. SiteTools. Version 3. Victoria, BC. (URL:http://www.for.gov.bc.ca/hre/sitetool/release.htm). Bulmer, C., Venner, K., and Prescott, C. 2007. Forest soil rehabilitation with tillage and wood waste enhances seedling establishment but not height after 8 years. Can. J. For. Res. 37:1894-1906. Greacen, E.L., and Sands, R. 1980. Compaction of forest soils: a review. Aust. J. Soil Res. 18:163-189. Kozlowski, T.T. 1999. Soil compaction and growth of woody plants. Scand. J. For. Res. 14:596:619. Letey, J. 1985. Relationship between soil physical properties and crop production. Adv. Soil Sci. 1:276-294. McBride, R.A. 2002. Atterberg limits. In Methods of soil analysis. Part 4. Physical methods. Edited by J.H. Dane and G.C. Topp. SSSA Book Series no. 5. SSSA, Madison, WI. pp. 519- 527. Ministry of Forests and Range. 2007. The state of British Columbia’s forests, 2006. B.C. Min. For. Range, Victoria, B.C. 157 Ministry of Forests and Range 2008. 2007/08 annual service plan report. B.C. Min. For. Range. Victoria, B.C. Robert, J.A., and Lindgren, B.S. 2006. Relationships between root form and growth, stability, and mortality in planted versus naturally regenerated lodgepole pine in north-central British Columbia. Can. J. For. Res. 36:2642-2653. Zhao, Y.H., Krzic, M., Bulmer, C.E., and Schmidt, M.G. 2008. Maximum bulk density of British Columbia forest soils from the proctor test: Relationships with selected physical and chemical properties. Soil Sci. Soc. Am. J. 72:442-452. 158 APPENDICES Appendix 1: Relationship between mineral particle density and total sand for forest soils from interior British Columbia. y = 7E-05x + 2.63 R2 = 0.02 2.36 2.46 2.56 2.66 2.76 2.86 2.96 0 200 400 600 800 1000 Total Sand (g Kg-1) M in er al  p ar tic le  d en si ty  (M g m -3 ) 159 Appendix 2: Maximum bulk density, WMBD obtained by the Proctor test, and other soil physical and chemical properties tested for the MBD study†. Loc MBD WMBD ρS TC OxOM CL FSI MSI CSI VFS FS MS CS VCS AlO FeO LL PL 1 1.74 0.173 2.79 12.2 14.9 417.9 116.9 162.4 133.9 78.0 57.1 19.6 10.4 3.7 0.10 0.20 0.349 0.203 1 1.75 0.173 2.74 9.7 10.5 378.4 116.1 157.2 141.6 80.8 74.6 28.1 18.2 5.0 0.11 0.31 0.340 0.203 1 1.77 0.151 2.71 13.2 11.0 299.7 77.5 135.8 142.8 115.1 150.7 46.6 25.7 6.0 0.11 0.34 0.288 0.187 1 1.70 0.172 2.78 13.9 11.4 396.0 95.4 139.6 157.3 90.6 77.5 22.0 16.3 5.2 0.13 0.40 0.380 0.210 2 1.61 0.199 2.71 22.6 27.2 367.0 113.0 198.2 156.7 80.8 51.0 13.7 11.9 7.7 0.20 0.73 0.316 0.217 2 1.58 0.210 2.68 24.9 32.0 357.6 121.1 159.9 165.9 84.7 60.2 18.7 23.9 7.9 0.16 0.72 0.299 0.220 2 1.61 0.190 2.71 20.8 24.5 315.0 127.1 207.5 165.4 86.3 60.9 18.0 13.5 6.5 0.17 0.66 0.288 0.219 2 1.62 0.180 2.68 21.1 26.6 351.6 114.3 167.8 179.3 104.3 50.8 10.5 12.3 9.3 0.16 0.65 0.301 0.218 3 1.70 0.175 2.75 14.2 20.5 383.4 121.9 145.1 129.9 87.0 79.0 27.0 19.5 7.2 0.13 0.54 0.292 0.204 3 1.72 0.175 2.73 17.1 21.9 362.8 109.5 169.6 139.0 96.0 80.0 24.4 14.2 4.5 0.11 0.39 0.325 0.208 3 1.66 0.182 2.76 20.5 24.3 374.9 110.8 140.9 168.2 87.7 81.9 21.9 10.0 3.7 0.10 0.37 0.338 0.224 3 1.67 0.177 2.71 17.5 27.0 361.6 97.2 169.3 145.7 81.3 100.2 22.9 18.3 3.4 0.11 0.44 0.319 0.214 4 1.36 0.271 2.59 35.4 50.1 227.7 64.4 210.4 267.0 148.5 42.0 16.2 15.2 8.7 0.20 0.74 0.392 0.306 4 1.62 0.197 2.79 13.9 22.5 129.5 64.5 188.5 320.8 168.7 48.9 20.9 34.8 23.5 0.15 1.01 0.237 0.205 160 Loc MBD WMBD ρS TC OxOM CL FSI MSI CSI VFS FS MS CS VCS AlO FeO LL PL 4 1.46 0.213 2.69 29.9 44.0 246.9 105.6 278.5 191.7 106.8 23.9 17.4 18.9 10.4 0.27 0.83 0.317 0.255 4 1.72 0.158 2.67 9.8 19.5 166.4 82.7 217.3 321.5 118.9 25.1 19.4 26.8 21.9 0.13 0.86 0.195 0.171 4 1.43 0.250 2.54 35.2 42.2 132.6 83.2 216.9 353.1 156.9 21.4 11.3 15.5 9.2 0.21 0.99 0.348 0.303 4 1.64 0.180 2.66 8.4 14.8 87.8 62.0 274.9 391.0 153.1 11.1 5.2 8.2 6.7 0.10 0.59 0.210 4 1.23 0.331 2.60 55.3 76.7 303.0 153.0 265.6 170.0 73.8 15.4 5.4 7.0 6.7 0.27 1.17 0.519 0.339 4 1.58 0.201 2.70 15.2 26.2 214.1 80.1 243.2 287.4 142.9 14.8 4.4 6.8 6.3 0.16 0.93 0.255 0.237 4 1.59 0.194 2.60 16.8 22.6 98.3 71.4 222.4 350.7 213.2 17.2 7.6 10.4 9.0 0.11 0.56 0.240 0.211 4 1.72 0.164 2.67 4.3 8.7 93.8 60.8 268.8 346.2 188.7 17.3 7.2 9.8 7.3 0.08 0.43 0.192 4 1.36 0.278 2.57 37.2 47.3 145.0 71.7 237.0 274.3 196.4 35.7 12.5 17.1 10.2 0.14 0.59 0.345 0.292 4 1.71 0.158 2.65 6.3 12.3 101.8 65.9 230.6 375.2 169.2 31.3 7.4 9.7 8.9 0.08 0.49 0.190 0.174 5 1.77 0.154 2.77 17.1 2.6 319.2 110.6 152.6 171.6 97.2 84.5 25.5 21.0 17.7 0.15 0.37 0.327 0.194 5 1.74 0.166 2.68 15.8 4.4 282.7 107.7 127.4 209.0 123.7 85.7 26.7 20.7 16.5 0.13 0.39 0.330 0.198 5 1.73 0.168 2.68 14.2 21.1 296.5 121.5 156.7 181.4 129.4 69.9 20.8 14.4 9.7 0.11 0.36 0.274 0.197 5 1.83 0.136 2.69 12.0 10.6 277.3 108.8 152.7 187.6 112.5 78.6 21.6 45.8 15.1 0.13 0.42 0.264 0.186 6 1.89 0.121 2.71 12.5 3.0 223.8 78.8 145.1 186.5 119.2 126.7 50.2 36.5 33.2 0.10 0.23 0.246 0.164 161 Loc MBD WMBD ρS TC OxOM CL FSI MSI CSI VFS FS MS CS VCS AlO FeO LL PL 6 1.98 0.101 2.70 11.2 2.2 160.7 43.2 108.8 169.2 139.5 211.4 89.3 47.4 30.5 0.08 0.15 0.189 0.155 6 1.91 0.120 2.70 11.2 6.3 173.9 68.6 116.2 170.6 143.9 187.7 68.8 41.6 28.7 0.08 0.20 0.207 0.154 6 1.84 0.138 2.77 10.7 2.2 258.4 70.0 119.9 163.5 151.3 144.5 49.2 26.7 16.7 0.12 0.31 0.259 0.166 7 1.50 0.205 2.74 22.7 41.5 142.1 75.8 168.8 50.4 77.0 132.0 116.0 100.0 138.0 0.25 0.65 0.317 0.240 7 1.44 0.214 2.60 29.2 46.1 219.1 84.2 151.1 68.6 85.0 124.0 88.0 79.0 101.0 0.16 0.53 0.289 0.270 7 1.49 0.227 2.56 31.0 46.0 222.3 76.0 120.3 74.5 71.0 128.0 116.0 88.0 104.0 0.37 0.72 0.302 0.216 7 1.01 0.429 2.50 75.8 65.9 268.0 85.6 130.8 72.3 98.8 132.9 59.6 88.6 63.4 0.45 1.11 0.592 0.426 8 1.68 0.168 2.72 12.3 15.2 241.5 87.3 129.0 155.3 88.1 94.9 46.9 63.4 93.6 0.33 0.64 0.289 0.204 9 1.64 0.208 2.76 8.3 10.4 247.7 167.8 269.8 219.3 64.2 15.9 5.7 7.2 2.2 0.19 0.62 0.250 0.214 9 1.71 0.169 2.69 4.2 2.6 346.5 115.2 156.8 176.6 66.2 56.7 25.0 26.7 30.2 0.17 0.33 0.342 0.198 10 1.87 0.136 2.81 4.3 3.2 299.1 72.6 148.7 125.8 72.2 84.5 46.5 58.0 92.7 0.17 0.34 0.293 0.174 10 1.64 0.156 2.63 19.8 24.1 136.1 74.7 153.5 165.4 90.1 111.7 68.3 87.9 112.4 0.20 0.46 0.238 0.222 10 1.58 0.165 2.65 16.5 21.9 120.8 83.3 190.4 181.1 114.8 101.2 33.2 104.8 70.3 0.29 0.55 0.258 10 1.89 0.107 2.65 5.0 8.2 74.1 60.1 129.8 153.8 137.5 200.0 87.9 86.6 70.1 0.09 0.22 0.154 10 1.82 0.125 2.68 5.0 8.6 104.1 67.8 125.7 181.2 154.0 178.0 62.2 68.6 58.4 0.08 0.22 0.166 162 Loc MBD WMBD ρS TC OxOM CL FSI MSI CSI VFS FS MS CS VCS AlO FeO LL PL 10 1.48 0.215 2.65 21.1 28.9 49.5 47.0 148.0 206.8 192.9 134.7 36.5 109.1 75.5 0.32 0.33 0.292 10 1.53 0.189 2.64 11.7 17.8 52.0 43.9 165.3 291.5 210.5 126.1 34.7 50.2 25.8 0.22 0.27 0.247 10 1.53 0.188 2.60 15.0 16.9 41.4 41.9 126.0 167.4 136.6 176.8 97.4 116.1 96.4 0.43 0.34 0.266 10 1.59 0.159 2.61 12.8 16.7 39.6 17.9 92.0 157.9 166.7 175.4 44.5 198.9 107.2 0.26 0.27 0.245 11 1.59 0.201 2.68 14.2 27.8 156.7 98.1 214.1 95.8 107.6 135.7 61.9 76.5 53.8 0.32 0.57 0.247 0.225 11 1.61 0.173 2.69 9.3 14.9 97.9 75.3 195.7 159.1 101.0 134.0 96.0 66.0 75.0 0.26 0.60 0.224 0.203 11 1.51 0.201 2.84 25.4 29.0 163.8 71.3 180.5 120.4 96.0 126.0 103.0 67.0 72.0 0.26 0.62 0.270 0.226 11 1.49 0.221 2.69 17.3 28.8 238.4 103.0 225.1 122.0 75.4 99.1 51.2 52.1 33.7 0.40 0.76 0.304 0.247 12 1.56 0.164 2.62 19.9 24.0 84.8 84.2 192.1 183.9 87.2 107.2 62.0 83.5 115.0 0.21 0.46 0.251 0.241 12 1.35 0.235 2.66 47.0 52.2 255.8 127.1 178.3 143.8 66.3 67.5 40.0 43.3 78.0 0.16 0.57 0.391 0.316 13 1.46 0.215 2.62 34.6 37.8 148.2 81.5 191.5 153.8 96.0 120.0 92.0 61.0 56.0 0.42 0.86 0.259 0.223 13 1.72 0.150 2.65 8.4 14.3 106.5 92.3 191.9 153.3 104.0 127.0 98.0 64.0 63.0 0.29 0.68 0.192 0.168 13 1.37 0.258 2.68 23.7 44.4 461.8 272.3 123.1 52.8 27.3 25.9 12.0 12.7 12.1 0.27 0.60 0.355 0.269 13 1.45 0.289 2.68 11.8 40.6 556.0 202.5 121.0 49.6 22.6 22.8 9.5 9.1 6.9 0.26 0.58 0.395 0.272 13 1.73 0.162 2.75 13.8 23.7 171.7 116.5 255.2 126.7 95.7 99.4 43.4 44.3 47.3 0.15 0.45 0.238 0.181 163 Loc MBD WMBD ρS TC OxOM CL FSI MSI CSI VFS FS MS CS VCS AlO FeO LL PL 13 1.69 0.167 2.69 14.8 24.2 192.5 114.1 241.0 178.1 106.6 83.5 28.7 27.8 27.5 0.17 0.53 0.209 0.173 13 1.87 0.140 2.73 2.7 12.0 214.3 140.7 235.2 163.8 100.2 75.0 25.3 24.0 21.5 0.15 0.46 0.186 0.153 13 1.50 0.220 2.67 27.9 37.0 187.2 183.2 201.7 96.2 70.8 82.2 40.6 71.9 66.2 0.24 0.64 0.265 0.233 13 1.82 0.120 2.66 7.4 13.7 105.1 80.0 211.2 181.6 126.6 132.9 61.7 53.3 47.5 0.12 0.42 0.162 0.155 13 1.96 0.120 2.74 1.8 12.2 149.2 68.0 265.2 131.4 118.8 118.8 46.0 54.1 48.5 0.16 0.53 0.157 0.135 14 1.23 0.359 2.64 45.9 68.6 552.9 131.7 195.4 73.6 23.8 8.1 6.2 5.7 2.5 0.48 1.18 0.508 0.386 14 1.38 0.280 2.71 34.0 47.5 367.4 105.9 206.5 204.4 62.9 19.7 14.1 16.3 2.9 0.29 1.15 0.397 0.305 14 1.37 0.286 2.74 26.2 43.0 545.2 119.0 173.5 105.7 29.3 8.7 7.7 7.7 3.2 0.34 0.95 0.437 0.322 14 1.46 0.267 2.83 14.1 21.9 609.5 120.8 155.8 60.1 20.2 16.0 8.7 7.0 1.7 0.29 0.75 0.542 0.288 14 1.38 0.294 2.77 21.8 40.8 692.5 164.8 62.9 44.5 12.2 10.9 5.7 4.8 1.6 0.32 0.83 0.557 0.323 14 1.44 0.279 2.93 11.8 22.2 702.5 157.4 41.7 69.1 12.8 10.5 2.5 2.3 1.3 0.30 0.62 0.547 0.302 15 1.76 0.145 2.83 8.6 22.8 166.9 67.9 163.3 108.2 92.6 122.5 89.2 92.1 97.3 0.17 0.84 0.232 0.198 15 1.70 0.160 2.72 9.7 18.9 81.6 91.0 221.8 120.6 90.0 119.0 95.0 86.0 95.0 0.14 0.53 0.224 0.194 15 1.73 0.146 2.77 7.3 20.7 76.4 74.4 197.3 164.9 97.0 116.0 88.0 84.0 102.0 0.15 0.63 0.222 0.207 16 1.32 0.292 2.70 22.0 37.1 82.4 95.1 293.1 201.6 110.9 68.0 26.7 47.4 74.8 0.53 0.60 0.387 0.326 164 Loc MBD WMBD ρS TC OxOM CL FSI MSI CSI VFS FS MS CS VCS AlO FeO LL PL 16 1.42 0.215 2.64 15.8 26.2 120.5 110.7 297.1 164.4 97.8 59.1 27.7 48.1 74.6 0.29 0.57 0.315 0.249 16 1.72 0.140 2.77 11.8 22.6 85.7 98.3 237.4 148.6 90.0 106.0 82.0 75.0 77.0 0.20 0.43 0.232 0.192 16 1.48 0.245 2.68 17.4 36.4 173.0 78.8 269.4 110.7 107.0 91.0 57.0 52.0 61.0 0.35 0.74 0.340 17 1.42 0.232 2.61 25.7 23.7 110.0 90.5 245.1 141.5 96.0 84.0 69.0 79.0 85.0 0.25 0.50 0.308 0.300 17 1.63 0.175 2.65 13.8 28.9 185.1 147.6 274.3 139.9 71.7 52.0 29.4 44.3 55.7 0.22 0.43 0.278 0.221 17 1.58 0.202 2.58 13.4 26.3 154.5 103.0 291.8 157.2 84.9 56.3 24.8 55.8 71.6 0.21 0.43 0.276 0.250 17 1.55 0.185 2.65 14.4 27.2 168.1 80.5 279.1 124.1 91.3 68.1 37.2 66.5 84.9 0.17 0.46 0.260 0.236 18 1.47 0.229 2.56 16.2 32.5 101.0 92.1 291.3 120.5 89.0 82.0 54.0 70.0 100.0 0.33 0.51 0.324 18 1.44 0.227 2.73 19.9 29.2 127.8 100.2 284.9 134.8 95.1 61.0 32.9 66.1 97.3 0.38 0.49 0.326 0.285 18 1.44 0.229 2.63 20.8 34.9 153.6 111.1 247.9 145.7 90.5 60.5 32.3 61.4 97.1 0.41 0.57 0.317 0.270 18 1.63 0.160 2.67 11.9 24.1 103.6 122.4 248.8 122.3 80.0 81.0 78.0 81.0 83.0 0.18 0.37 0.237 0.223 19 1.35 0.219 2.62 38.6 40.5 60.0 53.0 190.2 114.9 91.0 120.0 111.0 109.0 151.0 0.40 0.40 0.378 19 1.55 0.143 2.72 31.2 34.1 85.5 45.9 150.7 120.9 97.0 186.0 59.0 125.0 130.0 0.19 0.34 0.268 0.224 19 1.97 0.100 2.81 6.5 9.5 93.1 68.8 284.4 146.7 92.0 100.5 76.0 63.5 75.0 0.12 0.30 0.187 0.167 19 1.85 0.115 2.83 7.9 16.6 65.7 45.2 177.4 143.9 119.2 143.8 87.6 108.0 109.1 0.12 0.30 0.187 0.179 165 Loc MBD WMBD ρS TC OxOM CL FSI MSI CSI VFS FS MS CS VCS AlO FeO LL PL 19 1.84 0.112 2.67 6.9 16.8 158.1 76.8 186.2 131.9 95.0 124.0 88.0 68.0 72.0 0.14 0.28 0.203 0.153 19 1.77 0.121 2.72 10.9 20.5 60.0 63.5 182.5 88.0 97.0 139.0 120.0 116.0 134.0 0.34 0.29 0.216 20 1.24 0.293 2.68 56.4 62.0 140.9 59.6 193.0 93.7 85.0 142.0 95.0 82.0 109.0 0.27 0.37 0.550 0.467 20 1.15 0.315 2.54 59.0 46.8 150.3 64.6 203.1 103.4 115.7 133.7 50.1 72.2 106.9 0.29 0.36 0.562 0.475 20 1.39 0.251 2.48 31.9 55.6 222.0 76.2 184.1 99.8 103.7 112.8 51.8 71.4 78.3 0.21 0.38 0.456 0.336 20 1.50 0.186 2.61 14.5 38.5 195.9 85.7 202.5 91.9 97.0 108.0 75.0 67.0 77.0 0.25 0.37 0.346 0.248 21 1.24 0.200 2.59 28.6 34.9 51.7 59.4 265.1 193.9 119.0 85.0 56.0 69.0 94.0 1.12 0.53 0.435 21 1.27 0.300 2.48 16.4 33.0 60.3 74.4 241.3 164.0 114.0 85.0 67.0 83.0 111.0 1.27 0.54 0.369 21 0.91 0.472 2.35 63.3 57.3 130.3 73.2 295.9 135.5 105.5 65.7 31.6 58.4 103.8 0.46 0.31 0.607 21 1.00 0.389 2.48 58.9 54.7 120.9 80.9 238.1 132.3 103.4 61.9 33.7 71.3 157.6 0.46 0.31 0.569 22 1.55 0.197 2.76 14.0 20.0 19.3 14.0 53.5 62.4 72.4 341.9 176.4 236.7 23.4 0.58 0.30 22 1.57 0.188 2.66 11.0 19.3 65.2 15.9 47.2 31.1 72.9 337.3 236.5 174.5 19.3 0.69 0.34 0.209 22 1.57 0.193 2.97 13.5 22.0 51.1 26.4 65.9 59.5 83.8 273.1 109.3 268.1 62.8 0.54 0.34 0.245 22 1.69 0.163 2.62 7.5 14.2 36.5 34.3 85.3 36.1 104.9 299.8 142.8 196.7 63.7 0.57 0.30 22 1.54 0.211 2.72 10.6 17.2 38.8 36.5 78.4 68.1 132.3 410.1 145.1 70.9 19.8 0.55 0.32 166 Loc MBD WMBD ρS TC OxOM CL FSI MSI CSI VFS FS MS CS VCS AlO FeO LL PL 22 1.60 0.186 2.72 8.8 12.6 67.9 10.4 55.3 37.6 104.5 328.1 162.5 209.0 24.8 0.63 0.32 0.211 23 1.28 0.303 2.68 28.3 26.1 63.2 48.3 214.7 231.4 170.4 152.2 41.1 35.5 43.2 0.83 0.49 0.400 23 1.37 0.240 2.63 23.5 23.3 27.3 43.0 179.6 234.5 189.1 173.9 44.0 47.7 61.0 0.91 0.53 0.293 23 1.07 0.393 2.51 40.2 43.4 62.9 66.4 295.5 239.3 181.3 92.1 22.9 24.6 14.9 0.98 0.73 0.460 23 1.14 0.398 2.62 27.0 45.1 81.9 73.1 266.7 276.0 177.3 82.6 18.2 14.6 9.3 1.24 0.71 0.458 23 1.23 0.310 2.65 29.3 26.7 49.2 46.9 239.5 256.4 160.0 119.0 53.0 35.0 41.0 0.88 0.73 0.398 23 1.29 0.285 2.64 18.8 26.1 39.2 46.3 245.4 234.1 173.1 117.5 31.1 54.5 58.9 1.00 0.86 0.362 23 1.22 0.328 2.58 27.0 24.4 57.3 63.6 213.5 216.2 157.4 91.9 24.9 72.8 102.3 1.25 0.64 0.392 23 1.29 0.290 2.70 18.9 28.4 53.6 55.4 248.7 242.3 155.9 99.0 35.3 52.0 57.9 1.35 0.62 0.330 24 1.42 0.254 2.61 15.0 25.3 75.5 73.4 193.6 131.4 87.0 139.0 94.0 91.0 115.0 0.36 0.31 0.365 24 1.70 0.085 2.75 5.2 12.0 54.9 30.4 149.6 101.2 78.0 168.0 177.0 116.0 125.0 0.12 0.14 0.212 24 1.35 0.271 2.60 21.6 17.9 53.5 52.6 237.3 160.7 111.1 56.7 51.9 117.5 158.6 0.38 0.30 0.422 24 1.62 0.151 2.63 9.4 21.3 43.4 48.9 192.9 175.8 79.1 68.4 97.2 154.6 139.5   0.316 0.253 25 1.36 0.268 2.53 27.6 28.0 213.8 108.5 297.0 154.7 118.0 59.7 17.9 17.0 13.5 0.23 0.19 0.382 0.295 25 1.39 0.244 2.65 38.1 21.8 274.3 93.1 240.0 135.1 108.7 66.7 24.8 23.7 33.4 0.24 0.21 0.381 0.266 167 Loc MBD WMBD ρS TC OxOM CL FSI MSI CSI VFS FS MS CS VCS AlO FeO LL PL 25 1.12 0.343 2.42 50.6 50.3 203.3 102.7 259.7 131.3 112.7 77.7 33.8 39.9 38.8 0.20 0.21 0.482 0.390 25 1.32 0.236 2.51 45.1 30.2 220.5 85.4 210.3 206.2 95.7 68.0 30.0 32.4 51.5 0.22 0.23 0.393 0.308 25 1.47 0.220 2.59 11.7 29.6 198.2 100.7 308.6 169.6 121.7 53.4 17.0 16.0 14.7 0.16 0.13 0.299 0.237 25 1.61 0.179 2.64 10.0 25.8 255.9 91.5 300.7 156.8 98.8 49.5 17.2 15.3 14.3 0.13 0.10 0.306 0.193 25 1.42 0.233 2.53 22.2 25.8 194.3 105.0 280.1 127.6 113.0 85.0 46.0 27.0 22.0 0.17 0.14 0.354 0.268 25 1.43 0.220 2.71 34.8 23.0 269.7 113.7 279.9 121.0 77.7 57.7 23.4 22.2 34.7 0.20 0.17 0.376 0.259 26 1.42 0.241 2.60 18.2 34.3 168.9 77.9 326.4 189.1 104.3 55.8 18.6 20.5 38.3 0.17 0.19 0.344 0.298 26 1.64 0.180 2.61 12.7 33.9 220.6 69.9 287.5 161.4 103.9 59.6 21.0 26.9 49.1 0.21 0.19 0.309 0.223 26 1.51 0.209 2.59 22.1 33.7 183.3 93.9 298.5 180.5 113.3 66.3 22.3 21.6 20.2 0.16 0.21 0.296 0.266 26 1.66 0.161 2.66 11.7 27.5 305.0 99.1 222.9 150.0 92.6 70.6 24.5 18.8 16.6 0.20 0.23 0.286 0.200 26 1.39 0.251 2.65 17.6 32.3 191.9 121.2 379.7 166.3 94.3 26.4 6.6 6.4 7.2 0.21 0.20 0.349 0.292 26 1.47 0.232 2.62 10.0 32.4 229.0 118.1 335.3 170.1 109.6 20.6 5.2 6.0 6.2 0.32 0.22 0.312 0.250 26 1.62 0.194 2.70 52.4 11.4 259.5 108.3 206.5 127.6 101.8 93.2 35.0 34.2 33.7 0.09 0.20 0.293 0.217 26 1.56 0.207 2.75 47.6 15.2 218.5 105.1 237.1 146.7 96.2 91.1 33.1 30.8 41.3 0.10 0.18 0.298 0.225 27 1.07 0.440 2.55 53.4 70.9 406.8 155.9 273.2 75.0 45.2 36.3 4.6 2.5 0.7 0.39 0.47 0.593 0.452 168 Loc MBD WMBD ρS TC OxOM CL FSI MSI CSI VFS FS MS CS VCS AlO FeO LL PL 27 1.15 0.410 2.58 46.7 45.2 488.0 118.2 214.1 78.5 51.9 37.6 6.1 4.8 0.8 0.42 0.86 0.557 0.435 28 0.98 0.503 2.54 77.4 44.5 365.4 126.8 217.7 116.2 30.9 34.0 29.4 58.1 21.5 0.98 1.16 0.614 0.567 28 1.14 0.377 2.71 40.1 29.0 239.5 116.2 184.7 107.9 45.1 62.1 51.1 107.7 85.6 1.56 1.19 0.527 0.443 29 1.38 0.275 2.43 45.1 60.6 297.6 107.0 204.2 160.1 105.2 88.1 19.2 13.4 5.2 0.33 0.73 0.379 0.302 29 1.30 0.261 2.62 45.1 51.7 298.2 96.2 186.7 166.6 119.9 98.4 18.3 10.1 5.4 0.34 0.67 0.398 0.299 30 1.25 0.355 2.57 51.9 45.4 140.0 65.2 222.1 160.3 150.2 181.3 43.0 31.5 6.3 0.33 0.58 0.440 0.341 30 1.41 0.262 2.65 37.1 44.5 230.3 64.5 159.5 166.2 143.7 192.1 28.9 12.2 2.5 0.25 0.62 0.333 0.277 31 1.21 0.362 2.44 47.1 49.2 211.3 87.3 161.3 169.5 165.2 143.2 31.7 26.4 4.2 0.49 0.79 0.458 0.409 31 1.18 0.367 2.68 57.4 51.5 272.3 87.6 165.9 144.7 155.8 139.6 20.3 10.5 3.3 0.54 0.85 0.466 0.443 32 1.14 0.405 2.33 65.4 27.7 146.2 120.0 322.6 128.7 66.8 107.6 50.6 48.1 9.5 0.72 0.77 0.501 0.432 32 1.12 0.377 2.58 76.6 33.7 169.8 109.4 264.7 140.3 87.9 102.7 52.8 54.1 18.3 0.67 0.74 0.537 0.432 33 1.37 0.290 2.37 31.0 43.1 308.7 148.8 246.9 127.1 78.9 54.6 18.3 12.7 4.1 0.29 0.60 0.391 0.292 33 1.37 0.279 2.66 31.1 43.1 274.0 144.7 246.4 156.7 82.1 62.9 18.5 11.8 2.9 0.26 0.57 0.388 0.303 †MBD (maximum bulk density) and ρS (soil particle density), in Mg m-3;  WMBD (critical water content), LL (liquid limit), and PL (plastic limit), in kg kg-1; TC (total C), OxOM (oxidisable organic matter), CL (clay), FSI (fine silt), MSI (medium silt), CSI (coarse silt), VFS (very fine sand), FS (fine sand), MS (medium sand), CS (coarse sand), and VCS (very coarse sand), in g kg-1; AlO (Al-oxide) and FeO (Fe-oxide), in g 100g-1. See Fig. 3.1 for the names of geographic locations. 169 Appendix 3: Relationship between maximum bulk density (MBD) and particle density (ρs) for forest soils collected in British Columbia. ***significant at P < 0.001. P s (Mg m-3) 2.4 2.6 2.8 3.0 M B D  (M g m -3 ) 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 ***360,1473584110102 2 .Rn,ρ..MBD S ==+−= 170 Appendix 4: Relationship between maximum bulk density (MBD) and (a) total C (TC), and (b) oxidisable organic matter (oxOM) for forest soils collected in British Columbia. ***significant at P < 0.001, n=147. oxOM (g kg -1 ) 0 20 40 60 80 M B D  (M g m -3 ) 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 ***0.61R147,n oxOM,0.0116-1.8388MBD 2 === ***0.64R147,n ,e0.9807oxOM0.0009-9651.0MBD 2oxOM-0.0202 ==+= TC (g kg -1 ) 0 20 40 60 80 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 ***0.65R147,n TC,0.01101.7654MBD 2 ==−= ***0.70R147,n,e0.4379TC0.0068567.1MBD 2TC-0.0895 ==+−= a b 171 Appendix 5: Relationship between critical water content (WMBD) and (a) total C (TC), and (b) oxidisable organic matter (oxOM) for forest soils collected in British Columbia. ***significant at P < 0.001. 0 20 40 60 80 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0 20 40 60 80 W M B D  (k g kg -1 ) 0 .0 0.1 0.2 0.3 0.4 0.5 0.6 TC (g kg-1) oxOM (g kg-1) a b ***., RnTC, ..W MBD 6501470040013080 2 ==+= ***0.54 147,n oxOM,0.00390.1114W 2MBD ==+= R 172 Appendix 6: Relationship between maximum bulk density (MBD) and (a) Al-oxide (Al-O), and (b) Fe-oxide (Fe-O) for forest soils collected in British Columbia. ***significant at P < 0.001. Al-O (g 100g-1) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 M B D  (M g m -3 ) 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 Fe-O (g 100g-1) 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 ***36.0,147,4819.06648.1 ***53.0,147,1425.08551.03870.1 2 26600.7 ==−= ==−+= − RnAlOMBD RnAlOeMBD AlO ***17.0,147,3681.06960.1 2 ==−= RnFeOMBD a b  173 Appendix 7: Relationship between critical water content (WMBD) and (a) Al-oxide (Al-O), and (b) Fe-oxide (Fe-O) for forest soils collected in British Columbia. ***significant at P < 0.001. d 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 W M B D  (k g kg -1 ) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 ***24.0,1594.01430.0 2 =×+= RoFeWMBD oAl vs MBD Plot 1 Regr oFe vs MBD Plot 1 Regr 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 0.0 0.1 0.2 0.3 0.4 0.5 0.6 ***44.0),1(2647.00635.0 28270.3 =−×+= ×− ReW oAlMBD ***33.0,1672.01699.0 2 =×+= RoAlWMBD Al-O (g 100g-1) Fe-O (g 100g-1) a b 174 Appendix 8: Relationship between maximum bulk density (MBD) and (a) liquid limit (LL), and (b) plastic limit (PL) for forest soils collected in British Columbia. ***significant at P < 0.001. PL ( kg kg-1) 0.1 0.2 0.3 0.4 0.5 0.6 M BD  (M g m -3 ) 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 LL (kg kg-1) 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 ***72.0,144,753.1085.2 2 ==−= RnLLMBD ***93.0,113,368.0011.2164.1 ***87.0,113,454.2162.2 2121.6 2 ==−+= ==−= − RnPLeMBD RnPLMBD PL a b 175 Appendix 9: Relationship between critical water content (WMBD) and (a) liquid limit (LL), and (b) plastic limit (PL) for forest soils collected in British Columbia. ***significant at P < 0.001. 78.0,6650.00054.0 2 =×+= RLLW MBD 81.0,7559.00276.0 2 =×+= RPLW MBD 0.2 0.3 0.4 0.5 0.6 W M BD  (k g kg -1 ) 0 .0 0.1 0.2 0.3 0.4 0.5 0.6 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.0 0.1 0.2 0.3 0.4 0.5 0.6 PL (kg kg-1) LL (kg kg-1) ***78.0,144,671.00030.0 2 ==+= RnLLW MBD ***89.0,113,756.00280.0 2 ==+= RnPLW MBD a b  176 Appendix 10: Regression analysis between relative height growth index and relative bulk density (RBD) when data were grouped by presence of the surface organic material. Species†, growing seasons (number of observation) Intercept Coefficient and variable R2 P Surface organic material < 3 cm Fd 1 (n=6) --‡ -- -- -- Fd 2 (n=6) 0.85 0.31 RBD-1 0.85 0.009 Fd 4 (n=10) -- -- -- -- Fd 5 (n=6) -7.15 19.15 Log (RBD) + 7.63 RBD-1 1.00 0.008 Fd7 (n=16) -1.20 1.34 RBD-1 0.63 0.000 Pl 1 (n=9) -- -- -- -- Pl 2 (n=9) 0.76 -1.90 Log (RBD)  0.80 0.001 Pl 4 (n=19) 1.08 -0.81 RBD2 0.53 0.000 Pl 5 (n=18) 1.28 -0.68 RBD2 0.22 0.047 Pl 7 (n=10) 1.10 -0.77 RBD2 0.63 0.006 Pl 8 (n=24) 1.73 -1.13 RBD2 0.50 0.000 Sx 4 (n=10) -- -- -- -- Sx 7 (n=10) -- -- -- -- Surface organic material >= 3 cm Fd 1 (n=3) -- -- -- -- Fd 2 (n=3) -- -- -- -- Fd 4 (n=17) -7.86 9.45 RBD2 – 25.82 Log (RBD) 0.38 0.037 Fd 5 (n=3) -- -- -- -- Fd 7 (n=20) 2.38 -2.43 RBD 0.34 0.007 Pl 1 (n=6) -- -- -- -- 177 Species†, growing seasons (number of observation) Intercept Coefficient and variable R2 P Pl 2 (n=6) 5.58 -3.97 RBD2 - 1.71 RBD-1  0.91 0.025 Pl 4 (n=20) 1.83 -1.52 RBD 0.45 0.001 Pl 5 (n=9) -5.11 20.82 RBD - 16.71 RBD2 0.86 0.003 Pl 7 (n=17) -- -- -- -- Pl 8 (n=9) -- -- -- -- Sx 4 (n=17) -2.03 2.66 RBD-1 0.33 0.015 Sx 7 (n=10) -1.53 2.04 RBD-1 0.30 0.023 †Fd, interior Douglas-fir; Pl, lodgepole pine; and Sx, hybrid white spruce. Number following the name indicates number of growing seasons in the field. ‡not applicable. 178 0.0 0.5 1.0 1.5 2.0 2.5 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 Fd5: R2=0.78** a 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 R el at iv e he ig ht  g ro w th  in de x Sx4: R2=0.39*** b Appendix 11: Relationship between relative height growth index and bulk density for (a) Douglas-fir (Fd) at experiment 2, (b) hybrid white spruce (Sx) at experiment 4, and (c) lodgepole pine (Pl) at experiment 1. Number following species name indicates number of growing seasons in the field. Data distributed above the dotted y = 1.0 line indicate better height growth than in the pre-disturbed conditions. **, *** Significant at P < 0.01, 0.001, respectively.                             179 0.0 0.5 1.0 1.5 2.0 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 Bulk density (Mg m-3) Pl5: R2=0.70*** c       180 Appendix 12: GLOSSARY Atterberg limits – water contents of fine-grained soils at different states of consistency, which includes liquid limit (LL), the water content corresponding to the arbitrary limit between the liquid and plastic states of consistency of a soil; plastic limit (PL), the water content corresponding to an arbitrary limit between the plastic and semisolid states of consistency of a soil; and shrinkage limit (SL), the water content above which a mass of soil material will swell in volume, but below which it will shrink no further. Cohesion – the mutual attraction of like molecules that holds a solid or liquid together. It decreases with rise in temperature. Cohesion is different from adhesion which refers to the molecular attraction that holds the surfaces of heterogeneous particles and substances in contact. Compression – a process that describes the increase in soil mass per unit volume or to changes in internal water pressure due to externally applied load. External static or dynamic loads may be applied in the form of vibration, rolling, trampling, etc., while internal forces per unit area may be water pressure or water suction caused by a hydraulic gradient. Consolidation – compression in saturated soils. Compressibility – the resistance against volume decrease when the soil is subjected to a mechanical load. It is described by the shape of the stress-strain curve. Friction – the force resisting the relative lateral (tangential) motion of solid surfaces, fluid layers, or material elements in contact. Kurtosis – a measure of the peakedness of the soil particle size distribution. The kurtosis has a value of 3 for normal distributions, while high value of kurtosis indicates the predominance of a small number of adjacent particle sizes. Landing – areas of cutblocks where harvested trees are processed and loaded onto trucks. Oedometer – an instrument for measuring the rate and amount of consolidation of a soil specimen under pressure. Oxidisable organic matter – the weight difference before and after treatment of the sample with hydrogen peroxide (H2O2) and expressed as the gravimetric fraction of the original oven dry (105 oC) soil mass. It is also referred to as the Loss-on-pretreatment. Particle density – the mass of solid soil particles per unit volume of the solid soil particles. 181 Skewness – a measure of the asymmetry of the soil particle size distribution. The skewness has a value of 0 for normal distributions, while the negative skew indicates a high portion of large particle sizes and the positive skew indicates a high portion of small particle sizes. Tensile strength – the normal force per unit area required to detach or pull apart one section of soil from another. Timber production forests – the forests from which timber has been, or is expected to be, harvested – exclude protected forests, other reserves and forests that are uneconomical for timber production. Void ratio – the volume of pores divided by the volume of the solids of a soil. It generally varies between 0.3 – 2.0. Void ratio is the generally preferred index in soil engineering and mechanics.

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0070825/manifest

Comment

Related Items