Seasonal water relations in native and reconstructed mine soils : implications for ponderosa pine establishment by Karin Marie Jennings A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Soils Montana State University © Copyright by Karin Marie Jennings (1998) Abstract: Reclamation at the Rosebud Mine in Colstrip, Montana is generally considered to be successful based on the establishment and high productivity of cool season grasses. However, survival of ponderosa pine in pine reclamation sites varies between zero and 80 percent, with overall pine survival only 20 to 25 percent. Ponderosa pine frequently die shortly after planting, generally during the first two years. Competition with grasses for limited soil water is believed to reduce pine survival during this period. This thesis focuses on available soil water and competition for available soil water among plant species at the Rosebud Mine. Seasonal soil water status and soil physical and hydrologic properties of six native sites and six reclamation sites are quantified and compared. Secondary objectives were to evaluate whether a more suitable substrate for establishment and survival of ponderosa pine could be created, and if so, to recommend one or more soil profiles. Soil water was measured in the field during parts of the 1996 and 1997 growing seasons with a neutron moisture meter (NMM). Field descriptions of soil profiles and site characteristics were completed. Soil water retention was characterized using a pressure plate apparatus. The ERHYM-It computer simulation model was modified and used to extrapolate beyond the measured seasonal soil water data, using the measured soil physical and hydrologic properties in combination with 34 years of climate data from Colstrip, Montana. Differences in soil physical and hydrologic properties were measured between native and reclamation sites, including higher mean soil bulk density (greater than 1.4 g cm-3) for reclamation sites, which is consistent with the effects of reconstruction practices and the sandier soils. Reclamation sites contained more soil water, especially early in the growing season, than native sites. Despite lower plant available water holding capacity, reclamation sites experienced greater soil water depletion, with more than twice as much as native sites. Grass productivity appeared to be greater on reclamation sites, perhaps related to greater measured soil water contents. Based on the results of this study and relevant literature, several strategies are suggested to create more favorable conditions for survival of ponderosa pines. Establishment and productivity of grasses immediately surrounding ponderosa pine seedlings should be reduced to decrease competition for limited soil water. A soil profile which promotes deeper storage of soil water is generally expected to favor pines rather than grasses. Continued management to control grasses is recommended even for apparently established saplings, especially during periods of lower than average precipitation. Planting ponderosa pine into soil conditions better suited to the production of cool season grasses may hot be the best use of available resources. Adjustment of final bond release criteria may be reasonable, in some instances, to allow species that support an approved post-mine land use to take precedent.  SEASONAL WATER RELATIONS IN NATIVE AND RECONSTRUCTED MINE SOILS: IMPLICATIONS FOR PONDEROSA PINE ESTABLISHMENT by Karin Marie Jennings A thesis submitted in partial fulfillment o f the requirements for the degree of Master o f Science > Soils MONTANA STATE UNTVERSITY-BOZEMAN Bozeman, Montana May 1998 CVV447 APPROVAL of a thesis submitted by Karin Marie Jennings This thesis has been read by each member o f the thesis committee and has been found to be satisfactory regarding content, English usage, format, citations, bibliographic style, and consistency, and is ready for submission to the College o f Graduate Studies. Jon M. Wraith Committee Chair sr-/9~?x Date Approved for the Department o f Plant, Soil and Environmental Science Jeff Jacobsen Interim Department Head ?) i Lih r Date Approved for the College o f Graduate Studies Joseph J Fedock Graduate Dean Date iii STATEMENT OF PERMISSION TO USE In presenting this thesis in partial fulfillment o f the requirements for a master’s degree at Montana State University-Bozeman, I agree that the Library shall make it available to borrowers under rules o f the Library. I f I have indicated my intention to copyright this thesis by including a copyright notice page, copying is allowable only for scholarly purposes, consistent with “fair use” as prescribed in the U.S. Copyright Law. Requests for permission for extended quotation from or reproduction o f this thesis in whole or in parts may be granted only by the copyright holder. Date Signature ACKNOWLEDGMENTS I would like to acknowledge the assistance o f my committee members. Dr. Jon Wraith, Dr. Paul Hook, Dr. Tom Keck and Dr. Roger Sheley, for their technical support and guidance. Special thanks to my major advisor, Dr. Jon Wraith, whose general good nature and availability to answer questions was greatly appreciated. I would also like to acknowledge Dr. Bhabani Das for his assistance with laboratory methods, Mike Mullin for assistance with some especially ugly field work, Greg Millhollin for providing information about the Rosebud Mine and their reclamation practices, and my husband, Stuart Jennings, for his love and encouragement throughout this adventure. Thank you. TABLE OF CONTENTS 1. INTRODUCTION...................... I 2. LITERATURE REVIEW ................................................................................................. 4 Land Reclamation at the Rosebud Mine...................... ........... Previous Research on Pondero$a Pine at the Rosebud Mine Distribution o f Ponderosa Pine on Western Landscapes . . . Site Characteristics ......... 8 Root Distribution o f Ponderosa Pine ........................................................................... 9 Soil Physical Properties and Soil W a te r ................ ............................... .................... 12 Competition Between Pondetosa Pine and Grass for Below Ground Resources............................................................................. 16 3. OBJECTIVES.................... 21 4. MATERIALS AND M ETH O D S.................................................................................... 23 Field Measurement Sites . . . ......................................... ......................................... 23 Neutron Moisture Meter Cahbration and Bulk Density M easurem ents........................................................................................ 28 Soil Water Retention and Plant Available W ate r...................................................... 30 Computer Simulation Modeling o f Long Term Seasonal Soil Water Status ............................................................................... 32 Recommendations for Soil Profile Design ................................i ............................. 40 5. RESULTS AND DISCUSSION .................................................................................... 41 Field Measurement Sites ................. 41 Neutron Moisture Meter CaUbration and Bulk Density M easurem ents........................................................................................ 43 Soil Water Retention and Plant Available W ate r...................................................... 48 Soil Water S ta tu s..................... 52 Computer Simulation Modeling o f Long Term Seasonal Soil Water Status ......................... ............... ............. .. 70 Model Sensitivity A nalysis............................. ...........- .......... ........................ 70 RunoflT Curve Number ........................... 70 Soil Initial A bstraction........................... 71 Effective P recip ita tion ............................................................................... 72 V Page vo t" Transpiration C oefficient........................................................................... 72 Root Depth D istributions......... ........................................................ 73 Model Simulation Results ................................................................................. 74 Model In p u t................................................................................................. 74 Model Output ............................................................................................. 77 Recommendations for Soil Profile Design ...................................................... 106 6. SUMMARY AND CONCLUSIONS ........................................................................... 109 7. LITERATURE C IT E D .............................................................................. 114 APPENDICES ...................................................................................................................... 121 Appendix A— Site Description Forms ........................................................................ 122 Appendix B—Model Predicted Soil Water Contents During the Growing Season of ' Selected Years, for Native and Reclamation Sites (Figures I SE through I SE) .................................................................. 147 Appendix C—Model Predicted Soil Water Status by Wetness Class • (Tables 14A though 1 4 F ).................................................................................... 156 vi vii LIST OF TABLES 1. Summary of site characteristics ........................................................................ 26 2. Soil order and soil series designation for each study s i t e ............................................................................................... 27 3. Summary o f dominant grass species o f Native and reclamation site s ............................................................................... 44 4. Site groupings for neutron moisture meter calibration ........... ...................... 45 5. Mean profile (0 to 90 cm) soil bulk density for each study s i t e .................................................. 46 6. Laboratory measured plant available water holding capacity (PAWHC) for native and reclamation s i te s ...................................... 49 7. Percent effective overwinter precipitation during October I, 1996 to April I, 1997 for each Site, calculated from neutron moisture meter and daily precipitation d a ta ....................... ............................................... 54 8. Selected site characteristics used in Computer simulation modeling ........................................................................ 74 9A. Computer simulation modeling site input File values for native sites .................................. ............................................... 75 9B. Computer simulation modeling site input File values for reclamation s i te s .................................................... .................... 76 IOA Soil water retention results used in native site Input files obtained by fitting measured water retention data to van Genuchten’s (1980) equation ...................................... 78 I OB. Soil water retention results used in reclamation Site input files obtained by fitting measured water retention data to van Genuchten’s (1980) equation ...................................... 79 Table Page viii 11. Correlation o f measured and model predicted soil water contents for 1996 and 1997 ............................................................. 81 12. Summary o f thickness, field capacity water content and wilting point water content based on pressure plate measurements for each soil layer modeled ............................................................................... 87 13. 34-year precipitation summary for Colsttip, MT weather station (no. 1 9 0 5 )................................................................................. 88 14A-F. Model predicted 34 year mean (standard error) number o f days per month within each soil wetness c la s s ........................... ............................................. Appendix C X LIST OF FIGURES Figure Page 1. Field study site location map, Colstrip, Montana ........................................... 24 2. Example o f differences in grass productivity on native and reclamation sites, May 1 1 ,1997 ................................................... 42 3. Mean measured soil bulk density by d e p th ...................................................... 47 4. Pressure plate results: measured depth equivalent plant available soil water holding capacity by depth for each s i te ............................................................. 50 5. Distribution o f 1996 and 1997 overwinter and growing season precipitation, Colstrip, M o n tan a ............................................ 53 6. Mean field-measured soil water content for all native and reclamation s ite s ........................................................................... 56 7. Mean incremental soil water depletion o f near surface soil horizons (0 to 70 cm) .................................................................... 57 8 A. Soil profile (0 to 90 cm) starting water content based on 1996 neutron moisture meter measurements .................................. 59 SB. Soil profile (0 to 90 cm) starting water content based on 1997 neutron moisture meter measurements .................................. 60 9A. Soil profile (0 to 90 cm) starting plant available water based on 1996 neutron moisture meter measurements as a percent o f plant available water bolding capacity . . . ; ............................................................................... 62 9B. Soil profile (0 to 90 cm) starting plant available water based on 1997 neutron moisture meter measurements as a percent o f plant available water holding capacity ........................................................................................ 63 ix X I GA. Comparison o f soil profile (0 to 90 cm) water depletion from native and reclamation sites based on 1996 neutron moisture meter m easurem ents............................................. 65 I OB. Comparison o f soil profile (0 to 90 cm) water depletion from native and reclamation sites based on 1997 neutron moisture meter m easurem ents.................................. .. 66 I I A. Comparison o f soil profile (0 to 90 cm) water depletion from vegetated and non-vegetated reclamation sites based on 1996 neutron moisture meter measurements ........................................................................... 68 I IB. Comparison o f soil profile (0 to 90 cm) water depletion from vegetated and non-vegetated reclamation sites based on 1997 neutron moisture meter measurements .......................................................................... 69 12A. Measured and model predicted soil water content o f native sites during 1996 to 1997 NMM measurement period ............................................................................. 82 12B. Measured and model predicted soil water content o f reclamation sites during 1996 to 1997 NMM measurement period ............................................................................... 83 13. Daily precipitation recorded at the Colstrip, MT weather station (no. 1905) during 1996 and 1997 NMM measurement periods.................................................... 85 14. Distribution of monthly precipitation for selected years, and mean daily precipitation of 34-year re c o rd ........................................................................................ 89 15 A. Model predicted soil water contents during the growing season of selected years, for native site 493D-A with pine ro o ts ......................................................................90 15B. Model predicted soil water contents during the growing season o f selected years, for native site 493D-A with grass r o o ts ............................... 91 v • Xl I SC. Model predicted soil water contents during the growing season o f selected years, for reclamation site 4901-C with pine ro o ts ......... ................................................92 15D. Model predicted soil water contents during the growing season o f selected years, for reclamation site 4901-C with grass r o o t s ...........................................................93 I SE-L. Model predicted soil water contents during the growing season of selected years for native and reclamation sites ..................................................................Appendix B 16 A. Relative root depth distribution o f pine and grass for native site input values........................................................................ 96 16B. Relative root depth distribution of pine and grass for reclamation site input values ............................................................. 97 17A. Model predicted mean monthly number o f days soil water content was equal to or greater than a given matric potential during the growing season (34-year record): native site 121E-C with pine or grass roots.........................................................99 17B. Model predicted mean monthly number o f days soil water content was equal to or greater than a given matric potential during the growing season (34-year record): native site 183E-C with pine or grass ro o ts . ....................................................100 17C. Model predicted mean monthly number o f days soil water content was equal to or greater than a given matric potential during the growing season (34-year record): native site 493D-A with pine or grass roots.................................... ................. 101 I TB. Model predicted mean monthly number o f days soil water content was equal to or greater than a given matric potential during the growing season (34-year record): reclamation site 4888-A with pine or grass roots................................................................... 102 Xll I TE. Model predicted mean monthly number o f days soil water content was equal to or greater than a given matric potential during the growing season (34-year record): reclamation site 3915-C with pine or grass roots................................................................... 103 17F. Model predicted mean monthly number o f days soil water content was equal to or greater than a given matric potential during the growing season (34-year record): reclamation site 4901-C with pine or grass roots................................................................... 104 xiii ABSTRACT Reclamation at the Rosebud Mine in Colstrip, Montana is generally considered to be successful based on the establishment and high productivity of cool season grasses. However, survival o f ponderosa pine in pine reclamation sites varies between zero and 80 percent, with overall pine survival only 20 to 25 percent. Ponderosa pine frequently die shortly after planting, generally during the first two years. Competition with grasses for limited soil water is believed to reduce pine survival during this period. This thesis focuses on available soil water and competition for available soil water among plant species at the Rosebud Mine. Seasonal soil water status and soil physical and hydrologic properties o f six native sites and six reclamation sites are quantified and compared. Secondary objectives were to evaluate whether a more suitable substrate for establishment and survival o f ponderosa pine could be created, and if so, to recommend one or more soil profiles. Soil water was measured in the field during parts o f the 1996 and 1997 growing seasons with a neutron moisture meter (NMM). Field descriptions o f soil profiles and site characteristics were completed. Soil water retention was characterized using a pressure plate apparatus. The ERHYM-H computer simulation model was modified and used to extrapolate beyond the measured seasonal soil water data, using the measured soil physical and hydrologic properties in combination with 34 years o f climate data from Colstrip, Montana. Differences in soil physical and hydrologic properties were measured between native and reclamation sites, including higher mean soil bulk density (greater than 1.4 g cm'3) for reclamation sites, which is consistent with the effects o f reconstruction practices and the sandier soils. Reclamation sites contained more soil water, especially early in the growing season, than native sites. Despite lower plant available water holding capacity, reclamation sites experienced greater soil water depletion, with more than twice as much as native sites. Grass productivity appeared to be greater on reclamation sites, perhaps related to greater measured soil water contents. Based on the results o f this study and relevant literature, several strategies are suggested to create more favorable conditions for survival of ponderosa pines. Establishment and productivity o f grasses immediately surrounding ponderosa pine seedlings should be reduced to decrease competition for limited soil water. A soil profile which promotes deeper storage o f soil water is generally expected to favor pines rather than grasses. Continued management to control grasses is recommended even for apparently established saplings, especially during periods o f lower than average precipitation. Planting ponderosa pine into soil conditions better suited to the production o f cool season grasses may hot be the best use o f available resources. Adjustment o f final bond release criteria may be reasonable, in some instances, to allow species that support an approved post-mine land use to take precedent. I CHAPTER I INTRODUCTION Regulations governing coal mining in the United States require that mined lands be “reclaimed”, or revegetated to a condition as productive or more productive than pre-mine conditions. Federal regulations require the establishment o f “diverse, effective, and permanent vegetative cover o f the same seasonal variety native to the area, or species that support the approved post mining land use” (Section 515.19 SMCRA, 1977). Montana revegetation regulations require that productivity, cover and diversity be similar to the native vegetation (Coenenberg, 1982). The Rosebud Coal Mine, operated by Western Energy Company, is located near Colstrip, located in the eastern Montana ponderosa pine savanna vegetation type (Payne, 1973). Native vegetation surrounding the mine is a mosaic of mixed prairie grassland and pine woodland with Idealized areas o f riparian vegetation. Native landscape is characterized by sandstone ridges and rolling prairies. Sandstone ridges are frequently capped by hard, erosion-resistant porcelanite (heat-fused shale and clay from the roof and floor of burned out coal seams) and are dominated by ponderosa pine (Firmsponderosa Laws. var. scopulorum) trees. 2 Since 1968, Western Energy Company has disturbed about 4,300 hectares (10,500 acres) o f land mining coal at the Rosebud Mine (Montana Power Company, 1995). About 1,750 hectares (4,200 acres) have been revegetated. The majority o f mined lands at the Rosebud Mine will be returned to “multipurpose native vegetation”, including areas o f ponderosa pine. Most of the mined land was native rangeland, providing forage for livestock and wildlife (Coenenberg, 1982). Although final bond release has not yet been sought for any o f the reclamation sites where ponderosa pine establishment is required, reclamation at the Rosebud Mine has generally been considered successful based on the establishment and high production o f cool season grasses and forbs (Keck et al., 1993). Amaximum o f920 permitted hectares (2,272 acres) o f ponderosa pine habitat may be disturbed by the Rosebud Mine (Martin, 1990). While reestablishment o f ponderosa pine is occurring in a few areas o f the mine, mortality is high. Pine survival varies between zero and 80 percent at sites planted with ponderosa pine. The overall average survival rate of ponderosa pine on sites requiring pine is approximately 20 to 25% (personal communication: Pete Martin, Western Energy Company, 1997). High mortality has resulted in five tree densities far below the 40 trees per hectare (100 trees per acre) expected to be required for final, phase HI bond release on most sites (Martin, 1990). Young pine trees in reclamation areas frequently die during the first two years after planting (Martin, 1990; Richardson, 1981). It is during this time when competition for limited resources, primarily soil water, is high and most detrimental (Larson and Schubert, 1969). However, trees may also die when eight to ten years old and thought to be established. 3 Establishing woody species is generally difficult in arid climates. Competition with grasses for limited soil water is believed to limit survival o f ponderosa pine seedlings (Baumbauer and Blake, 1984; Larson and Schubert, 1969; Potter and Green, 1964; Richardson, 1981). Competition decreases when the pines reach sapling stage, apparently due to differences in the spatial distribution o f pine and grass root systems (Lee and Lauenroth, 1994; Potter and Green, 1964). Despite competition with grasses, native stands o f ponderosa pine are expanding away from sandstone/pprcelanite outcrops as individual pines successfully establish, survive and reproduce in areas o f deeper soil that are dominated by grasses. The Rosebud Mine has experienced continued difficulty with establishment o f ponderosa pines in reclamation areas. Poor pine establishment is a great concern to operators o f the Rosebud Mine for several reasons, including repercussions for the overall success o f the reclamation program, economic effects because o f the expense o f repeated plantings, and possible difficulty in obtaining final bond release for a potentially large area o f the mine. The problem of ponderosa pine establishment at the Rosebud Mine is the subject of this thesis. 4 CHAPTER 2 LITERATURE REVIEW Land Reclamation at the Rosebud Mine One o f the fundamental objectives in land reclamation is the rapid establishment o f vegetative cover to stabilize surface soils. Another objective is that this cover be diverse, effective, permanent, and similar to the pre-mine community or a different, but approved post­ mining land use. To achieve these general goals. Western Energy Company currently applies a two-phase seeding sequence at the Rosebud Mine, as outlined by Coenenberg (1982) and summarized in the following paragraphs. The first step in reclamation is redistribution o f the salvaged soil materials, which are primarily stripped and directly hauled to recontoured spoil areas (reclamation sites) or occasionally stockpiled in separate topsoil or subsoil storage areas. Replacement o f soil in a reclamation site occurs either as “single-lift” or “double-lift” . Single-lift refers to the placement o f one layer of topsoil over spoil material (replaced geologic material from below the soil resource and above the coal seam). In contrast, double-lift refers to two layers, subsoil then topsoil, placed over spoil material. The replaced soil materials are then chisel plowed to reduce compaction, break up soil clods, and prepare the surface for seeding. 5 The next step is seeding. Four main seed mixes are used for reclamation at the Rosebud Mine: upland, supplemental, conifer and lowland. “Upland” contains primarily cool-season grasses and forbs. “Supplemental” contains primarily warm-season grasses and forbs. “Conifer” is also comprised o f warm-season grasses and forbs, although a slightly different mixture than the supplemental mix. The conifer mix is only seeded in areas planned for ponderosa pine. Rocky Mountain juniper and/or skunkbush sumac. “Lowland” contains perennial grass species adapted to more mesic environments. Seeding is followed by planting o f shrub or tree species on designated reclamation sites. Seeding occurs in multiple phases depending on the community desired for the site. Efforts are made to seed and plant during early spring when natural moisture is most dependable (April to June) or late fall when seeds or plants are considered dormant. “Upland” and “supplemental” seed mixes are seeded at the same time. “Conifer” and “lowland” seed mixes are seeded alone. Sites with slopes near the maximum (20 percent) are mulched with native grass hay, which helps to temporarily protect the surface from erosion. Ponderosa pine and Rocky Mountain juniper are planted as tubeling or bare root stock with a Vermeer tree spade. These seedlings are propagated by a contracted nursery from locally collected seeds. Various strategies for establishment o f ponderosa pine have been tested by Western Energy Company. These include: planting o f bare root and containerized stock in single-lift and double-lift soils; planting in areas with different seed mixes, in both newly seeded areas and locations where herbaceous vegetation was already established; chemical spraying to reduce competition from grass and forb species; protection from mammal depredation with 6 Martin (1990) summarized Western Energy Company’s preferred strategy for the establishment o f ponderosa pine in reclamation areas as “planting 1-0 containerized pine seedlings grown from locally gathered seeds, with a treeplanter..., into shallow soil (or subsoil) newly seeded with the “conifer” seed mix, followed by an application o f simazine (herbicide) the following spring and cattle grazing two to three years after planting.” The numbers “ 1-0” indicate the age o f the seedling in years since germination, then the number o f times the seedling has been transplanted. In this case, the seedlings are one year old and have never been previously transplanted. Previous Research on Ponderosa Pine at the Rosebud Mine Establishment of ponderosa pine at the Rosebud Mine was evaluated between 1979 and 1987 by personnel o f the University o f Montana’s School o f Forestry (Martin, 1990). This research, performed at the Rosebud Mine, focused on the following topics: o ponderosa pine ecology, including native stand structure and regeneration (Richardson, 1981) o summer climatic influences (Vance and Running, 1985) o establishment and early growth o f seedlings on mine soil (Danielson, 1986) o effect o f grass control on ponderosa pine seedlings (Baumbauer and Blake, 1984) 0 root distribution and shodtroot characteristics on mine soil (Thamams, 1987b), and observations o f root egression o f container stock (Thamams and Blake, 1984) plastic tubing; and protection from moisture loss due to excess radiation with shade cards and “Terra-Mats” (Martin, 1990). Despite these strategies mortality o f young trees remains high. 7 6 heritability o f drought resistance (Riley, 1984) o effects o f seed stratification (Woods and Blake, 1981) ° local genetic variation (Woods, 1982; Woods et al., 1983) and the application of genetic analysis to select seed for reclamation (Woods et al., 1984). According to Martin (1990), the most important discoveries from this research include the occurrence of inherited drought resistance in ponderosa pines o f the Colstrip area (Riley, 1984), the ability to identify specific pine trees with superior survival qualities (Woods, 1982), and the documentation o f extensive root development unimpeded by planting technique, materials used or reconstructed minesoils (Thamarus, 1987b). Distribution o f Ponderosa Pine on Western Landscapes Throughout the western United States ponderosa pine grows under a wide variety of ecological conditions (Schubert, 1974). However, in southeastern Montana and in North Dakota, ponderosa pine occurs almost exclusively on the top o f knolls or exposed sandstone or porcelanite (Potter and Green, 1964). These landscape features are interrupted by valleys o f deeper, finer textured soils. Although sandstone or porcelanite outcrops are the primary locations o f ponderosa pine, pine are slowly encroaching into areas o f deeper soil, away from outcrops in locations with a constant nearby seed source (Potter and Green, 1964; Richardson, 1981). Historically, periodic fires probably confined the pines to outcrops. Among other factors, the relatively recent practice o f fire suppression may contribute to the observed encroachment and apparent shift in the position o f ponderosa pine on the landscape. 8 Site Characteristics Ponderosa pine in the Colstrip area are mostly found on coarse textured soils or rocky substrates. The majority o f ponderosa pine trees on native sites at the Rosebud Mine grow in areas with 50% or more rock in the substrate (Stout, 1980). Blake and Running (1986) also stated that most native pine stands are found on coarse textured soils, ofteii with large rock fragments. Richardson (1981) found native ponderosa pine near Colstrip to be “largely associated with Entisols and to a lesser extent Aridisols” (Aridisols in this area have since been reclassified as Inceptisols). Entisols and Inceptisols are generally weakly developed, rocky or skeletal and often erosive. Potter and Green (1964) observed that a sandy soil, with deeper and more rapid penetration o f rainfall, favors the establishment o f pine Seedlings over that in heavier silts and clays. In a study at the Rosebud Mine, Stark (1985) found no significant soil texture differences between reclamation areas and undisturbed forest soils. Stark found that percent clay of replaced topsoil and subsoil in reclamation areas was highly variable, averaging 28% clay, 44% sand and 28% silt. Native forest areas averaged 25% clay, 40% sand and 35% silt. In comparison. Keck and Wraith (1996) found reconstructed soils at the Rosebud Mine to have lower mean percent clay, ranging from 23 to 25 percent. Topsoil, subsoil and spoil horizons had mean clay contents o f 23%, 24% and 25%, respectively. These clay contents are most similar to the percent clay o f native soil found by Stark (1985). Differences in results for percent clay o f reconstructed soils may relate to the area sampled, number o f samples taken and method o f analysis. Keck and Wraith (1996) analyzed 174 samples from 9 a 30 hectare (75 acre) portion o f Area E o f the Rosebud Mine, whereas Stark (1985) took 240 samples from the much wider, and unspecified, “Colstrip area.” In summary, the overall range o f textures in reconstructed soils at the Rosebud Mine is not different than in native soils from which they were constructed, though there are abrupt textural changes between horizons in reconstructed soils compared to native soils (Keck, 1993). Root Distribution o f Ponderosa Pine The distribution o f ponderosa pine roots varies with substrate. Cox (1959) compared the distribution o f 51- to 78-year old ponderosa pine tree roots growing in three soil types having different textures. He found the greatest number o f roots in a medium-textured (silt and clay loam) soil and the smallest number in a fine-textured (clayey) soil. Over 70% of the measured roots were in the upper 61 cm (24 in) o f soil for all textures. In all soils, roots penetrated at least 1.2 m (4 ft) deep, which was the extent of excavation. A larger component o f understory vegetation present at a site resulted in lower density o f pine roots within the A- horizon(Cox, 1959). Studies o f 30- to 70-year old ponderosa pine tree roots on native sites near the Rosebud Mine also indicated that the majority o f fine and lateral roots are located within the upper 46 cm (18 in) o f the substrate (Stout, 1980). He further observed that roots tend to grow deeper where the substrate is fractured rock or coarse to medium textured soils. Curtis (1964) also observed many roots o f a 60 year old pine in the central Idaho area within cracks and crevices in the bedrock or hardpan. Zwieniecki and Newton (1994) found at least one quarter to one 10 third o f the total root length o f 12 year old ponderosa pine roots located within the metasedimentary rock layer in southwest Oregon. Potter and Green (1964) observed differences in the distribution o f ponderosa pine (predominantly 20 to 50 years old) roots on sandstone or porcelanite outcrops and downslope areas in southwestern North Dakota. Pine roots on outcrops were confined to the horizontal and vertical cracks in the platy porcelanite, extending to great depths (observed to 7.6 m [25 ft] below the surface). On downslope positions, where the amount o f fine soil material increased, root systems were more extensive and widespread but not as deep. Even further from the outcrops the depth o f fine soil exceeded the depth of penetration o f moisture. This resulted in a permanently dry subsoil lacking roots. Fibrous grass roots dominated the top 20 cm (8 in) o f soil in these areas. Thamarus (1987b) excavated six-year-old pine seedlings from reclamation areas at the Rosebud Mine and found most fine roots, laterals, and mycorrhizal associations were within the topsoil, which varied in depth from 20 to 30 cm. Lateral roots were observed to extend up to one meter from seedling stems. Taproots o f the seedlings had grown through the subsoil into the spoil to depths greater than one meter (39 in). Based on measurements o f several thousand 1-0 and 2-0 ponderosa pine seedlings obtained from nurseries, the average extent (depth) o f the root system before transplanting was 22.5 cm (9 in) and 27.7 cm (I I in), respectively (McDonald and Fiddler, 1989). Therefore, during a six year period the roots o f ponderosa pine seedlings may extend more than 70 cm (28 in) in length, depending on conditions. 11 Comparison of root distribution of one to three year old ponderosa pine seedlings in soils derived from metamorphic and limestone (sedimentary) parent materials indicated deeper root penetration on the coarser textured metamorphic soil (van Haverbeke, 1963). Average penetration for one, two and three year old seedlings on metamoiphic soil was 29 cm, 31 cm and 34 cm, respectively. Average root penetration for seedlings on the limestone soil was 26 cm for all three ages. Van Haverbeke (1963) attributed much o f the difference in rooting depth in the two soils to a denser subsoil (no soil bulk density provided) and more abundant grass cover for the limestone soil Soil texture and bulk density have been shown to affect root growth arid development o f pine seedlings. In soils with high percentages o f silt and clay lateral and vertical root development as well as growth of pine seedlings was restricted (Potter and Green, 1964; van Haverbeke, 1963). Seedling growth at the Rosebud Mine was observed to be most rapid on soils with greater than 50% sand, haying a compacted horizon at 60 cm (24 in) depth (Stark, 1985). Increased soil bulk density (1.12 g cm"3 compared with 0.80 g cm"3 within the upper 30 cm [12 in] of soil) reduced young ponderosa pine stand volume and reduced annual shoot growth by 43% in two year old seedlings and 13% in 15 year old trees (Helms, 1983). Thamarus (1987b) observed that soil layer interfaces and localized areas o f soil compaction in reclamation areas at the Rosebud Mine did not impede growth or development o f pine seedling roots. This review indicates substantial pine root plasticity and the importance o f soil structure and density. 12 Soil Physical Properties and Soil Water It is well documented that plant available soil water is a primary factor limiting the establishment and growth o f ponderosa pine in many locations (Heidmann and King, 1992; McDonald and Fiddler, 1989; Richardson, 198.1; Riegel et al., 1992; Running and Danielson, 1984; Schubert, 1974; Shainsky and Radosevich, 1986; Stark, 1982, 1985). Competition for soil water is apparently most detrimental when ponderosa pine trees are young and an extensive root system has not developed. The highest tree mortality typically occurs during the first one to two years after planting (Larson and Schubert, 1969; Martin, 1990; Richardson, 1981; van Haverbeke, 1963). One of the more important variables related to the distribution o f soil water, in addition to texture, is soil structure. The original structure o f disturbed soil, which had generally developed over hundreds or thousands o f years, may be altered or destroyed (Schafer et al., 1979). Disturbance o f soil structure can affect aeration and soil water retention and movement and therefore plant growth. Potter et al. (1988) stated that soil structural units and the associated interaggregate pore spaces are among the most important soil properties disrupted during mining and reclamation because of their importance for root penetration and growth. Soil bulk density is closely related to soil structure, texture, porosity, aeration and water­ holding capacity. Bulk density can indicate the relative degree p f soil compaction, the resulting changes in ability o f a soil to transmit water and gases to plant roots, and the ability of roots to penetrate the soil. In natural soils, bulk density generally increases with depth as 13 organic matter content, root and biotic activity, and porosity decrease (Sutton, 1991). A soil may naturally possess high bulk density, depending on the soil texture (e.g., bulk density of sand is about 1.6 g cm'3) and particle arrangement. Presence o f rocks and sand in soil favors high bulk densities, whereas the content o f fine fractions favors relatively low bulk density and high total porosity. Reconstructed soils generally have an artificially higher bulk density than the native soil material because o f compaction due to the size and weight o f equipment used for soil salvage and redistribution. Reconstructed soils also generally lack large pore spaces, such as those between peds. The bulk densities o f native soils o f the Colstrip area are not well characterized. Stark (1985) reports a native soil bulk density o f 1.03 g cm"3, but this value seems too low for the types o f weakly developed, often rocky soils (Entisols and Inceptisols), around Colstrip. Mean bulk density for reconstructed topsoil, subsoil, and spoil materials in Area E o f the Rosebud Mine has been reported as 1.54, 1.67 and 1.79 g cm'3, respectively (Keck and Wraith, 1996). Also at the Rosebud Mine, Keck (1993) determined that bulk density o f spoil material in Area A ranged from 1.6 to 1.95 g cm"3 after reclamation was complete. At a given soil texture, higher soil bulk density generally results in lower soil water holding capacity and also lower plant available water. Penn et al. (1987) measured soil water content and matric potential on a restored mine and an undisturbed site. They found significantly lower plant available water in both topsoil and subsoil layers on the reclaimed site compared to the undisturbed site. They attributed this result to the loss of soil structure, specifically the reduced volume o f mesopores, and suggested that drought is more likely to occur on the reclaimed than undisturbed site. Sharma and Carter (1993) observed that matrix 14 and preferential water flow rates o f pre-mine compacted soils and post-mine reclaimed soils were about one to two orders o f magnitude lower than those o f undisturbed pre-mine soils. In addition, they noted that the redistribution o f soil and spoil materials results in discontinuity o f pores at soil layer interfaces. Saturated hydraulic conductivity (Ks) is a measure o f water flow rate in soil under water- saturated conditions, and is strongly influenced by pore size distribution and pore continuity. Mean Ks was determined to be 25 to 50 percent less for reconstructed soils than for native soils at the Rosebud Mine (Hepfher et al., 1996). Saturated hydraulic conductivity decreased with soil depth in reconstructed soils. Hepfher et al. (1996) stated that decreased Ks was likely due to higher measured bulk density in these lower layers. Measured Ks at the Glenharold Mine in North Dakota indicated similar results, with Ks o f the reconstructed topsoil material 25 percent lower than that of the undisturbed A horizon. Saturated hydraulic conductivity o f the reconstructed subsoil materials was less than 10 percent that o f an undisturbed B horizon (Potter et al., 1988). Reconstructed topsoil and subsoil layers at the Rosebud Mine had a narrower pore size distribution than native soils (Hepfher et al., 1996). The narrower pore size distribution V indicates lack o f structural development in the reclaimed sites, and was consistent with bulk density measurements. Furthermore, the mean effective water transmitting pore size was greater for native than reclaimed topsoil layers (Hepfher et al., 1996). Potter et al. (1988) likewise found significant differences in pore volume distribution between undisturbed and reconstructed soils, with the greatest difference observed in pores having radii >15 pm, and most evident at lower profile depths. 15 The effects on soil water of variable topsoil and subsoil thicknesses over spoil materials have been studied in reclaimed mine lands. Other factors being equal, thicker topsoil and subsoil replacement depths will store more water than shallower replacement depths (Power et al., 1981; Stark and Redente, 1985). However, Schroeder (1995) found no significant difference in soil water content between native and reclaimed sites in North Dakota, even though native sites had uneven depths and reclaimed sites had uniform depths of topsoil and subsoil replaced. The amount of water in a given soil was site specific, depending on climate, soil textural and structural attributes, and reclamation practices. Topography is another important factor that influences soil water content. Wollenhaupt and Richardson (1982) found that concave slopes accumulated more soil water than convex sites. Greater Soil water was also found at downslope positions (Schroeder, 1995). Greater plant-available soil water holding capacity was measured at top and middle slope positions in native soils in the Colstrip area than for the bottom slope position (Hepfiier et al., 1996), based on the “m” parameter o f the van Genuchten (1980) soil water retention model. The “m” parameter is related to the width o f the pore size distribution, and potentially to soil structural development. Although this result might appear to be counter to expectations, it was hypothesized that the result of erosive deposition of soil particles from up slope positions cduld fill in larger pore spaces in bottom slope positions. This results in narrower pore size distribution which influences soil water retention and availability to plants. Rock materials can also influence water availability. Numerous studies have concluded that various types o f underlying weathered rock are an important source o f plant-available water for ponderosa pine and other species when surface soils are dry (Arkley, 1981; Jones 16 and Graham, 1993; Stark, 1983; Wang et al.„ 1995; Zwieniecki and Newton, 1994). Wang et al. (1995) determined that bedrock and dispersed rock fragments o f slightly metamorphosed sandstone, siltstone, and shale provided an important storage reservoir. Bedrock in their study held 71 to 80% o f all water stored at >-2.0 Mpa (-20 bar) matric potential between the surface and 3.0 m depth. Competition Between Ponderosa Pine and Grass for Below Ground Resources Cofnpetitive success is often determined by early resource capture and the capacity o f a species to maintain productivity in a competitive environment (Shainsky and Radosevich, 1986). Water is required for most plant physiological processes. Soil water is one o f the most frequent controls o f plant growth and community structure, especially in arid or semi- arid environments (Coffin and Lauenroth, 1991). Soil water has been repeatedly identified as a primary factor limiting the establishment and growth of ponderosa pine (Richardson, 1981; Riegel et al., 1995; Running and Danielson, 1984; Schubert, 1974; Shainsky and Radosevich, 1986; Stark, 1982, 1985). Competition for available soil water is most critical during the first one to two years after planting (Larson and Schubert, 1969; Martin, 1990; Richardson, 1981; van Haverbeke, 1963). During this period the root systems of pine seedlings are not well developed, and most photosynthate produced is used for taproot growth and extension (Richardson, 1981). Significant development o f lateral roots generally does not occur until the vertical roots reach a zone o f available soil water (McDonald and Fiddler, 1989). Usually, this is not until the second or third growing season, when lateral roots may double or triple in length (van Haverbeke, 1963). 17 Limited root growth o f ponderosa pine was observed where interspecific competition, especially from grasses, was heavy (vanHaverbeke, 1963). Interspecific competition was also attributed to a decrease in secondary lateral roots within the top 46 cm (18 in) (Curtis, 1964). Baumbauer and Blake (1984) observed significantly greater seedling growth rates after competing vegetation, primarily grasses, was removed by chemical application at the Rosebud Mine. This observation has been supported by many others (McDonald and Fiddler, 1989; Riegel et al., 1992; Riegel et al., 1995; Sands and Nambiar, 1984). Potter and Green (1964) noted that once ponderosa pines reached the sapling stage interspecific competition with grasses was significantly reduced. The shifr in competitive advantage from grasses to pines near the sapling stage is largely attributed to greater development o f root systems by this stage. The extensively branching root systems o f woody plants, including ponderosa pine, allows nearly exclusive access to the deeper soil layers compared to grasses (Lee and Lauenroth, 1994). The intensive root systems o f grasses are more adapted to exploiting resources concentrated in smaller volumes nearer the soil surface. A comparison of the spatial distribution o f grass roots and roots o f woody species provides an explanation for the dominance o f grass in the shortgrass steppe ecosystem. Limited and variable water availability in this ecosystem and the concentration of precipitation in the summer lead to the majority o f water remaining in upper soil horizons (Lee and Lauenroth, 1994). Woody plants in semiarid regions are favored by conditions that promote storage o f water deep in the soil. Grasses are generally favored by water being available mainly in upper soil layers during the growing season. 18 Soil depth pan affect inter- and intraspecific competition. For example, Sheley and Larson (1995) observed that unrestricted soil depth permitted resource partitioning between species, with intraspecific competition having the most significant influence on each species. Yellow starthistle (Centcmrea solstitialis L.) had an advantage over cheatgrass (Bromus tectorum L.) in deep soils because its taproot morphology enabled continued resource uptake When adequate deep moisture was available and surface soils had dried. However, under conditions of restricted soil depth, interspecific competition was more significant. Apparently the relatively shallow, fibrous rooting system o f cheatgrass was better suited for resource capture in shallow soil. In addition to spatial partitioning of resources, temporal partitioning is also an important aspect of inter-plant competition. Zwieniecki and Newton (1994) noted temporal and spatial differences in root function o f conifers. Shallow roots absorbed water only during the wet season. After the surface soil dried to the wilting point, the deeper roots became the main source o f water and nutrients. Cool season and warm season grasses are known to have different periods of growth, which result in different levels o f competition between ponderosa pine and each grass type (Larson and Shubert, 1969). Western Energy Company is taking advantage o f this strategy by seeding warm season grasses before cool season grasses to encourage successful establishment o f the warm season grasses, and also by only seeding a variety o f warm season grasses and forbs into areas to be planted with ponderosa pine (Coenenberg, 1982; Martin, 1990). Despite the well-documented interspecific competition between ponderosa pine and grasses, native stands o f ponderosa pine are successfully moving into areas o f deeper soils 19 dominated by grasses, away from the sandstone/porcelanite outcrops in the Colstrip area (Potter and Green, 1964; Richardson, 1981). Although many situational distinctions exist between pines locally expanding from areas o f high pine density onto deep soils and planting pines in reclamation sites, the above observations suggest a potential for establishment o f ponderosa pine in reclamation areas at the Rosebud Mine. However, pine mortality continues to be high and if it continues will be detrimental to the overall success o f reclamation efforts. Specifically a problem exists in obtaining final bond release for acreage designated for ponderosa pine establishment. What factors may allow the expansion o f native stands o f ponderosa pine, yet still restrict the establishment o f ponderosa pine in similarly deep reconstructed minesoils? According to Archer (1989), “quantitative and historical assessments suggest that woody-plant abundance has increased substantially in arid and semiarid grasslands over the last 50 to 300 years in many parts o f the world.” Partial explanations for the conversion of savannas to woodlands within the past century are fire suppression, overgrazing, and climatic changes, which have interacted in complex ways (Archer, 1989). The expansion of native ponderosa pine stands is greatly enabled by the presence of mature trees which provide a nearby and constant seed source, a canopy for shading, and other benefits to seedlings, including root/mycorrhizal interactions (Richardson, 1981). Nearby trees may also improve the nutrient conditions in their immediate surroundings, providing more suitable conditions for the establishment o f ponderosa pine seedlings (Skarpe, 1992). The balance between grasses and woody vegetation may be regulated by the ratio o f topsoil to subsoil wetness (Archer, 1989). Factors reducing the ratios o f topsoil to subsoil 20 water could cause savannas to develop into woodlands (i.e., reducing topsoil wetness and/or. increasing subsoil wetness). Climatic influences which could increase topsoil/subsoil water ratio include an increase in annual rainfall, shifts from small, frequent precipitation events to large, infrequent events, and/or a shift toward increased winter precipitation (Archer, 1989). In addition, grazing limits the ability o f grasses to competitively exclude the invasion and establishment o f woody vegetation by decreasing the transpiring leaf area and root initiation and extension o f grasses, which subsequently decreases their ability to take up water. Grazing of grasses may therefore enhance percolation o f water to the subsoil (Archer, 1989; Skarpe, 1992), increasing the availability of soil water for the more extensive pine roots. Grazing can also increase surface soil wetness, enhancing woody seedling establishment and growth (Skarpe, 1992). 21 CHAPTERS OBJECTIVES At the Rosebud Mine, a maximum of 920 permitted hectares (2,272 acres) o f pine habitat may eventually be disturbed (Martin, 1990). Western Energy Company is required by federal and state laws to reestablish ponderosa pine over some portion o f this disturbed area. Much effort has gone toward pine establishment, yet mortality rates continue to be high, resulting in tree densities far below the 40 trees per hectare (100 trees per acre) expected to be required for final (phase HI) bond release (Martin, 1990). A variety o f factors appear to affect the successful establishment o f pines, including: I) available soil water and competition for available soil water among plant species; 2) site suitability (slope, aspect, soil physical and chemical characteristics, etc.); 3) handling o f seedlings; 4) planting technique and timing; and 5) animal and insect predation. This study is focused on the first, and to lesser extent, the second factors listed. The objectives o f the study were to: I) quantify and compare soil physical and hydrologic properties o f selected reclamation and native sites at the Rosebud Mine; 2) compare seasonal soil water status in reconstructed soil profiles to that o f selected native sites supporting ponderosa pines; 3) evaluate whether a more suitable soil substrate for ponderosa pine tree establishment might be reasonably constructed (i.e., considering economics and the 22 current regulatory laws), I f the outcome o f this objective is positive, one or more reconstructed soil profiles having' physical and hydrologic properties potentially more conducive to the establishment and survival o f ponderosa pines will be recommended, using materials and resources reasonably available to the Rosebud Mine. I hypothesized that seasonal soil water status and soil physical and hydrologic properties are different for reconstructed mine soil and native soil. Furthermore, my intentions were to provide suggestions for at least one reconstructed soil profile that could be reasonably implemented at the Rosebud Mine to improve establishment and survival o f ponderosa pines. 23 CHAPTER 4 MATERIALS AND METHODS Field Measurement Sites Native and reclamation sites were selected based on pre- and post-mine soil surveys, air photograph's, and mine planning maps. Native and reclamation sites were located within or immediately outside Areas A, B and C of the Rosebud Mine (Figure I). The objectives o f site selection were to: I) encompass a range of soil textures present in the area both naturally and following mining and reclamation; 2) include some reclamation sites in which pine establishment appears successful; 3) select reclamation sites that were greater than three years old; and 4) select sites with reasonable access. A total o f 12 sites were selected; 6 in native areas, and 6 in reclamation areas. The term ‘native’ as used here means undisturbed by mining activities. All o f the native sites supported ponderosa pine trees. Half o f the reclamation sites supported ponderosa pine regeneration. Numerical references given native sites are those o f the soil map units identified in the Rosebud Soil Survey for these sites (Soil Survey Staff, draft manuscript), followed by a letter designation for the area o f the mine it is located in or adjacent to, then a direction to differentiate sites within similar soil series. For reclamation sites, references are those used Figure I . Field study sites located at Colstrip, Montana. 25 by the Rosebud Mine to identify specific fields, similarly followed by a letter designation for the area o f the mine in which it is located. The Rosebud Mine’s numbering system tracks when and where reclamation occurred. The first digit indicates the quarter-year and the second and third digits indicate the year o f reclamation. The fourth digit sequentially identifies all the fields reclaimed during that quarter. The native sites were purposely not located on sandstone/porcelanite outcrops because o f the difficulty o f installing neutron access tubes in this rock substrate and the lack o f a comparable substrate in reclamation areas, both currently and in the future. Sampling native sites focused on substrates supporting ponderosa pines that were more similar to substrates that are and could potentially be reconstructed following mining. Soil and landscape features at each site Were characterized through direct sampling following the standard soil profile description and classification process used by the USDA Natural Resource Conservation Service (NRCS). Observations o f soil properties included horizon thickness, color, texture, structure, soil consistence (workability), lime (calcium carbonate), root and pOre distributions, rock fragments, and soil pH. Profile descriptions were accomplished by excavating one 90 cm soil pit at each site, and using a bucket auger to sample an additional 90 cm to a total depth o f 180 cm where possible. Soils were classified to the subgroup level based on criteria in Keys to Soil Taxonomy (Soil Survey Staff, 1994). Landscape was characterized by identifying the dominant plant species at each site, determining canopy coverage classes by species from ocular estimates, and measurement o f slope steepness and slope direction (aspect). Table I summarizes some of the main characteristics o f the selected sites. Table 2 Table I. Summary of site characteristics. Site Name/ Reference Number Site Type Location Dominant Land Use Year Reclaimed Pine Trees Present Dominant Soil Texture of Profile (weighted average) % Hard Coarse Fragments (70-90 cm) 121E-C(s) Native Area C Pine Woodland N/A Yes Loam 15 121E-C(n) Native N: of Area C Pine Woodland N/A Yes gravely Loam 30 183E-C(s) Native Area C Pine Woodland N/A Yes Loam/Silt Loam 20 I83E-C(n) Native N. of Area C Pine Woodland N/A Yes Silt Loam 2 493D-A(e) Native Area A . Rangeland/Open Pine Woodland N/A Yes Clay Loam/ Silty Clay Loam trace 493I>-A(w) Native Area A Rangeland/Open Pine Woodland N/A Yes Silty Clay/Soft Sandy Shale trace 4888-A Reclamation Area A Pine Reclamation/ Rangeland 1988 Yes Silty Clay Loam/ Sandy Clay Loam 28 4822-B Reclamation Area B Reclamation Rangeland 1982 No Sandy Clay Loam 30 2856-B Reclamation AreaB Reclamation Rangeland 1985 No Sandy Clay Loam- varied: SL, LS1 SCL1 SiCL 40 4881-C Reclamation Area C Reclamation Rangeland 1988 No Loam 40 4901-C Reclamation Area C Pine Reclamation/ Rmigeland 1990 Yes Clay Loam/ Sandy Loam 30 3915-C Reclamation Area C Pine Reclamation/ Rangeland 1991 Yes Loam/Sandy Loam 20 27 presents the soil classification for profiles sampled to the family level. Appendix A provides site description forms with riiore specific information for each site, including landscape and vegetation information and detailed soil descriptions. Table 2. Soil order and family designation for each study site. Site Reference Soil Order Family Designation Native Sites 12IE-C (s) Inceptisol fine loamy, mixed, frigid Aridic Ustochrept 121E-C (n) Inceptisol fine loamy, mixed, fiigid Aridic Ustochrept 183E-C (s) Inceptisol fine loamy, mixed, frigid Aridic Ustochrept 183E-C (n) Inceptisol fine loamy, mixed, Aridic Ustochrept 493D-A (e) Inceptisol fine, montmorillonitic, fiigid Aridic Ustochept 493D-A (w) Entisol fine, montmorillonitic (calc.), fiigid, shallow Aridic Ustorthent Reclamation Sites 4888-A Entisol fine loamy, mixed (calc.), fiigid Aridic Ustorthent 4822-B Mollisol fine loamy, mixed Aridic Haploboroll 2856-B Mollisol fine loamy, mixed Aridic Haploboroll 4881-C Entisol fine loamy, mixed (calc.), fiigid Aridic Ustorthent 4901-C Entisol fine loamy, mixed (calc.), frigid Aridic Ustorthent 3915-C Entisol coarse loamy, mixed (calc.), fiigid Aridic Ustorthent Based on qualitative descriptions, the most distinct site characteristic is land use (Table I). All native sites selected have ponderosa pine trees present and are considered woodlands or open woodlands. None are grazed by cattle, but they do experience grazing by deer, elk and antelope. Reclamation sites are all classified as rangeland, with those sites havitig 2 8 ponderosa pine trees present classified additionally as pine reclamation. All reclamation sites experience periodic grazing by cattle, with grazing leases overseen by Western Energy Company. Neutron Moisture Meter Calibration and Bulk Density Measurements Following site selection, twb neutron moisture meter access tubes were established at each site to measure seasonal soil water status. Access tubes were installed May 14 and 15, 1996 to a depth o f two meters, wherever possible. Shallower installation depths, were sometimes required because o f hard bedrock material. Five-cm (2 in) diameter, thin-walled polyvinyl chloride (PVC) pipe was used for access tubing. Soil water content was measured in the field using a neutron moisture meter (NMM) (CPN Model 503DR), every 20 cm (8 in) to a depth o f up to 2 m (6.5 ft). Neutron count ratios were taken at two week intervals (except for one three week interval) during May 15 through September 12,1996, and at three week intervals during March 30 through May 11 ,1997. Atotal of 12 NMM access tubes were installed in the six native sites and 12 tubes in the six reclamation sites, with two tubes ‘paired’ a few meters apart at each site. Within each o f the six reclamation sites, soil water status was also assessed without live vegetation. To accomplish this, two additional access tubes were installed at each reclamation site followed by repeated spraying with Roundup® (Glycophosphate, N-[phosphonomethyl] glycine) to remove live vegetation within an approximate 1.5 m radius of the access tubes. The purpose was to determine the seasonal soil water status in reconstructed soil profiles without the influence o f established vegetation, primarily grasses. 29 The neutron moisture meter was calibrated by collecting soil samples for gravimetric analysis o f volumetric water content. NMM count ratios were measured at depths o f20, 40, 60, and 80 cm just prior to collecting soil samples at the same depths adjacent to the access tubes. Measured soil water content was then related to the field count ratio collected by the NMM to derive the tube-specific calibration. Soil samples were collected June 24 and 25, 1996 for calibration at the.“wet” end, and again September 12 to 14, 1996 for calibration at the “dry” end o f the soil water content range. Soil samples were collected by pressing or carefully pounding an aluminum ring o f known volume (152 cm3) into the soil. The ring was carefully removed to obtain the soil sample. It was occasionally necessary to chisel away soil surrounding the ring in order to remove the ring with an intact core. Soil cores were placed in labeled plastic Ziplock® bags, stored in a cooler containing ice and weighed within 12 hours. Wet soil mass was obtained with a Sartorius B3100P balance, The soil was later oven-dried for at least 24 hours at 105° C, and dry soil mass obtained from the same samples. Volumetric water content (6) was calculated as the volume of water per bulk volume of soil. Volume o f water was obtained as mass of moist soil minus mass o f oven-dry soil, divided by density o f water. Soil bulk density was calculated as the ratio o f the mass o f oven-dry soil to its bulk volume, which assumes a core having dimensions equal to the sampling ring. Calibration o f the NMM to each soil material was accomplished by linear regression o f measured volumetric soil water content against field measured count ratios. The coefficient o f determination (r2) was used as the goodness o f fit criterion. In order to obtain optimal palibratipn, readings and samples were separated into eight groups based on soil horizon 30 (topsoil and subsoil, and spoil if frorii reclamation sites) then soil texture class where appropriate. The resulting Hnear caUbrations were used to convert NMM count ratios measured during 1996 and 1997 to 0. Field soil water contents measured by NMM were compared based on soil horizons, textures, depths, and sites. The characteristics o f seasonal soil water status that were compared include total soil profile water content, and soil water depletion at each site. Soil water status over time was determined by comparing soil water changes between each NMM reading (incremental soil water depletion). Comparison of beginning and ending soil water content for each NMM measurement interval (May 15 to September 12, 1996 and March 30 to May 11 ,1997) provided cumulative soil water depletion for each site. The purpose o f the above comparisons was to determine if relationships exist between seasonal soil water content and soil physical properties or site characteristics, with particular emphasis on differences between native and reclaimed sites. Soil Water Retention and Plant Available Water Soil cores collected in June and September, 1996 for caHbration o f the neutron moisture meter were also used for laboratory measurement o f soil water retention. Water retention at one-third bar and 15 bar pressure was measured using a pressure plate apparatus. Disturbed samples were repacked to initial soil bulk density, as described by KJute (1986). Each soil sample was individually sieved through a 2 mm screen to remove coarse fragments and repacked to the initial bulk density (previously measured from the field core samples) by pressing a specific mass of soil into rigid 7.3 cm (2.9 in) inside diameter PVC segments (rings) 31 to attain a standard volume o f 23 cm3, with a sample height o f I cm. By packing a variable mass o f soil into a consistent volume, the previously measured bulk densities were replicated and all samples had uniform dimensions, To obtain a representative soil sample, soil was redistributed and sampled using the cone method prior to weighing and packing. Repacked soil within the rings was saturated with 3 pM calcium sulfate (CaSO4) solution on the pressure plate for at least 24 hours. Paired samples for each of four depths at each site were prepared for 15 bar soil w ater measurements (total o f 72 samples). Four samples o f each depth for each site were prepared for one-third bar measurements (total o f 144 samples). A greater number o f samples were analyzed for the one-third bar pressure because retention results at this low pressure are sensitive to differences in soil pore size distribution. It was recognized that any original soil structure was destroyed by sampling, crushing and sieving dried samples. Thus, samples at one-third bar are expected to be different (wetter) than for the field condition. Fifteen bar samples were allowed to equilibrate at the applied pressure for I days. One-third bar samples equilibrated for 2 to 3 days. These relatively short equilibration times were possible because the sample height was only I cm (Klute, 1986). Plant-available soil water holding capacity was calculated for each soil horizon at each site based on soil water retention data, by taking the difference between measured volumetric soil water contents at one-third bar and 15 bar pressures. The value obtained was corrected for percent coarse fragments and converted to equivalent depth o f water based on the soil horizon depth (Marshall et al., 1996). Percent coarse fragments and horizon depths were obtained from soil profile descriptions for each site. 32 Computer Simulation Modeling o f Long Term Seasonal Soil Water Status Computer simulation modeling was used to extrapolate beyond the measured seasonal soil water data using the measured soil physical and hydrologic properties in combination with 34 years o f climate data from Colstrip, MT. The intent of the model exercise was to: I) estimate seasonal soil water within the soil profile relative to the anticipated location and demands o f ponderosa pine root systems relative to grass root systems; and 2) determine whether a more suitable soil substrate might be reasonably created that would promote a more favorable time- and depth-dependent soil water status for pines. Finally, if the outcome of Objective two were positive, to recommend one or more reconstructed soil profile designs. The computer simulation model used was an upgraded version o f Ekalaka Rangeland Hydrology and Yield Model (ERHYM), known as ERHYM-II (Wight, 1987). TheERHYM- II model was selected because it is a fairly simplified “tipping bucket” soil water model, based on the water balance equation and thus requires minimal input data. Conceptually, with a “tipping bucket” model, soil water is redistributed from one soil layer to the next when the water content o f a particular soil layer exceeds its plant available water holding capacity. In this manner, any soil water in excess o f field capacity (one-third bar pressure) is redistributed, or “tipped”, to the adjacent and lower soil layer(s). Or “bucket(s)” . Some salient characteristics o f a “tipping bucket” model include: I) it does not allow upward flow o f soil water; and 2) water additions, depletions and redistributions are instantaneous (daily time step) rather than gradual, as would actually occur in the field. The ERHYM-II model was developed specifically for application to rangeland 33 environments. It incorporates a rangeland crop coefficient curve that was developed using lysimeter data from a mixed prairie range site in southeastern Montana (Wight, 1987). Weltz and Blackburn’s (1993) analysis o f the ERHYM-II model on south Texas rangelands determined that the model has the potential to simulate the annual water balance o f semiarid rangeland plant communities where runoff and deep drainage are limited components o f the water balance. ERHYM-II is a climatically-driven water balance model that functions on the plant community level (Weltz and Blackburn, 1993). Components o f the water balance are associated with changes in precipitation, evaporation, transpiration, runoff and soil water routing, occurring on a daily time-step basis. A primary advantage o f a “tipping bucket” model is limited input information requirements, or more specifically, that required input data is limited in scope and generally readily available. This was critical to application to the Rosebud Mine, where only basic climatic and soil hydrologic data are available, ruling out use o f more complex, mechanistic models. Disadvantages o f a “tipping bucket” model are associated with its assumptions and simplifications, and therefore the caution required in applying resulting interpretations to the natural system. The ERHYM-II model code was modified for input of actual climate data instead of using the model’s ability to generate stochastic climate data. Thirty-four years (January 1964 through May 1997) o f climate data from the Colstrip weather station were acquired from Climatedata Summary o f the Day, Western, a computerized database (National Climatic Data 34 Center, 1997) and via electronic mail from the Western Regional Climate Center in Reno, NY. Climate data from the Colstrip station included daily maximum and minimum air temperatures and daily precipitation totals. Pan evaporation data were obtained for the Yellowtail Dam weather station, located approximately 75 miles southwest o f Colstrip. This is one o f the closest weather stations collecting this information and was considered most ecologically similar to the Colstrip station. Daily maximum and minimum temperature records were also collected from the Yellowtail Dam weather station and compared to daily temperature records from the Colstrip station to confirm the applicability o f Yellowtail Dam evaporation data to the Colstrip area. Climate data were reformatted for use as input files to the ERHYM-II model. Where daily temperature data were missing from the climate record, a value was provided by averaging the temperature o f the preceding and following days. I f greater than 2 days in a row were missing (for example no records were available for the entire year of 1975), daily temperatures from a year determined to have average temperature closest to the year with missing data was substituted. Data were substituted in whole or part depending on the number of days missing. For air temperature, values from 1970 were substituted for missing information in 1973 and 1974; 1977 data were substituted for missing data for years 1975, 1980 and 1981; and 1995 data were substituted for missing data during 1997. A similar method was used for substitution o f missing precipitation data, however, only for years or periods missing several days in a row. For precipitation data, 1970 was 35 substituted for 1973 and 1974; and 1995 was substituted for 1997, in whole or in part. The 34-year average was substituted entirely for 1975 (which had no climatic record). The root water uptake algorithms in ERHYM-II were modified from the original version in an attempt to more closely simulate what is known about water uptake by plant root systems. Plant available water is defined in ERHYM-II as soil water held between 1/3 (field capacity) and 15 bar (permanent wilting point) tensions. The model was altered such that soil water was considered to be freely available to plants above 60 percent (0.6) plant available water, but water availability decreased linearly below this point until zero at the permanent wilting point. This concept follows that o f Doorenbos and Kassam (1979) and is commonly used in computer simulation models. A second modification was to constrain the proportional amount o f daily transpirational demand met from a given soil layer based on the proportional root system density in that layer. A natural growth function, y = l-exp(-bx), was incorporated into the model. In this equation, “y” is proportional daily transpirational water that a given layer could provide, and “x” is the proportional root density in that layer (ROOTF variable in ERHYMrII). A value o f 5 for “b” was selected as providing reasonable constraints to soil water uptake. This modification prevented an unreasonably large amount o f water being removed from a soil layer having very few roots present. The constraint on water uptake is small until relative root density decreases to about 0.4 or less, and increases in severity as ROOTF approaches zero. Simulation models based on more deterministic principles generally calculate depth- dependent root water uptake as proportional to its availability (i.e., soil water potential) and 36 the root density at that location. Hence, we surmised this to be an appropriate modification for our purposes. A final modification to ERHYM-II was to allow multiple iterations within the soil water uptake module. The algorithm cycled through the four soil depths, in order from top to bottom, until either daily transpirational demand had been met or no plant available water remained in soil layers haying roots present. Calculations in this algorithm were constrained by the modifications noted above. Three native sites and three reclamation sites were selected for computer simulation modeling. Sites were selected to encompass the range of measured profile water holding capacity and plant available water holding capacity. The range of dominant soil texture was represented. Site-specific information regarding soil properties and soil water content were organized into the input format for the model. Site specific information collected or measured during field and laboratory investigations, and used in the ERHYM-H model for each soil horizon, included horizon thickness, bulk density, percent rock fragments, soil water content at field capacity, and soil water content at wilting point. Because the model required mass (kg/kg) soil water content, volumetric measurements were converted to mass equivalents using measured soil bulk density (Marshall et al., 1996). Horizon thickness for each of four soil layers was input as measured during profile description. The only exception was that the top layer could not exceed 30.5 cm in the model and was therefore divided and entered as two equal soil layers if greater than this limit. 37 Site specific input data that were derived or estimated included air-dry soil water content and a runoff curve number (RCN)- The amount o f water below permanent wilting point, but above air-dry water content which can be evaporated from the upper soil layer, was estimated from Table 4 o f the ERHYM-II model description and user guide (Wight, 1987). This is based on measured water content at -15 bar and the soil.texture o f the top 30 cm (12 in) o f each site. Site specific runoff curve numbers were derived using NRCS methodology (Chapter 2 o f the Engineering Field Manual), which takes into account the soil hydrologic group. Cultural practice, and vegetation cover and condition. The soil hydrologic group for each site was determined based on field descriptions and criteria in the National Soil Survey Interpretations Handbook (Soil Survey Staff, 1992). Due to its structure, the model could not simulate two different root systems at one time. In order to predict seasonal soil water distribution with pine and grass roots on the same site the model was run twice for each site with the same soil and climatic input values, changing only the relative root density within each soil horizon (ROOTF value). This allowed comparison but did not simulate competition. Another important consideration is that the model was unable to reasonably simulate potential temporal differences in pine and grass root ‘activity’. Temporal niche separation may be a critical aspect o f below ground interactions o f these plants. Relative root density with depth for two year old pine seedlings was estimated from studies in the literature that were conducted at the Rosebud Mine (Richardson, 1981; Thamarus, 1987a and b) and North Dakota (van Haverbeke, 1963). Relative root depth density for grasses was estimated from root abundance recorded during profile description 38 work using the specific native and reclamation sites selected for modeling. The ROOTF input values for each soil layer at a site were dependent on the characteristics o f each root system (pine or grass) and the horizon thicknesses specific to the native or reconstructed soils being modeled. Several input and output modifications were made to the model. Input parameters CROPCO, TRANCO, STRGRO and ENDGRO were disengaged. An evaporation pan coefficient was applied to the climate input data, which made a crop coefficient (CROPCO) unnecessary. Evaporation pan coefficient for the Yellowtail Dam weather station was estimated to be 0.55, based on photographs o f the immediate surroundings o f the weather Station (Jensen, 1973). The transpiration coefficient (TRANCO) was modified as an input variable in order to test its affect on model results. Similarly, decimal fractions to estimate effective Overwinter precipitation (that portion o f total ovenyinter precipitation that remained in the soil by April I) was set Up as a model variable to test model sensitivity to its estimated ( value. 1 The model was altered to accept climate input files including daily precipitation, potential evaporation, and total overwinter precipitation in addition to daily maximum and minimum temperatures. Overwinter precipitation was added to climate input files because simulations were only conducted during the growing seasons (April I through September 30) for 34 years. The model is not designed or suitable for use during winter months in Montana. Four new lines were added at the end o f each site input file to provide measured and/or predicted water retention data for matric potentials o f 0, -1/3, -I , -2, -5 and -15 bar [Q ^0 _1/3 -i,-2,-s and-is bar)] f°r each o f the four soil layers. Measured water retention data were O ^ 173 39 -is bar), using the pressure plate apparatus. W ater retention data at B0jf0j _1( .2 amj _5 baf) were predicted using nonlinear optimisation (Wraith and Or, 1998) o f the parametric model developed by van Genuchten (1980), The purpose of adding this information to the input files was to more clearly define model-predicted seasonal soil water status between field capacity and wilting point, by providing soil wetness classes to summarize model output. Monthly summaries o f the total number o f days within each soil water retention class, as predicted by the model over the 34-year simulation period, were added as output. Several assumptions were made regarding model input information. Climate data recorded at the Colstrip weather station was assumed to be the same for all field sites. The value used for FURCAP (surface water storage capacity) was input as zero because it was assumed to be taken into account in the runoff curve number. SIA (soil initial abstraction coefficient for runoff curve number) was input for all sites as 0.2 in (0.5 cm ), as suggested by the ERHYM-H user manual (Wight, 1987) and based on model sensitivity analyses. Input values which relate to the seasonal relative plant growth curve were selected based on obtaining a shape with dates o f “green-up” and senescence that corresponded with qualitative observations from staff at the Rosebud Mine (personal communication: Greg Millhollin, Western Energy Company) and those during 1996 and 1997 field work. Model sensitivity analyses were performed for the following model parameters: runoff curve number (RCN), soil initial abstraction (SIA), effective overwinter precipitation, transpiration coefficient (TRANCO), and relative root depth distribution (ROOTF). Model runs for the sensitivity analysis were conducted using selected site input files and changing only one input value at a time for each o f these parameters, then comparing the results to 40 determine how Sensitive the model was to a particular input. This was used to help determine the most reasonable values to input in some cases. Each native and reclamation site Was also simulated separately for only 1996 and 1997, corresponding to field NMM measurement periods. Running these years separately allowed the model simulation to begin with measured 0, rather than the predicted 0 resulting from the previous 32 or 33 year simulations. The purpose o f simulating 1996 and 1997 separately was to evaluate agreement o f model-predicted and field measured soil water contents. Because NMM measurement depths were different from modeled soil horizons, the NMM measured 0 for each soil layer was conformed to the modeled soil layer thicknesses by c|epth-weighted average to allow Comparison o f measured and predicted 0 for each site. Recommendations for Soil Profile Design The ERHYM-II computer simulation model used to extrapolate beyond the field measured data was determined to be unsuitable for evaluating the potential seasonal soil water status o f various “designed” soil profiles. Therefore, recommendations for creation o f soil profiles at the Rosebud Mine are based on information obtained from literature and on observations from the 1996 and 1997 field studies. 41 CHAPTERS RESULTS AND DISCUSSION Field Measurement Sites The most distinct soil physical characteristic differences between native and reclamation sites were soil texture and percent hard coarse fragments o f the lower horizons (70 to 90 cm depth). O f the sites investigated, native soils were, in general, finer textured with greater percent silt than reconstructed soils (refer to “Dominant Soil Texture o f Profile” in Table I). Reconstructed soils overall tended to be more coarse textured than the native soils sampled. In addition, reconstructed soils had greater percent hard coarse fragments in the. lower horizons than native soils (refer to “Percent Coarse Fragments” in Table I). Another difference observed between native and reconstructed soils was the abrupt textural changes frequently encountered between soil horizons (topsoil, subsoil and spoil) in reclamation sites (see site description forms in Appendix A). Although not quantified, reclamation sites had substantially greater grass biomass production than native sites (Figure 2). The most common dominant grass species in native sites is bluebunch wheatgrass (Elymus spicatus (Pursh) Gould), followed by green needlegrass (Stipa viridula Trin.) and little bluestem (Andropogon scoparius Michx.), as 42 Native site 493D-A(w) (top photo) Reclamation site 3915-C (bottom photo) Figure 2. Example o f differences in grass productivity on native and reclamation sites. May 11, 1997. 43 indicated on the site description forms in Appendix A and summarized in Table 3. Prairie sandreed (Calamovilfa longifolia (Hook.) Scribn.) is the most common dominant grass species in reclamation sites, followed by bluebunch wheatgrass and thickspike wheatgra?s (Elyrtms lanceolatus (Schribn. & Smith) Gould). Bluebunch wheatgrass is a cool-season, native, perennial range grass which is valuable for forage. Prairie sandreed is a warm-season, native perennial o f little value for forage, but is strongly rhizomatous making it useful for erosion control. Table 3 summarizes some characteristics of the other dominant grass species, as provided by Lovell (1992). Neutron Moisture Meter Calibration and Bulk Density Measurements Neutron moisture meter (NMM) calibration results are presented in Table 4, which also outlines the site numbers, soil horizon(s) and dominant soil texture(s) that were grouped together as having similar calibration relationships. R2 values obtained from NMM calibration for the sites sampled range from 0.61 to 0.94, with a mean value o f 0,85±0.04. The highest possible r2 value is 1.0, indicating 1:1 correlation of measured and predicted values. R2 values from calibration o f reconstructed subsoil horizon readings were all below the mean r2. All other site/horizon/soil texture groups were above mean. The low r2 values which were obtained for reconstructed subsoil likely result from the variety o f soil textures present in these horizons within and between reclamation sites. The measured data collected were considered insufficient to do reliable location-specific calibrations for each site. Table 3. Summary o f dominant grass species o f native and reclamation sites. Dominant Grass Species Native Sites Reclamation Sites Common Name Latin Name Type m e ­ ets) 121E- c # 183E- 183E- C ( S ) C(n) 493D- A(e) 493D- A(w) 4888- A 4822- 2856- 4881- 4901- B B C C 3915- C bluebunch wheatgrass E lym u s sp ic a tu s C,N,P X X X X X X X prairie sandreed C a la m o v ilfa lo n g ifo lia W,N,P X X X X X green needlegrass S tip a v ir id u la C,N,P X X X X little bluestem A n d ro p o g o n sc o p a r iu s W,N,P X X X thickspike wheatgrass E ly m u s la n ceo la tu s C,N,P X X X crested wheatgrass A g ro p y ro n c r is ta tu m QkP X X sideoat grama B o u te lo u a cu r tip en d u la W,N,P X X cheatgrass B ro m u s tec to ru m C,I,A X X Type: A=annual, C=cool season, I=Introductd, N=native, P=perennial, W=warm season (Lovell, 1992). Dominant grass species based cm ranking from Site Description Forms located in Appendix A. This table includes only those grass species ranked as dominant on more than one site.______________ 45 Table 4. Site groupings for neutron moisture meter calibration. Site Type Site Number(s) Horizon Dominant Soil Texture(s) Linear Regression I2 Value Native 121E-C(s), 183E-C(s), 493D-A(e) Topsoil CL, SiL, L 0.89 Native 121E-C(s), 493D-A(e) Subsoil SiCL, SiL 0.89 Native 183E-C(s) Subsoil L 0.94 Reclamation 4888-A, 4822-B, 2856-B. 4881-C, 4901-C, 3915-C Topsoil fSCL, fSL, SiCL, L 0.89 Reclamation 4822-B, 4881-C, 4901-C,3915-C Subsoil L/SCL, SL, CL, L 0.82 Reclamation 2856-B Subsoil LfS, fSCL 0.61 Reclamation 4888-A Subsoil SiCL 0.84 Reclamation 4888-A, 4822-B, 2856-B, 4881-C, 4901-C, 3915-C Spoil SL, SiCL, grSCL, SL 0.94 Statistics: range 0.61 to 0.94; mean 0.85; std. error mean 0.04. The mean measured soil profile bulk density for each site is presented in Table 5. Mean profile soil bulk density o f native sites is lower than reclamation sites, with native sites having a range o f 1.15 g cm"3 to 1.37 g cm"3 and reclamation sites ranging from 1.45 g cm"3 to 1.61 g cm"3. Thus, mean profile soil bulk density for all native soil horizons was less than 1.4 g cm"3, whereas mean profile soil bulk density for all reconstructed soil horizons was greater than 1.4 g cm"3 Increased soil bulk density is expected to constrain root growth and reduce shoot growth in pines (Helms, 1983; Potter and Green, 1964, van Haverbeke, 1963). 46 Table 5. Mean profile (0 to 90 cm) soil bulk density for each study site. Native Sites Mean Soil Bulk Density (g/cm3) 121E-C(s) 1.15 183E-C(s) 1.16 493D-A(e) 1.37 Reclamation Sites 4888-A 1.49 4822-B 1.45 2856-B 1.45 4881-C 1.55 4901-C 1.61 3915-C 1.61 Statistics: Native Sites mean 1.23, std. dev. 0.10; Reclamation Sites mean 1.53, std. dev. 0.07. Figure 3 illustrates mean measured soil bulk density by depth for native and reclamation sites. Mean soil bulk density o f native sites 121E-C(s) and 183E-C(s) were similar to one another at all depths and consistently lower than that measured for site 493D-A(e). Mean bulk density for reclamation sites in Area C (4881-C, 4901-C and 3915-C) is higher than the other reclamation sites sampled in Areas A and B (4888-A, 4822-B and 2856-B) at all depths except 0 to 30 cm, where 4822-B has higher bulk density than 4881-C. For most reclamation sites, soil bulk density increases with depth. However, soil bulk density decreases consistently with depth for 4822-B. Sites 4901-C and 3915-C increase in bulk density to 70 cm depth, then decrease. The higher soil bulk densities measured for reconstructed soils are consistent with expectations because o f soil disturbance, reconstruction and use o f heavy equipment that is 47 Figure 3. Mean measured soil bulk density by depth. Native Sites I 1'4 S 1.2 t 1.0 C0) Q 0.8 m 0.6 I 0.4 0.0 0-30 30-50 50-70 70-90 Soil Depth (cm) Q 121 E-C(S) ■ 183E-C(s) D 493D-A(e) Reclamation Sites i 1.0 C