WPCO 2 ZBxHP LaserJet 4/4MHPPCL5EXR PQXPDefault Para6GNDefault Paragraph Font( * ȯ7 EPPendnote text6GNendnote text =( 8 * ȯ7 EZPheading 2,he6GNheading 2,heading 2,28 * ȯ7 E#XxP7XP#?? US#ixP7P#2 @>@>endnote refe6GNendnote reference 8 * ȯ7 ESSfootnote tex6GNfootnote text=( 8 * ȯ7 EZ P footnote ref6GNfootnote reference 8 * ȯ7 ES S toc 1 6GNtoc 1To9 =( 8 * ȯ7 E. $4` hp x (#2  n toc 2 6GNtoc 2To9 =( 8 * ȯ7 E.$4 ` hp x (#toc 3 6GNtoc 3To9 =( 8 * ȯ7 E .$4 ` hp x (#toc 4 6GNtoc 4To9 =( 8 * ȯ7 E .$4 <` hp x (#toc 5 6GNtoc 5To9 =( 8 * ȯ7 E .$4<` hp x (#2  Jtoc 6 6GNtoc 6To9 =( 8 * ȯ7 E .$4` hp x (#toc 7 6GNtoc 7To9 =( 8 * ȯ7 E ZPtoc 8 6GNtoc 8To9 =( 8 * ȯ7 E.$4` hp x (#toc 9 6GNtoc 9To9 =( 8 * ȯ7 E.$4` hp x (#2&'& z#$&index 1 6GNindex 1To9 =( 8 * ȯ7 E.$ 4 ` hp x (#index 2 6GNindex 2To9 =( 8 * ȯ7 E.!$"4 ` hp x (#toa heading6GNtoa heading =( 8 * ȯ7 E.#$$` hp x (#caption 6GNcaptionTo9 =( 8 * ȯ7 EZ%P&2 ,X'~P()&*_Equation Ca6GN_Equation Caption 8 * ȯ7 EP'P(heading 36GNheading 3o9 =( 8 * ȯ7 E)*#XxP7XP#??  US#ixP7P#heading 46GNheading 4o9 =( 8 * ȯ7 Ee+[,  heading 56GNheading 5o9 =( 8 * ȯ7 El-b.??  US21>,nX- ./heading 66GNheading 6o9 =( 8 * ȯ7 Ef/\0??USStyle1 6GNStyle1To9 =( 8 * ȯ7 E12#x6X@X@#?? US#ixP7P#eis 2 6GNeis 2To9 =( 8 * ȯ7 Ei3_4?? USBody Text 6GNBody Texto9 =( 8 * ȯ7 Ef5\6??US2MO 210<1ЫXR PQXP `CG TimesXXxP7XP2xA`Arial (TT)XixP7P,xA`Arial (TT)XxP7XP2xA`Arial (TT)XixP7P,xA`Arial (TT)x6X@X@<6X9`("Courier New (TT)XixP7P,xA`Arial (TT) m PQ P `CG Times ` PQP `CG TimesXR P QXP `CG TimesX` P QP `CG TimesXR P QXP `CG TimesX` P QP `CG TimesXR P QXP `CG TimesX` PQP `CG TimesXR PQXP `CG TimesX` PQP `CG Times m PQ P `CG Times D PQP `CG TimesXR PQXP `CG TimesXD PQP `CG TimesXR PQXP `CG TimesX m PQ P `CG Times ` PQP `CG TimesXR PQXP `CG TimesX m PQ P `CG Times ` PQP `CG TimesXR PQXP `CG TimesX m PQ P `CG Times ` PQP `CG TimesXR PQXP `CG TimesX` PQP `CG Times m P Q P `CG Times ` P!QP `CG TimesXR P"QXP `CG TimesX m P#Q P `CG Times ` P$QP `CG TimesXR P%QXP `CG TimesX m P&Q P `CG Times ` P'QP `CG TimesXR P(QXP `CG TimesX` P)QP `CG TimesXR P*QXP `CG TimesX` P+QP `CG TimesXR P,QXP `CG TimesX2OO3|R h yxxdddy *# m PQ P# Ã Evaluation of Effects of Acidic Deposition to Terrestrial Ecosystems in Class I Areas of the Southern Appalachians #` PQP# A Report to the Southern Appalachian Mountains Initiative (SAMI) Christopher Eagar #XR P QXP#USDA Forest Service#` P QP# Helga Van Miegroet #XR P QXP#Utah State University#` P QP# Samuel B. McLaughlin #XR P QXP#Oak Ridge National Laboratory#` PQP# Niki S. Nicholas #XR PQXP#Tennessee Valley Authority#` PQP# # m PQ P# March 1996yx0dddy K#D PQP#SAMI Terrestrial Ecosystems Report  -pp27  < Page  yxdddy Kn#XR PQXP#yxdddy #D PQP#     &Contents#XR PQXP# ă    X` hp x (#%'0*,.8135@8: Output (Nodvin et al, 1995)]. It is not clear at this point to what extent this removal is due to SO4 adsorption, biological immobilization by plants or microorganisms, or to microbial reduction of SO4 in parts of the soil or watershed that exhibit reducing conditions (e.g., riparian zones). The dominance of the organic S fraction in the total soil S pool would indeed suggest some role of biological processes in S retention (e.g., Harrison et al., 1989a; Mitchell et al., 1992). Fitzgerald and Autry (1992) suggested that the capacity of SA spruce soils to accumulate S through net microbial immobilization could be large, especially when considering the entire soil profile. However, they also indicated that the potential for abiotic SO4 adsorption generally exceeded organic S formation in this ecosystem type and that there may be large spatial variability in S retention both across sites (horizontal) and with soil depth (vertical). Mitchell et al. (1992) concluded that the role of biological processes in regulating SO4 flux is most important in sites with low atmospheric S inputs. These seemingly contradictory findings make a general evaluation of the SO4 retention capacity of highelevation ecosystems particularly difficult. 4.1.2 Relationship of soil to soil solution chemistry  The combination of climatic conditions (low temperature, high precipitation), vegetation (conifer litter), and parent material (low weatherability and base replacement capacity) results in podzolization as the natural soil forming process in these SA sprucefir ecosystems and causes soils in these regions to be naturally acid, exchangeable base cation reserves to be low, and Al solution levels to be inherently high (especially in the upper soils). These conditions, in turn, have a profound effect on the sensitivity of soils and soil solutions to atmospheric deposition inputs and predispose the system to potential Al toxicity and Al induced inhibition of cation uptake, rather than to cation depletion per se. Indeed, given the already low base status, it is unlikely that these soils would significantly acidify further. Although further base cation depletion is possible with the input of strong acid anions (e.g., NO3 , SO4), this is unlikely to be the major pathway of soil solution change (e.g., Johnson and Fernandez, 1992). More profound effects are expected in Al chemistry, particularly in terms of amounts of potentially toxic monomeric Al3+ forms. The mechanisms of Al mobilization are described in detail in Reuss and Johnson (1986) and Johnson and Fernandez (1992). In acid soils with low base saturation, any increase in mobile anion concentration (SO4 or NO3) in the soil solution will preferentially displace Al3+ ions (over divalent and monovalent cations such as Ca2+, Mg2+ or K+) from exchange sites because of their abundance on the exchange complex and their charge. Joslin et al. (1987) reported Al levels of 1535 mol/L in the organicrich O and A horizons increasing to 3550 mol/L at greater soil depth under birchspruce forests in the Raven Fork Watershed of the Great Smoky Mountains National Park. Aluminum levels generally in the range of 50 mol/L and occasionally above 100 mol/L were measured in the upper A horizons during the IFS study in the Smokies and at Mt. Mitchell, NC. Average Al concentrations did not change considerably with soil depth but peak values declined significantly (Smithson, 1990; Johnson et al., 1991; Van Miegroet, unpublished data). Joslin and Wolfe (1992) reported similarly high Al average and peak concentration for the O and upper mineral soils (025 cm) at spruce sites on Whitetop Mountain, VA. At highly polluted sites, there may be a major anion shift from naturally organic aciddominated to strong acid aniondominated soil solutions and a concomitant shift in Al species form organically complexed Al to monomeric Al3+ forms. This anion/Al shift can occur to various degrees in the SA soil systems and/or at different parts of the soil profile, and organic acids may continue to exert some influence on leaching and Al chemistry even in polluted areas (e.g., Joslin et al., 1987). Although total Al levels in solution may be very similar between pristine and heavily polluted areas (see examples in Fernandez, 1992; Johnson and Fernandez, 1992), significant differences in Al species composition are expected. This issue of Al speciation needs to be considered when evaluating soil solution chemistry data. High Al levels in solution by themselves are not deleterious and often occur naturally in part of the soil profile, but when a high fraction of that Al is in monomeric form the possibility of Al toxicity, inhibition of cation uptake, or leaching loss to streams is increased.  4.1.3 Role of atmospheric deposition in changing soil water chemistry  The two main mobile anions of importance in causing Al mobilization in SA soils are SO4 and NO3, the former because of typically low SO4 adsorption capacity of SA soils coupled to high atmospheric inputs, the latter because most high elevation ecosystems are N saturated (see discussions in sections 4.1.1, 4.2.1 and 4.2.3). In addition to atmospheric NO3 inputs, mobile NO3 also originates from nitrification of other atmospheric N input forms (e.g., NH4+) or from mineralization and nitrification of the inherently large organic N pools in these systems (Joslin et al., 1987; Johnson et al., 1991; Van Miegroet et al., 1992). Soil solution measurements at different sprucefir ecosystems throughout the SA have consistently documented (1) that SO4 and NO3 are the dominant anions in rooting zone solutions, (2) that Al dominates the cation fraction, and (3) that most of the Al in the upper mineral soil solution (A horizon) occurs as Al3+ (e.g., Joslin et al., 1987; Smithson, 1990; Joslin and Wolfe, 1992 & 1993; Van Miegroet et al., 1990; Johnson et al., 1991). These studies have also indicated that there is less seasonal variability in SO4 solution concentrations than in NO3 solution concentration and NO3 has higher concentration peaks. Because Al mobilization is driven by peak events rather than by annual anion loads or average anion concentrations, the degree of temporal variability in total anion concentration is of critical importance in driving Al levels above potentially toxic levels. This also explains why the role attributed to SO4 in the mobilization of Al differed among studies with the concentration of SO4 being either a poor predictor of Al concentrations (where SO4 concentrations were relatively stable, e.g., Smithson, 1990, Joslin and Wolfe, 1992) or a good predictor (where SO4 concentrations fluctuated considerably, e.g., Johnson et al., 1991). However, all studies consistently showed a strong correlation between Al levels and NO3 levels in solution and concluded that internal N dynamics, and particularly of occurrence of periodically large NO3 peaks were a major driver in Al chemistry. Comparisons between atmospheric inputs and soil leaching of SO4 or between SO4 concentrations in throughfall and those in soil solutions at different SA sites indicate that the SO4 measured in rhizosphere solutions in the high elevation sites of the SA is primarily derived from atmospheric inputs, and that greater soil SO4 fluxes are generally associated with greater throughfall inputs to the forest floor (e.g., Joslin and Wolfe, 1992; Mitchell, 1992b; Van Miegroet, unpublished data). Smithson (1990) found that sites at Mt. Michell, NC with greater cloud water impact typically had higher soil solution SO4, although differences in soil solution fluxes were less pronounced than those in total atmospheric S input. Joslin and Wolfe (1992) made a similar observation at Whitetop Mt, VA. In areas that contain pyrite bearing minerals (e.g., Anakeesta formations), SO4 may also be internally released through weathering (Rochelle et al., 1987; Cook et al., 1994), but this internal source is likely more important within the context of stream water acidification than for rooting zone chemistry. The situation is somewhat different for N. Although atmospheric input and leaching output fluxes are often similar in magnitude and would suggest that NO3 leaching losses are entirely derived from atmospheric deposition, a closer look at N fluxes in and out of the upper soil suggest a significant contribution from internal processes (N mineralization + nitrification) to the overall NO3 signal, and particularly to its temporal variability (Joslin et al, 1987 & 1989; Johnson et al., 1991, Jamison et al., 1994). Cook et al. (1994) found streamwater NO3 concentrations were typically higher during the dormant season, which provides further evidence for the biological control of N retention and release. In a shortterm field study near the IFS spruce site in the Smokies, throughfall and Ahorizon solutions were collected at nearly the same location. These paired comparisons showed a strong direct link between throughfall and lysimeter SO4 levels and a consistent increase in soil solution NO3 concentration over those in the throughfall solutions (40100 mol/L) . However, it was not possible to unequivocally establish the origin of the soil solution NO3 based on isotopic N signature (Van Miegroet, Unpub. Data; Garten and Van Miegroet, 1995).   4.1.4 Uncertainties and implications to policy decisions  The observations discussed in section 4.1.3 may have major implications for policy and especially for the effectiveness of regulatory decisions on the health of Class I ecosystems because of the sources of strong anion input and the role in Al mobilization differs substantially between SO4 and NO3. Whereas the SO4 fluxes in sprucefir systems are largely due to atmospheric inputs, NO3 seems to be derived from both atmospheric and internal sources, is largely under biological control, and tends to show much larger seasonal variations. Consequently, high SO4 levels may be responsible for elevated Al background levels, but it is the NO3 peaks that are responsible for causing the potentially harmful Al peaks. Curtailing SO4 input alone may not eliminate the problem as long as N dynamics (especially nitrification) dominate the solution chemistry in the SA soils. At best, such regulatory measures may lower the magnitude of the Al peaks and the frequency with which they exceed set threshold values. Soil and solution chemistry data available to date suggest that more attention should be directed towards understanding internal N dynamics within the framework of local stand dynamics and as affected by anthropogenic influences.  4.2 Nutrient cycling in high elevation sprucefir ecosystems  Although few mineral cycling studies have been conducted in the sprucefir zone of the SA, the data that are available point at some interesting general trends with respect to nutrient distribution, atmospheric input fluxes, internal cycling, leaching losses below the rooting zone, and especially the role of stand dynamics and forest disturbances on these cycling patterns. Some of the details can be found in Johnson et al. (1991) and in Johnson and Lindberg (1992).  4.2.1 Atmospheric deposition regime  One of the significant outcomes of the IFS was a more accurate quantification of atmospheric deposition processes and the recognition that the highelevation sprucefir systems in the SA are subject to some of the highest atmospheric N and S inputs in the eastern United States. At the IFS spruce Tower site near Clingmans Dome in the Great Smoky Mountains total deposition of N was 27 kg haé1 yré1 and of S 36 kg haé1 yré1 (Lindberg, 1992; Lindberg and Lovett; 1992; Lovett, 1992; Lovett and Lindberg, 1993). A large fraction of these pollutants are deposited via dry deposition (around 25% of total S, little less than 50% of total N) and through cloud and fog impaction (almost 50% for S, 30% for N). In that cloud deposition is generally proportional to cloud immersion times, and the frequency of cloud immersion at the IFS site is lower than at many other mountain summits in the SA, atmospheric deposition levels at other sites could be even higher than those observed in the Smokies (Lovett and Lindberg, 1993). Smithson reported annual S deposition rates of 20 kg S haé1 yré1 for a low cloud and 49 kg S haé1 yré1 for a high cloud impact site on Mt Mitchell, NC, but no N deposition values are available for that location . Throughfall fluxes at Whitetop Mt, VA ranged from 40 to 47 kg haé1 yré1 for S and were a little less than 20 kg haé1 yré1 for N (Joslin and Wolfe, 1992). Considering that throughfall fluxes generally underestimate total N inputs at high elevation sites by as much as 15% due to canopy retention (Lovett, 1992), actual N deposition rates are probably in the range of those observed in the Smokies. As the cloud base typically occurs at or above a given elevation [approximately 1800 m elevation in the Smokies (Lindberg and Owens, 1993)], and because of the influence of cloud exposure on deposition in mountainous regions (Lovett et al., 1992), one would expect atmospheric input rates to the forest floor to increase with elevation, especially when comparing locations above versus below the cloud base elevation. While a comparative throughfall study in the Smokies found this trend to be true for S deposition, treetotree variability in throughfall fluxes within each site were too large to reveal any significant N input gradient with elevation (Lindberg and Owens, 1993). More importantly, Lindberg and Owens (1993) showed that canopy gap formation influences the deposition regime within these highelevation forests by significantly increasing water and ion (SO4 and NO3) throughfall fluxes below mature edge trees The stand structure (i.e., formation of gaps and edges) also has a significant impact on the elevational gradient in throughfall fluxes. Whereas total SO4 fluxes underneath the edge trees nearly doubled when crossing the cloud base elevation, the elevational gradient for NO3 followed the reverse pattern. Total throughfall fluxes below the edge trees were about 30% higher below the cloud base compared to the higherelevation site and this was attributed to greater wind penetration and HNO3 (dry) vapor deposition at the lower site. However, elevational patterns were found to be less consistent with time than those for SO4. A followup study at the Noland Divide Watershed (Shubzda, Unpublished Data) has further verified this elevational trend (increase in SO4 input and decline in NO3 input with increasing elevation).  4.2.2 Nutrient pools and fluxes  Only a few intensive nutrient cycling studies have been conducted in the SA (Weaver, 1972; Johnson et al., 1991; Johnson and Lindberg, 1992). From these studies and comparisons with other sprucefir ecosystems in North America (Johnson and Fernandez, 1992) a number of distinct features emerge. Organic matter is a major component in the nutrient pools and nutrient dynamics of this ecosystem type. The forest floor, and to a lesser degree the upper (organicrich) portion of the mineral soil, constitute a major reservoir of available nutrients (N, S, Ca, and Mg). Although this observation is common in many ecosystems, the situation is further accentuated for Ca and Mg in the SA by the extremely low exchangeable base cation levels in the mineral soil (see section 4.1.1). A second distinct feature of these highelevation ecosystems is the dominance of hydrologic fluxes in the biogeochemistry of most elements: deposition, throughfall, and leaching (all hydologic fluxes) generally exceed uptake and return by vegetation (biological fluxes). Low uptake rates are typical for mature forests where nutrient increment is primarily associated with wood increment. Recent heavy mortality of fir by balsam woolly adelgid infestation would tend to further decrease the role of vegetation uptake in nutrient retention. Input/output calculations indicate net loss or close to zero retention for N, Mg, and S, but net accumulation of atmospherically deposited Ca. Such low biological retention of S and N has major repercussions on anion leaching losses as already discussed in section 4.1.2. At the IFS sites in the Smoky Mountains, there was a net (albeit small) annual loss for Mg through leaching ( 2.5 kg haé1 yré1) which constituted 4% of the soilexchangeable pool and 0.01% of the total soil pool. Since there is no information on the rate of Mg weathering release in these soils, it is difficult to assess whether or not this process is a significant factor in the Mg supplying power of the soil. Net leaching of all base cations together accounted for an insignificant portion of the exchange capacity and Ca appeared to be accumulating in this system at a rate of 4 kg haé1 yré1 (Johnson, 1992). Collectively, these observations tend to suggest that at this point in time soils in the SA are not significantly acidifying through base cation stripping even at current high atmospheric deposition rates, but that Al is the major cation mobilized from the exchange complex, as predicted by theoretical considerations (see Reuss and Johnson, 1986). The increase in Ca and Mg flux below the canopy over atmospheric deposition inputs (2 kg haé1 yré1 for Ca and < 1.5 kg haé1 yré1 for Mg) supports the contention that these ecosystems are undergoing foliar leaching and that this may be an important factor in causing base cation (Ca and Mg) stress in these forests (see Section 4.3.1). In the case of Mg, net foliar leaching input to the forest floor appears to be larger than litterfall return. These rates could be even more important where hydrologic fluxes increase due to enhanced cloud interception associated with increasing elevation or at forest edges. At Whitetop Mt., VA, Joslin et al. (1988) found foliar leaching during acid cloud episodes equal to 16% of foliar Mg and as much as 36% of foliar Ca. The predominance of throughfall input to the forest floor (hydrologic input) over litterfall returns (particulate input) is unusual for Mg and rather exceptional for Ca. It illustrates the Capoor status of the vegetation (and the foliage in particular), and suggests that deficiencies may be further aggravated in the future through a feedback mechanism between the base status of the foliage and the base supplying capacity of the forest floor that is produced from the foliage.  4.2.3 Stand dynamics and N saturation Nitrogen saturation in SA sprucefir ecosystems is indicated by the following signs: (1) high NO3 levels in solution throughout the year (Johnson et al., 1991; Joslin and Wolfe, 1992), (2) NO3 leaching rates closely in balance with atmospheric N deposition rates (Johnson et al., 1991; Van Miegroet et al., 1992b), (3) N mineralization rates in excess of plant uptake requirements (Johnson et al., 1991; Van Miegroet et al., 1992 a&b), and (4) lack of tree response to N fertilization (Joslin, 1994). This condition is triggered by the combination of (1) large N pools with low C/N ratios in soil and forest floor which favor net N mineralization and nitrification, (2) high atmospheric deposition rates, and (3) low tree N uptake rates especially in mature forests which have low net primary productivity and/or having recently experienced heavy fir mortality (Johnson et al., 1991; Van Miegroet et al., 1992a). The role of stand dynamics in belowground N dynamics of the SA was first documented by Silsbee and Larson (1981; 1982; 1983) through stream survey in the Smoky Mountains. The NO3 concentrations in streamwater consistently increased with elevation, but were generally lower in streams draining watersheds that had been heavily logged prior to the establishment of the National Park compared to those that had not been logged or only lightly logged and contained more oldgrowth. A recent followup survey (Flum and Nodvin, 1995) showed: (1) NO3 levels are highest and dominate the anion load in streams draining watersheds containing sprucefir on nonpyrite bearing bedrock (2) NO3 levels in streamwater tend to decline below the sprucefir zone, and (3) sharp drops in NO3 stream concentrations at the sprucefir/hardwood ecotone more typically occur where the hardwoods have been logged. These observations formed the basis for a comparative study in the Smoky Mountains where N input via throughfall, N mineralization, and NO3 leaching were compared across sprucefir ecosystems at different developmental stages (Jamison et al., 1994). A preliminary analysis of the data shows significantly higher NO3 concentrations in lysimeter solutions under mature to overmature stands compared to those containing considerable regeneration in the understory. This difference in NO3 leaching can in part be attributed to a difference in stand vitality and tree uptake rates, as N inputs via throughfall were not significantly different among sites. Mature and overmature stands also showed the highest net nitrification rates, whereas net mineralization was not significantly different between stands (Jamison, personal communication).Because of high nitrification potential in these Nrich soils, changes in stand structure, and especially the formation of gaps through tree mortality and windthrow, are likely to impact belowground N turnover and accentuate N saturation. There are several reasons why an increase in N mineralization and nitrification would be expected. First, any factor causing a decline in tree N uptake (including mortality) will reduce competition for available N and increase microbial N turnover. Indirect evidence to that effect can be derived from IFS. The Becking site had a higher proportion of dead standing biomass, higher N mineralization rates, higher NO3 baseline levels in the soil solutions, and lower temporal fluctuations than the Tower site (Johnson et al., 1991; Van Miegroet et al., 1992a). In the classification of N saturation stages (see Stoddard, 1994), the decline in biologically induced temporal fluctuations in NO3 levels and the increase in the overall baseline NO3 concentrations in streams is used as an indicator for increasing N saturation within the watershed. A second reason for expecting an increase in N leaching with gap formation is the effect of soil temperature increases on N mineralization rates. Joslin and Wolfe (1993) observed significantly higher NO3 leaching rates at the warmer, more insolated portion of a gap compared to the shaded (cooler) area. Johnson et al., (1991) also showed the role of seasonal temperature fluctuations in causing periodicity in soil N transformations. A third possible contributor to increased N leaching is soil disturbance caused by windthrow. There is limited direct evidence on the role of soil disturbance on N dynamics, but this issue is frequently debated within the context of sampling methodologies for N availability and especially the relationship between field and laboratory assays (e.g., Binkley and Hart, 1989). A small experiment in the Smokies suggests that soil disturbance can significantly accelerate nitrification, especially in Nrich soils (Van Miegroet, 1995). Furthermore, it is not uncommon to observe surges in nitrification following lysimeter installation (e.g., Johnson et al., 1991). On the other hand, wild hog rooting in beech forest, which tends to mix the upper horizons and is in that respect similar to soil mixing associated with windthrow, did not significantly affect N mineralization rates (Bloss, 1987). Finally, the presence of edges in the stand structure tends to increase N inputs to the forests floor via throughfall (Lindberg and Owens, 1993), which in turn could lead to greater NO3 leaching rates. For all the reasons stated above, areas in the SA that have suffered heavy fir mortality due to infestations by the balsam woolly adelgid or that are experiencing a decline in tree growth (whether naturally due to advanced stand age or induced by cation deficiencies) would seem particularly prone to the negative effect of accelerated NO3 leaching and a further deterioration of the soil as a growing environment. 4.3 Forest Nutrition  4.3.1 Causes of nutrient deficiencies and uncertainties  There are several hypotheses that relate observed changes in spruce growth rates and carbon (C) allocation to soil chemical properties, and especially atmospherically induced changes in these properties. One hypotheses that is most frequently proposed is that sprucefir forests in the SA are experiencing base cation deficiencies (especially Ca and Mg), either as a result of cation depletion in the upper soils or through inhibition of cation uptake triggered by elevated Al levels in the rhizosphere. Another hypothesis poses that root activity and function may be adversely affected by Al toxicity. A third hypothesis attributes low foliar base levels to a depositioninduced acceleration in foliar leaching. Central to proving causality between atmospheric deposition and a decline in forest nutrition and health is establishing (1) that trees are experiencing toxicity (e.g., Al) or nutrient deficiencies (Ca, Mg) and (2) that this condition is the direct result of atmospheric deposition. There are still a number of uncertainties in establishing a direct causal link between atmospheric deposition and tree nutrition, particularly in view of the fact (1) that it is still unknown what portion of the strong anion pulses that mobilize Al in the upper soil of SA forests and/or leach cations out of the upper soil (e.g., Bondietti et al., 1991) are derived from internal processes (i.e., N mineralization and nitrification) versus atmospheric deposition; and (2) that historically these soils have been subject to acidification and cation depletion during soil formation. Low base status may have been a natural condition preceding the advent of atmospheric deposition, and we lack the longterm data required to prove this condition has changed significantly in recent times in the SA. However, a prepollution natural condition of low base, acidic soils provides an important predisposing condition for mobilization of Al by atmospheric deposition of strong anions (Reuss and Johnson 1986). These high elevation forests receive the highest SO4 deposition in the U.S. which provides an elevated baseline of mobil anions in soil solution. This high baseline makes the seasonal NO3 peaks that are derived from both internal processes and atmospheric deposition more effective in mobilizing Al. Tree ring analysis and wood chemistry provides circumstantial evidence for a significant change in soil solution chemistry coincident with an increase in pollution (Bondietti et al., 1989). Proposed deleterious effects of high Al levels in rhizosphere solutions are largely based on seedling studies conducted under controlled laboratory conditions that established threshold values for inhibition of Ca and Mg uptake (100 mol/L; Raynall et al., 1990) or restricting root growth (200 mol/L; Thornton et al., 1987; Joslin and Wolfe, 1988) in red spruce seedlings. Whether these threshold values are the same for mature spruce in a forest environment is not clear. Nutrient deficiencysufficiency levels in red spruce foliage have likewise been largely based on a single greenhouse study with seedlings conducted in Canada in the early 1970's (Swan, 1971). These standards may or may not apply to field conditions. As noted by McLaughlin and Kohut (1992) comparisons between the nutritional status of Norway spruce seedlings (which behave quite similarly to red spruce seedlings) and mature trees suggest that more restrictive standards may be necessary in the field (see Table 1). Elevational trends have been used as circumstantial evidence for the role of atmospheric deposition in causing a decline in forest nutrition. For example, McLaughlin et al. (1990 & 1991) reported lower tree growth, lower foliar Ca and Mg levels, and higher foliar Al levels at the highest elevation sites along three elevational gradients in the Smoky Mountains, and found these foliar trends to be strongly correlated with soil extractable levels. Wells et al. (1989) also noted higher Al levels in spruce foliage at the highest elevations in the SA. However, one must exercise caution in attributing causality. First, total atmospheric deposition shows large spatial variability related to stand structure and does not always increase with elevation (e.g., Lindberg and Owens, 1993). On the other hand, cloud exposure does increase with elevation (Lindberg and Owens, 1993) and has been linked to a decline in foliar base cation levels both in field (Joslin et al., 1988) and greenhouse studies (McLaughlin et al., 1993). Secondly, soil solution Al levels (both total and monomeric), Ca and Mg concentrations, and Ca:Al and Mg:Al ratios were not found to be significantly different between the mid and highelevation sites at Clingmans Dome (Van Miegroet., et al., 1990). In the Whitetop study, higher soil solution Al levels were not translated into higher Al levels in red spruce foliage (Joslin and Wolfe, 1992). Finally, slope and rockiness may cause significant differences in soil stability, soil depth and soil properties which may in turn affect site fertility status, irrespective of deposition regime (see Johnson and Fernandez, 1992). A contributing and possibly critical factor to the observed decline in foliar base cations at the higher elevations may be the role of acidic cloudwater in leaching foliar nutrients (Joslin et al., 1988; Thornton et al., 1994). Based on work at Whitetop Mt., Joslin and Wolfe (1992) found significantly higher levels of NO3 and Al in the soil solution and 40% less fine roots in the upper soil of mature spruce forests under high cloud exposure compared to those having less cloud exposure. Except for Mg in the mineral soils, no consistent and significant differences in Ca and Mg solution concentrations were measured. Solution Ca:Al and Mg:Al ratios were significantly lower at the high cloud site (mostly driven by the differences in Al concentrations). Despite these significant differences in Al solution concentrations, no significant differences in foliar Al levels were observed. For two out of three observation years, newly formed foliage had significantly lower Ca, Mg and Zn concentrations at the high cloud site (Thornton et al., 1994). Elevated soil solution levels of Al have been shown to inhibit the uptake of Ca and Mg by red spruce roots (Thornton et al., 1987). A combination of altered soil and solution chemistry coupled to higher foliar leaching rates seems a plausible explanation for the reported elevational trends in foliar levels and particularly for the sharp decline above the cloud base level (see McLaughlin et al., 1991). Many of the proposed mechanisms for declining base cation nutrition are based on the assumptions that considerable root activity is located in the upper, most acid soils horizons, that available base cations pools are thus mainly confined to the forest floor, and that the deeper soil has little or no influence on tree nutrition in these forests (e.g., Robarge et al., 1989; Bondietti et al., 1991). Red spruce is a typical shallow rooting species and roots are most abundant in the upper horizons (e.g., Johnson et al., 1991; Joslin and Wolfe, 1992). This characteristic is common in conifer forests and may be related to the abrupt soil chemistry changes that occur between organic matterdominated and mineral horizons (see discussions in Fernandez, 1992 and Johnson and Fernandez, 1992). However, some data indicate that roots do extend deeper into the soil profile (where base saturation may be somewhat higher) and that such roots should therefore be able to extract nutrients from that part of the profile as well, especially when trees encounter limited base supplies in the upper soil (e.g., Johnson et al., 1991; Johnson and Fernandez, 1992). Root distribution and the influence of organic material on rhizosphere chemistry and seedling growth and nutrition is also discussed in Chapter 5. The differential response of seedlings to acid treatments depending on whether organic matter was part of the growth medium is important to note (e.g., McLaughlin et al., 1993; Thornton et al., 1994), in that it underscores the potentially attenuating influence of organics on Al toxicity and/or Al induced nutrient deficiency. Another area of uncertainty is the inability to accurately describe rhizosphere chemistry and its implications for tree nutrition. Most soil solutions discussed in the literature were collected with lowtension or zerotension lysimeters, and essentially reflect gravitational water that drains freely through the soil. Soil water collected in this manner may not accurately describe solutions that roots typically encounter, especially when soils dry out and water is held at greater tension within the smaller soil pores. One study in the Great Smoky Mountains attempted to characterize water held at greater tension in the soil through centrifugation and showed these solutions to be different in composition than solutions from lowtension lysimeters. Contrary to expectations, centrifuged solutions had higher Ca and Mg concentration and Ca/Al and Mg/Al ratios than corresponding lysimeter solutions (Van Miegroet et al., 1990). Although these are the only such data available for the SA, and there may have been disturbance effects associated with the centrifugation technique (see Van Miegroet, 1995), they nevertheless demonstrate that caution should be exercised when discussing rhizosphere solutions. Finally, there is the issue of Al speciation in organic versus mineral soils and its potential effect on roots. Questions that arise are: Where exactly are roots likely to experience the most deleterious effects of high Al levels? Do elevated Al levels represent the same "toxicity" in organic and mineral soils? What is the role of organic matter in complexing Al and mitigating Al toxicity? To what extent does the presence of strong acid anions diminish the complexing capacity of organics? This issue is relevant if one accepts the assumption that fine roots are preferentially distributed in the upper (organicrich) horizons and that tree nutrition is for the most part dependent on these surficial roots. This is indeed the zone where organic matter content is high, where most of the decomposition occurs, and where Al is much more likely to occur in the form of Alorgano complexes thought to be less toxic. It is at greater depth, where the mineral soil dominates but less fine roots are present, that Al is more likely to occur in the more toxic monomeric form. Such steep gradients in soil and soil solution chemistry, which also occur under natural conditions, may in fact be a driver in the shallow rooting pattern (e.g., Fernandez, 1992). There is little information on Al speciation in the SA and even less on the relative toxicity of the different forms. Smithson (1990) studied the Al fractionation through extraction of soils from several high elevation sites in North Carolina, but did not evaluate how these fraction changed with soil depth. Studies that deal with spruce responses typically discuss Al in terms of total concentrations. Joslin and Wolfe (1992) recognized this limitation in their study of soil chemistry and root distribution across a deposition gradient but argued that (1) when soil solution concentrations of Al are high, most of the Al is in inorganic form, and (2) seedling studies have shown similar correlations between red spruce root growth and total Al as were found for monomeric Al. Such simplification may be valid when comparing changes in chemistry of a particular soil or horizon in response to different external influences or when evaluating changes over time at a given site. They may not be appropriate, however, when the objective is to evaluate the effect of strong acid anion inputs on the soil as a growing environment. As indicated by acid mist applications to red spruce seedlings in pot studies, the presence/absence of organic matter may be critical to seedling growth response (see discussion in Chapter 5). Most trees naturally occur in an acid environment, and many in an environment with high Al levels, yet seem to tolerate these conditions. From this observation it follows that high Al levels per se are not sufficient in explaining the effects of strong acid inputs on tree nutrition. The central question then becomes why Al occurs in a monomeric rather than organically complexed from, even in environments where the formation of the latter should be favored. Alternatively, it may be necessary to reexamine longheld opinions as to the relative toxicity of different Al forms (including those that are organically bound).  4.3.2 Nutrient Concentrations  Foliar chemistry is typically used as a measure of the nutritional status of a tree or forest. Robarge et al. (1989), Johnson et al. (1991), and Joslin et al (1992) have summarized concentrations of base cations, Al and other micronutrients in spruce and fir needles in the SA. Tables 1 and 2 represent a recent update of that summary information. For most conifers that occur in managed forests, there is usually an extensive database derived from fertilizer field trials against which to evaluate existing foliar concentrations for deficiencies. Unfortunately, no such information is available for red spruce and the only guidelines currently available are based on seedling studies conducted under controlled laboratory condition (Swan, 1971). Even less information is available for Fraser fir. Throughout the SA, foliar Ca levels in spruce are considerably lower than those in high elevation sites in the Northeastern U.S., while other nutrient concentrations are quite similar throughout eastern highelevation forests. In addition, several studies have found declining Ca and Mg levels and increasing Al levels with elevation and/or with an increase in cloud water deposition (Wells et al., 1989; McLaughlin et al, 1991; Van Miegroet et al., 1993; Thornton et al., 1994). Ascertaining whether low Ca and/or Mg levels are within the deficiency range is problematic in that there are no existing standards for mature trees, and foliar symptoms associated of these possible deficiencies in the SA are neither specific enough nor pronounced enough to allow a reliable diagnosis (e.g., Robarge et al. 1989). Furthermore, lack of historical data for spruce and fir precludes pre versus postpollution time trends. It must also be reiterated that nutrient deficiency can occur in absence of substantial anthropogenic pollution and that some highelevation soils may be naturally deficient in these elements (see discussion in Robarge et al., 1989). Based on the thresholds established in Swan's study (1971), most Ca levels in current foliage (range: 2.1 1.0 mg/g) would appear well within the sufficiency range, whereas at some sites Mg levels (range: 0.5 0.9 mg/g) may be approaching the moderate deficiency range for red spruce (Table 1). However if we apply stricter field standards, as suggested by McLaughlin and Kohut (1992), most foliar Mg levels and some of the Ca levels may start falling within the moderate deficiency range. Robarge et al., (1989) compared foliar levels in spruce and fir from different location within the SA to published ranges for unmanaged lowelevation forests in the Northeastern U.S. and Canada and found that some Ca, Mg, Mn, and Zn levels fell below published minima. They also concluded that the foliar Al in Fraser Fir and red spruce in the SA are among the highest reported in the literature. Based on the available standards for red spruce seedlings (Swan, 1971) or recommended levels for Fraser fir Christmas trees (Shelton as cited in Robarge et al., 1989), N in spruce and fir would appear to be moderately deficient (Table 1 and 2). This is highly unlikely given that most highelevation forests in the SA have been found to be N saturated. Robarge et al (1989) and Johnson et al. (1991) concluded that N was probably not deficient since foliar N concentrations did not typically decline in older foliage. Such decline is indicative of internal N translocation which occur when N supplies are limited. This assessment was supported by the lack of foliar N response to N fertilization at Whitetop Mt., VA (Joslin and Wolfe, 1994). Robarge et al., (1989) further concluded that phosphorus (P) and potassium (K) were well within sufficiency range. Several researchers have used concentration ratios between elements in an attempt to detect possible nutritional shifts. This is a common practice in commercial forestry and the basis for the DRIS system (Diagnosis and Recommendation Integrated System, e.g., Hockman et al., 1989). Robarge et al. (1989) compared the Ca:Mg ratio in the SA (Ca:Mg  2) against published ratios for healthy stands (Ca:Mg  4) and concluded that although both Ca and Mg may be approaching deficiency, Ca limitations appear to be more stringent. Also in that same study, Ca:Al ratios in the SA were among the lowest yet reported (average Ca:Al=11 versus published range 24114). This is consistent with the low value reported by Joslin and Wolfe (1994) for Whitetop Mt. (Ca:Al=18).   4.3.3 Response to nutrient amendments  Based on the interpretation of foliar concentrations and foliar levels, it would appear that sprucefir in the SA are currently experiencing some base cation deficiency (Ca and/or Mg). Two fertilization studies that were conducted at Clingmans Dome (Van Miegroet et al., 1993) and Whitetop Mt. (Joslin and Wolfe, 1994), respectively, lend further support to this interpretation. In the first study, red spruce saplings below and above the cloud base received Ca, Mg, and Ca + Mg amendments for two consecutive years. Only at the high site (which was characterized by significantly lower pretreatment Ca and Mg levels) was there a significant improvement in Ca nutrition with the addition of Ca (Ca and Ca+Mg treatment). This positive response did not persist in the second treatment year, suggesting that the trees were experiencing incipient Ca deficiency. Magnesium fertilization did not cause a significant improvement in Mg nutrition at this site, but appeared to interfere with Ca uptake. At the lower site (generally characterized by better Ca and Mg nutrition) no significant response was observed following either Ca or Mg fertilization. Based on this study, a threshold for incipient deficiency of 1.7 mg/g was proposed for Ca, while Mg levels in the range of 0.6 mg/g were considered sufficient. A study using mature trees at Whitetop Mt. found similar trends, as well as some distinct differences in fertilizer response (Joslin and Wolfe, 1994). Pretreatment nutrient concentrations in current foliage suggested Mg (0.45 mg/g) and Zn (13.5 g/g) were in the deficiency range and Ca (1.75 mg/g) above deficiency range according to Swan's (1971) standards. Significant foliar growth responses and a concomitant increase in foliar concentrations of Ca, Mg, Zn, and Mn followed the addition of Ca and Mg applied alone or together. Both cations appeared to exert a synergistic effect on the nutrition of other elements in mature spruce trees (including Zn). Calcium and Zn appeared most responsive to the treatments, and were considered most likely to be growthlimiting. Joslin and Wolfe (1994) proposed that the positive response to the addition of Ca was due to the indirect positive effect exerted by Ca upon root growth and elongation and by counteracting Al antagonisms. 4.4 Impacts at other Southern Appalachian nonClass I Areas  To address the question to what extent the situation described for the high elevation sprucefir is also applicable to other forest ecosystems at lower elevations in the SA, one must compare deposition regimes, soil characteristics, and ecosystems dynamics among these systems. Research as part of the IFS at Coweeta Hydrological Laboratory in southeastern North Carolina and at Oak Ridge, TN provide an opportunity to compare low elevations with higher elevations (Swank and Crossley, 1988; Johnson and Lindberg, 1992). Forests at lower elevations are much less impacted by cloud water deposition and annual depositions rates of N and S tend to be a considerably lower than those observed at the high elevation sites. While wet deposition rates are generally fairly comparable across low elevation sites, S input via dry deposition is variable among locations and is strongly influenced by concentrations of airborne S (Lindberg, 1992). The concentration of S is determined by the S sources within the pathway of major air masses that reach a particular location. At Coweeta, white pine and hardwood sites located at 720m elevation received around 9 kg haé1 yré1 of S and 57 kg haé1 yré1 of N. A comparable loblolly pine site at 300 m elevation near Oak Ridge, TN received 15 kg haé1 yré1 of S and 10 kg haé1 yré1 of N (Lindberg, 1992; Lovett, 1992). This compares to high elevation sites in the Smokies which received 36 kg haé1 yré1 of S and 27 kg haé1 yré1 of N. Soil characteristics are also significantly different from those in the highelevation sprucefir forests. Lowelevation soils are mostly Ultisols and Inceptisols associated with Ultisols. They generally have higher base saturation and lower soil C and N contents. The latter is probably related to their land use and disturbance history (e.g., agriculture, logging, fire) which tends to limit organic matter accumulation in the soil. As a consequence, these ecosystems have a large N retention capacity, and little or no NO3 leaching occurs below the rooting zone. Only with major forest disturbances that eliminate or strongly curtail tree uptake processes can N saturation be induced, and even then the phenomenon is temporary (Swank and Crossley, 1988). Such a response corresponds to stage 0 or 1 of watershed N loss (Stoddard, 1994). A comparison of the belowground N dynamics among different vegetation types along an elevational gradient in the Smoky Mountains confirmed this lower N status in general terms (indicated by amount and form of extractable N), but also showed that potential N release rates through mineralization (determined by laboratory incubations) could vary strongly among vegetation types (Garten and Van Miegroet, 1994). Higher base saturation in mid and lowelevation soils make them susceptible to soil acidification due to cation loss rather than to Al mobilization and its effect on tree health and nutrition. There is evidence of recent change in the exchangeable base capital at the Walker Branch Watershed (Oak Ridge, TN) ( Johnson and Van Hook, 1989) and at Coweeta Hydrological Laboratory (Otto, NC) (Knoepp and Swank 1994) which has been attributed both to natural processes (Ca uptake by trees) and atmospheric deposition (SO4ĩinduced leaching of Mg). Crucial to these atmospheric deposition effects is the rate of SO4 deposition and the S retention capacity of the ecosystem. As long as S inputs are below annual plant requirements, S retention is close to complete and there is little danger of SO4ĩmediated leaching. With increasing S deposition, the rate of soil adsorption tends to become critical. In general, Ultisols are considered to be strong SO4 adsorbers due to higher sesquioxide contents in these highly weathered soils and are not classified as sensitive to the deleterious effect of atmospheric deposition (see Section 4.1.1). There may be considerable variation in the adsorption capacity within this soil order (Harrison and Johnson, 1992), however, and some Ultisols at low elevation were found to incompletely retain atmospheric S inputs (Johnson et al., 1986). From this information it appears that lower elevation sites are unlikely to show the same base cation deficiency/Al mobilization stresses that typically characterize the high elevation sites. This is likely the results of a combination of lower atmospheric deposition regimes, higher base cation reserves, and larger N and S retention capacity of these systems compared to the highelevation sprucefir ecosystems. It should be noted, however, that a decline in S and/or N retention capacity might result in accelerated cation leaching and lowering of the base saturation. Such base cation stripping, if occurring long enough or rapidly enough to exceed replenishment by weathering and decomposition, can ultimately result in soil conditions similar to those currently observed at the high elevation sites. Although the current information seems to indicate that SO4 breakthrough is more likely than N breakthrough in these lower elevation systems, especially those that have low soil N capital due to past land use or disturbances, the same impact can be expected whether caused by NO3 or SO4. The streamwater observations from the Fernow Watershed Experimental Forest in West Virginia illustrate this. Since the 1970's, a progressive increase in streamwater NO3 levels has been measured accompanied by a similar increase in Ca concentrations (Helvey and Kunkle, 1986; Edwards and Helvey, 1991). Using Stoddard's (1994) criteria, the watershed has lost most of the biological control over N retention and has reached stage 2 of N saturation, which is characterized by consistently high NO3 levels in the stream throughout the year. Although this watershed has been subject to higher N deposition rates (average throughfall N input of 22 kg N haé1 yré1 in the 1980's) than are typically measured at the lower elevation in the SA, these observation nevertheless show that forests of similar composition as those in the SA can at some point lose their ability to retain N. In order to fully ascertain the causes for this condition, a more extensive analysis of the ecosystem and soil properties would be required.  4.5 Links between terrestrial and aquatic impacts  Soil processes that play a critically role in the nutritional quality of the soil solution can also impact stream water quality. Elevated concentrations of specific ions such as NO3 or Al3+ can have a direct detrimental effect on aquatic ecology. Water quality deterioration can also occur through a decline in the acid neutralizing capacity (ANC) of stream waters caused by a surge in strong acid anion concentrations (irrespective of whether they are NO3, SO4 or Cl) without an equivalent increase in base cation concentrations (see discussion in Reuss and Johnson, 1986). Thus, in ecosystems where elevated anion loads (from atmospheric deposition or through internal release) cause accelerated cation stripping from the exchange complex, significant soil acidification may result without causing a significant change in stream water acidity (ANC). Conversely, in waters draining terrestrial ecosystems characterized by low base cation supplies (e.g., high elevation sprucefir) a decline in ANC would be expected with an increase in anion load without a measurable change is soil acidity. In addition to soil properties, hydrologic flow paths determine what ions are transported from the terrestrial to the aquatic ecosystem and the timing of transport. During stormflow and snowmelt most water bypasses deeper soil horizons and is strongly influenced by the water chemistry of upper, organicrich horizons with low SO4 adsorption capacity (e.g., Joslin et al., 1987; Cook et al., 1994). Baseflow chemistry reflects the chemical transformations of the percolating water along its longer and slower path towards the stream. However, the solution leaving the rooting zone can undergo chemical changes such as denitrification and S reduction in nearstream zones or weathering in deeper soil strata. Consequently, not all the NO3 and SO4 leached from the uplands portion of the terrestrial ecosystem necessarily ends up in the streams (see discussion Van Miegroet, 1994 and Section 4.1.1). As discussed in Section 4.1.1, the National Stream Survey (NSS) did not include high elevation (acid) stream reaches and information on possible links between the condition of highelevation forests and the chemistry of streams draining these areas is scarce. Data collected in the Great Smoky Mountains (e.g., Silsbee and Larson, 1981;1982;1983; Cook et al., 1994; Flum and Nodvin, 1995; Nodvin et al., 1995) substantiate (1) that SO4 and NO3 are indeed the dominant anions in highelevation streams, (2) that the ANC in these streams is generally low and declines with each anion peak as predicted, and (3) that spatial patterns in stream water NO3 concentration are consistent with the degree of N saturation of the watershed. However, Nodvin et al. (1995) could detect little Al to accompany these strong anion loads, suggesting additional exchanges between the solid and solution phase before reaching the stream. 4.6 Summary and Conclusions  High elevation sprucefir ecosystems have the following characteristics that make them particularly sensitive to the direct and indirect impacts of atmospheric deposition: X hp x (#%'0*,.8135@8: