Soil-solution chemistry in a coniferous stand after adding wood ash and nitrogen

Transcription

1 Soil-solution hemistry in a oniferous stand after adding wood ash and nitrogen Eva Ring, Staffan Jaobson, and Hans-Örjan Nohrstedt 153 Abstrat: Wood-ash appliations have been proposed to promote the long-term sustainability of forest prodution at inreased harvest intensities. Effets of wood-ash and nitrogen (N) appliation on soil-solution hemistry were studied for 9 years following appliation in a oniferous stand in Sweden. Crushed, self-hardened wood ash was applied at 3, 6, and 9 Mg ha alone, the lowest dosage both with and without 15 kg N ha. Pelleted wood ash (3 Mg ha ) and N were also applied alone. The soil solution was sampled by sution ups at 5 m depth. The rushed, self-hardened ash readily dissolved in water, as refleted in inreased soil-solution onentrations of sodium and sulphate. Signifiant (p <.5) elevations were also found for potassium, alium, aluminum, and total organi arbon. Vanadium, hromium, manganese, nikel, opper, zin, arseni, and lead were not signifiantly affeted by the ash treatments, but admium tended to inrease in the treatments with ash alone. From the fourth year onwards, the ph of the soil solution was lowered and the aluminum onentration raised in the plots given 9 Mg rushed ash ha. Fertilization with N alone temporarily inreased onentrations of inorgani N, admium, aluminum, and zin and dereased the ph. The rushed ash generally had longer lasting effets than N fertilization. Résumé : Des appliations de endre de bois ont été proposées afin de promouvoir la durabilité à long terme de la prodution forestière ave des intensités arues de réolte. Les effets de l appliation de endre de bois et d azote (N) sur la himie de la solution de sol ont été étudiés durant les neuf années suivant l appliation dans un peuplement résineux en Suède. La endre broyée autodurissante a été appliquée seule aux taux de 3, 6 et 9 Mg ha ;leplusfaible dosage a été appliqué seul mais aussi ave 15 kg N ha. De la endre granulée (3 Mg ha ) et N ont aussi été appliqués seuls. La solution de sol a été éhantillonnée ave des oupes à tension à 5 m de profondeur. La endre broyée autodurissante s est rapidement dissoute dans l eau omme l indiquaient les onentrations arues de sodium et sulfate dans la solution de sol. Des augmentations signifiatives (p <,5) ont aussi été observées pour potassium, alium, aluminium et arbone organique total. Les onentrations de vanadium, hromium, manganese, nikel, uivre, zin, arseni et plomb n ont pas été signifiativement affetées par les appliations de endre mais admium avait tendane à augmenter ave les appliations de endre seule. À partir de la quatrième année, le ph de la solution de sol a diminué et la onentration de aluminium a augmenté dans les parelles traitées ave 9 Mg de endre broyée ha.la fertilisation ave N seul a temporairement augmenté les onentrations de N inorganique, de adium, aluminium et zin et a diminué le ph. La endre broyée a généralement eu des effets à plus long terme que la fertilisation ave N. [Traduit par la Rédation] Ring et al. 163 Introdution Biofuels from forest fellings are renewable energy soures, the use of whih is onsistent with Swedish energy poliy and national environmental quality objetives regarding lean air. The extration of biofuels from forests sometimes implies inreasing the harvesting intensity, whih may negatively affet the long-term sustainability of forest prodution and ause unwanted environmental effets. For instane, wholetree harvesting, that is, the harvesting of all aboveground parts, poses a greater threat to the long-term sustainability of Reeived 17 June 25. Aepted 12 Otober 25. Published on the NRC Researh Press Web site at on 25 January 26. E. Ring, 1 S. Jaobson, and H.-Ö. Nohrstedt. 2 Skogforsk, Forestry Researh Institute of Sweden, Uppsala Siene Park, SE Uppsala, Sweden. 1 Corresponding author ( eva.ring@skogforsk.se). 2 Present address: Swedish Researh Counil for Environment, Agriultural Sienes and Spatial Planning (FORMAS), Box 126, Stokholm, Sweden. forest prodution than onventional stem harvesting, whih leaves the logging residues on site, beause it inreases the amounts of nutrients that are removed (Weetman and Webber 1972; Mälkönen 1976). The burning of forest biofuels produes large amounts of alkaline ash, whih ontain most of the inorgani nutrients and metals, that have aumulated in the trees. Nitrogen (N), on the other hand, is volatilized in ombustion and is not present in the ash. The omposition of the ash generated depends on the types of biofuel and boiler involved, the point where the ash is olleted within the plant (fly ash or bottom ash), and whether additional fuels, suh as petrohemial oil, are used. To date, most of the ashes produed in Sweden have been dumped in landfills, a pratie that may ause negative environmental effets on a loal sale. It has been suggested that biofuel ash should be reyled instead, that is, returned to forest land, to promote the long-term sustainability of forest prodution and (or) to ounterat anthropogeni aidifiation. In the short term, however, indiations of both inreases and dereases in stem growth have been found after wood-ash appliations (Jaobson 23). In a Swedish environmental impat assessment regarding forest biofuels, Egnell Can. J. For. Res. 36: (26) doi:1.1139/x NRC Canada

2 154 Can. J. For. Res. Vol. 36, 26 Table 1. Studied treatments at the 249 Riddarhyttan experiment. Treatment Abbreviation Moisture orretion of dosage (%) Control Control 3 Mg rushed, self-hardened ash ha 3Cr 25 6 Mg rushed, self-hardened ash ha 6Cr 25 9 Mg rushed, self-hardened ash ha 9Cr 25 3 Mg pelleted ash ha 3Pel 1 15 kg N ha N 15 kg N ha + 3 Mg rushed, self-hardened ash ha (simultaneous appliation) N + Cr 25 (rushed ash) 15 kg N ha + 3 Mg rushed, self-hardened ash ha (N applied 1 month before ash) NbeforeCr 25 (rushed ash) Note: Two treatments with organi pellets were inluded in the experimental design but exluded from the present paper. The dosages of ash pellets refer to dry masses. Nitrogen (N) was applied as NH 4 NO 3 with some dolomite and boron. et al. (1998) suggested that the effets of inreased harvest intensity should be ounterated by reyling wood ash and that the demand for N an be met by onventional forest fertilization with N when neessary. The timing of fertilization is likely to be ruial when applying ombinations of N fertilizer and alkaline ash. Simultaneous appliation of alkaline ash and ammonium (NH 4 + ) may inrease the formation of ammonia (NH 3 ) beause of the high ph of the ash. Owing to the potentially adverse effets on both the environment and human health of applying untreated ash to forested sites, it must be stabilized before it is applied to forests. The aim of stabilization is to modify the solubility of the ash omponents and the size of the ash partiles, that is, to form dense ash-partile agglomerates (Steenari et al. 1999). Diverse methods for doing this have been tested, for instane pelleting, granulation, and self-hardening and rushing (e.g. Ring et al. 1999). Self-hardening refers to the ability of most ash materials to solidify upon the addition of water (Steenari and Lindqvist 1997). In Sweden, self-hardening followed by rushing has beome the most ommon method for ash stabilization in reent years. However, a laboratory leahing test on self-hardened wood ashes showed that signifiant proportions of sodium (Na) and potassium (K) were present as salts in these ashes and were rapidly released from them (Steenari et al. 1999). Steenari et al. (1999) found that the simulated short-term release of the Na and K salts was not redued by any of the stabilization methods applied. The environmental effets of reyling biofuel ashes need to be fully understood before the pratie an be introdued in Swedish silviulture on a large-sale basis. Over the past 15 years, there has been extensive researh on reyling biofuel ashes (e.g. Egnell et al. 1998; Nohrstedt 21; Saarsalmi and Mälkönen 21; Lundström et al. 23b; Aronsson and Ekelund 24). However, the great variations in ash hemistry, stabilization methods used, ash dosage, appliation dates, and limate ompliate the generalization of observed effets (f. Steenari and Lindqvist 1997; Aronsson and Ekelund 24). Nevertheless, some general effets on forest soil proesses of liming and wood-ash appliation have been reognized. Lundström et al. (23b) reviewed a large number of liming and wood-ash experiments performed in European and North Amerian forests. They onluded that liming or wood-ash appliation generally inreased the leahing of dissolved organi arbon (C), dereased the ph in the deep mineral soil solution, and inreased the onentrations of aluminum (Al), sulphate (SO 4 2 ), and nitrate (NO 3 ), probably as a result of the high ioni strength that they generated and inreased mirobial ativity. In another review, Aronsson and Ekelund (24) identified issues related to aquati eosystems and their responses to terrestrial ash appliations as important topis for future researh. The present study was initiated to study the effets on soil-solution hemistry of applying a pelleted wood ash at 3 Mg ha and a rushed, self-hardened wood ash at three different dosages (3, 6, and 9 Mg ha ), the lowest dosage both with and without N. Materials and methods Site desription The 249 Riddarhyttan experimental site is loated on a forested slope (approximately 1%) in south-entral Sweden (59 48 N, 15 3 E). The altitude is 18 m a.s.l. A podzolized soil profile has developed on the sandy silty to silty moraine. The C/N ratio of the mor layer was 31 in the ontrol plots and the ph (H 2 O) was 4.. The unorreted annual preipitation is on average 728 mm and the annual temperature 4.7 C (Alexandersson et al. 1991). During most of the study period, the annual preipitation was slightly (3%%) lower than the long-term average, while in 1996 and 2, the preipitation was nearly 2% lower and 4% greater than the long-term average, respetively (data from the Swedish Meteorologial and Hydrologial Institute). The site was overed by a 6-year-old mixed stand of Pinus sylvestris L. and Piea abies (L.) Karst. The site quality lass was 7.7 m3 ha year and the stand was thinned before the experiment was established. Experimental design Ten treatments were tested in a randomized blok design with three bloks. The results from eight of these treatments are reported here (Table 1). Study plots sized 22 m 22m were laid out with similar tree basal area and number of stems per hetare. The bloking was largely based on a prinipal omponents analysis of soil-solution hemistry sampled on two oasions before treatment, but moisture onditions and some site fators were also taken into onsideration. The randomization within bloks was not omplete, whih might have affeted the results from the statistial analyses to some extent. Some plots deviated from the general pattern. The 26 NRC Canada

3 Ring et al. 155 treatment regarded as being of the least potential interest, N applied 1 month before 3 Mg rushed, self-hardened ash ha (NbeforeCr), was therefore applied to the most deviant plots. The ash and N were applied manually between 27 and 3 September In treatment NbeforeCr, the N fertilizer was applied on 1 September. Wood-ash harateristis The rushed, self-hardened ash originated from the ombustion of fuel onsisting of (on average over the season) 75% biofuels, 23% petrohemial oil, and 2% pine oil at ASSI Kraftliner in Piteå, northern Sweden. Bark aounted for about 5% of the biofuels and logging residues and byproduts from sawmills for the remainder. The ash was piled and left to self-harden for about 1 month. It was then rushed and sieved through a 1 mm mesh. A laboratory dry-sieving analysis of the ash showed that the length of over 6% (by mass) of the ash partiles exeeded 2. mm. Wet sieving and a simple leahing test revealed that a large proportion of the ash was readily soluble in water. Aording to an X-ray diffration analysis (Siemens D-5, PSD-detetor, 4 kv, 2 ma), the rushed, self-hardened ash onsisted of Na 2 SO 4, K 3 Na(SO 4 ) 2, Na 2 SO 4 1H 2 O, CaCO 3, and traes of SiO 2, among whih the sulphates dissolve readily in water. The pelleted ash originated from ombustion at the Ortviken pulp plant. It was pelleted by IKAB AB, Iggesund, inluding 8%% pine oil as a binding agent. The pellets dissolved quite readily in water aording to the leahing test already mentioned but less readily than the rushed, self-hardened ash. The pellets onsisted mainly of SiO 2, CaCO 3, and MgO, but Ca 2 Al 2 SiO 7, NaAlSi 3 O 8, KAlSi 3 O 8, and probably KCl and Fe 2 O 3 were also present. Generally, the ontents of major elements in the two ashes applied at experiment 249 were omparable with those of other wood fuel ashes (n = 12156) as ompiled by Steenari and Lindqvist (1997). Compared with these ashes, the rushed ash was high in sodium (Na) and sulphur (S), whereas the pelleted ash was rather high in Al (Table 2). Compared with the ash pellets, the rushed ash had higher onentrations of Na, potassium (K), S, boron (B), barium (Ba), nikel (Ni), lead (Pb), and vanadium (V) and lower onentrations of Al, iron (Fe), silion (Si), C, arseni (As), and merury (Hg). The high onentrations of Ni and V in the rushed ash were probably due to the oombustion of petrohemial oil. The N fertilizer was NH 4 NO 3 (27.2% N) with some dolomite and B supplied as olemanite (Table 2). Soil-solution sampling Six sution ups were installed in eah study plot: three Teflon ups (type PRENART Super Quartz supplied by PRENART Equipment ApS, and three erami ups as bakups. Sine the Teflon ups worked adequately, the results presented refer solely to the soil solutions that they retrieved. The Teflon and erami ups in eah plot were installed alternately at fixed ardinal points in a irle with a diameter of about 6matapproximately.5 m depth in the mineral soil and at an angle of 6 7 to the soil surfae. The sampling depth orresponded to the lower B or upper C horizon. The Teflon ups were installed in a slurry of water and quartz flour that onsisted of 99.5% SiO 2,.15% Al 2 O 3,.35% Fe 2 O 3, and.5% TiO 2 aording to the manufaturer. The sampled soil solutions were olleted in 1 L glass flasks, whih were buried at.5 m depth up to 1 m away from the sution ups. To speed up equilibration with the soil, the ups were sampled and the olleted soil solutions were disarded twie. All but one of the Teflon ups were installed 2 7 weeks before the sampling ommened, orresponding to 3.5 months prior to the treatments. During sampling, a sution of 7 8 kpa was applied by manual pumping and the soil-solution samplers were left to fill for 3 5 days. The soil-solution samples were olleted in aid-washed plasti flasks in the early part of the study period and thereafter in unwashed polyethylene flasks, whih had been tested and approved by the aredited laboratory used, and frozen as soon as possible after sampling. The soil solutions retrieved from the three ups on eah plot were mixed in equal proportions to reate a omposite sample representing the plot. In total, 97% of all potential samples in the studied treatments were obtained. However, some missing or inomplete samples, that is, samples retrieved from less than three Teflon ups, were obtained from nearly all study plots: in all, 23% of the olleted samples were retrieved from only two ups per plot and 3.5% from only one up. Most of the failures were due to some ups repeatedly failing to retrieve water. Three samplings were performed before and 22 after the treatments. Eah soil-solution sample was analysed unfiltered with respet to eletrial ondutivity (κ), ph, and onentrations of K, Ca, Al, Cu, Cd, Pb, NH 4 + -N, NO 3 -N, phosphate-p (PO 4 3 -P), and SO 4 2 -S. The sampling period after appliation overed 6 years for most variables, but for ph and K, Al, Cd, Pb, NO 3 -N, and SO 4 2 -S ontents it was 9 years. The following onstituents were only determined in a subset of the samplings: Na, magnesium (Mg), V, hromium (Cr), manganese (Mn), Ni, zin (Zn), As, hloride (Cl ), and total organi C (TOC). B was determined only in the ontrol and N treatments. The hemial analyses were performed by an aredited laboratory (KM Laboratorierna AB, urrently ALontrol AB in Linköping). The eletrial ondutivity was determined aording to Swedish standard SS-EN The ph value was determined with a ph meter equipped with a glass eletrode (ombined eletrode) aording to Swedish standard SS Na, K, Mg, and Ca were measured by optial plasma emission spetrometry, ICP-AES (SS-EN ISO ). Al and Mn were measured by either ICP-AES or by mass spetrometry (MS), ICP-MS (EPA 2.8). Cu, Pb, and Zn were analysed with an atomi absorption spetrometer equipped with a graphite furnae devie. However, from 1999 onwards, the ICP-MS tehnique was used. V, Cr, and As were analysed by ICP-MS (initially also with an AAS graphite furnae). For As and Cr, the dynami reation ell tehnique was employed using ammonia as ollision gas. Total N was onverted to NO 3 -N after digestion with peroxodisulphate and determined photometrially by the autoanalyser tehnique. NO 3 -N was redued to NO 2 -N and the reagents sulfanile amide and N- (1-naphtyle) ethylene diamide gave a oloured omplex measured at 545 nm. NO 3 -N was measured by the same priniple. First, the sum of NO 3 -N and NO 2 -N was determined followed by the separate analysis and subtration of NO 2 -N. NH 4 + -N was analysed with an autoanalyser. NH 4 + -N was reated with sodium saliylate and dihloroisoyanuri aid forming a blue omplex measured at 66 nm. PO 4 3 -P was 26 NRC Canada

4 156 Can. J. For. Res. Vol. 36, 26 Table 2. Chemial ontents of the studied produts. Crushed ash (n =4) Pelleted ash (n =2) Moisture ontent (%) na LOI (%) na Total onentration (% of dry mass) Na na K na Mg b Ca b Al na Mn.8 1. na Fe na Si na Ti na P.8.8 na S na C na Total onentration (ppm of dry mass) As B b Ba na Be <1.2 <1.2 na Cd <.1 Co na Cr Cu Hg na La na Mo 7.7 <5.9 na Nb <12 <12 na Ni <.5 Pb 11 7 <1 S < na Sn <24 <24 na Sr na V <2 W <24 <24 na Y na Yb na Zn Zr na Nitrogen (Skog-CAN) a (n =1) Note: The ash analyses were performed by Svensk Grundämnesanalys AB in Luleå, Sweden, and the analysis of the N fertilizer (Skog-CAN) by KM Laboratorierna AB (urrently ALontrol) in Uppsala, Sweden. The ash analyses were onduted as follows: the samples were dried at 15 C. As, Cd, Co, Cu, Hg, Ni, Pb, and Zn were measured after dissolving dried samples in HNO 3 H 2 O (1:1) in losed Teflon vessels in a mirowave oven. The other elements were measured after melting.125 g samples with.375 g of LiBO 2 and dissolving the resulting mixture in HNO 3. Conentrations were measured by ICP-AES and ICP-MS. S was determined using a Carlo-Erba analyser and by ICP- AES after digestion in HF and HNO 3. ICP-AES probably gave the best estimates of the S ontents of the ash produts, sine S appears to have been present in the ash produts mainly as SO 4 2. n, number of samples; na, not analysed. a Conentrations in ppm of fresh sample. b Aording to the manufaturer. analysed spetrophotometrially with an autoanalyser. In a solution aidified with sulphuri aid, orthophosphate formed a blue-oloured omplex together with molybdate, trivalent antimony, and asorbi aid (measured at 88 nm). SO 4 2 -S and Cl were analysed by ioni hromatography aording to Swedish (and international) standard SS-EN ISO Finally, TOC was determined by atalyti ombustion of the sample followed by spetrometri measurement (IR) of formed CO 2. Prior to the ombustion, the sample was aidified and aerated to remove inorgani C (SS-EN 1484). The onentrations of Cu, Cd, Pb, NH 4 + -N, NO 3 -N, and PO 4 3 -P were often below the limits of detetion. For V, all onentrations were below the limit of detetion. All values below the detetion limit for any variable were set equal to it. The valene of Cu, Cd, and Pb was assumed to be 2+ and that of Al 3+. Whitish flakes (of unertain omposition and origin) were o- 26 NRC Canada

5 Ring et al. 157 asionally observed in the early samples. The results from the samples with flakes are inluded in the data presented here. Leahing The total amount of K, NO 3 -N, and SO 4 2 -S leahed annually from some of the treatments was alulated by multiplying the annual mean onentration of the elements in the soil solution at 5 m depth by the estimated annual perolation in the area. The annual perolation was estimated as the annual preipitation measured at Swedish Meteorologial and Hydrologial Institute station 9549 Riddarhyttan orreted by +1% minus the annual mean evapotranspiration (5 mm) for the area (Brandt et al. 1994). Statistial analysis The data were analysed aording to the following mixed linear model (model 1) by means of SAS Pro MIXED ver. 8.2 (SAS Institute In. 1999): [1] y ijk = µ + α i + b j + d ij + τ k +(ατ) ik + e ijk where y ijk is the onentration of the variable under onsideration in the soil solution, µ is the overall mean, α i is the fixed effet of treatment, i = 1,..., 1, b j is the random effet of blok, j =1,2,3,d ij is the random effet of interation between treatment and blok, τ k is the fixed effet of time, k = 1,..., 22 (maximum), (ατ) ik is the fixed effet of interation between treatment and time, and e ijk is random error. The data related to all 1 treatments in Table 1 were analysed using model 1. To aount for the dependene between repeated measurements, a spatial ovariane struture on residuals was used (σ 2 ρ t i j ) in whih the dependene between samplings (ρ) delines with inreasing time intervals (t ij ) (SAS Institute In. 1999). The denominator degrees of freedom for the tests of fixed effets were estimated aording to a general Satterthwaite approximation (SAS Institute In. 1999). The need for transformation was judged by onsidering the skewness and kurtosis of the residuals and by visual interpretation of their distribution. The normality of the distributions of residuals improved after logarithmi transformation for all variables exept PO 3 4 -P and B. Cu also showed a better distribution of residuals after logarithmi transformation, but then the iteration did not onverge. The analysis was therefore performed on untransformed Cu onentrations. The ph values were onverted to H + onentrations, whih gave a loser to normal distribution of the residuals. Treatment means over bloks were alulated by the leastsquare means method and inluded an adjustment for missing values. The nonlogarithmi least-square means that are presented in the figures were alulated by running model 1 on untransformed onentrations. Model 1 was modified for appliation to single samplings by exluding the fators τ k and (ατ) ik, resulting in model 2. The effets of different treatments were ompared by alulating differenes of leastsquare means adjusted for multiple omparisons aording to Tukey Kramer (SAS Institute In. 1999). Pearson s orrelation oeffiients were alulated based on all postappliation observations within single treatments (n = 1 treatment 3 bloks number of samplings). Effets were onsidered statistially signifiant if p <.5. Results The onentrations of the measured elements in the soil solution of the ontrol plots during the study period were as follows (mean onentrations in µmol L for all samplings, 1 n 22, with standard deviations in parentheses): K, 7.8 (2.6); Ca, 53 (11); Al, 51 (11); Cd,.12 (.86); Pb,.44 (.41); Cu,.39 (.36); NH 4 + -N,.62 (.81); NO 3 -N,.44 (.24); PO 4 3 -P,.46 (.12); SO 4 2 -S, 14 (56); and TOC, 2.8 (.71) mg L. The median ph was 4.6 (minimum 4.5, maximum 4.8) and mean κ was 3.8 (.57) ms m. Aording to model 1, there was a signifiant effet of either treatment (α i ) or interation between treatment and time (ατ) ik for all of the studied variables exept the Mn, Cr, Pb, and As onentrations. Signifiant effets of treatment were found for κ, K, Ca, Al, Cd, NH 4 + -N, NO 3 -N, and SO 4 2 -S, all of whih were measured on oasions after appliation, and for Na, Cl, and TOC, whih were measured on 5 oasions after appliation. An effet of treatment was indiated for H +, Mg, Cu, and Cr (.87 < p <.13). Mn, Ni, Zn, As, and PO 4 3 -P were not signifiantly affeted by treatment, but signifiant interations between treatment and time were found for Ni, Zn, and PO 4 3 -P ontents. Suh interations were also present for κ, H +, Na, K, Mg, Ca, Al, Cu, Cd, NH + 4-N, NO 3 -N, SO 4 2 -S, and TOC. Most of the studied variables showed a distint temporal pattern, manifested in a signifiant effet of time for all variables. Soon after appliation, κ, Na, K, Mg, Ca, Mn, Cd, Cl, inorgani N, SO 4 2 -S, and TOC inreased (Figs. 1 and 2). These inreases were often followed by gradual delines in the ash treatments. In the pure N treatment, on the other hand, the onentrations shortly returned to pretreatment levels. Most of the effets of the N fertilization lasted for 1 2 years. 3, 6, and 9 Mg rushed ash (3Cr, 6Cr, and 9Cr) During the first year after appliation, a pronouned inrease in soil-solution κ was observed in the 3Cr, 6Cr, and 9Cr treatments (Fig. 1). The eletrial ondutivity was signifiantly elevated for about 1 year in the 3Cr treatment, 3 years in the 6Cr treatment, and at least 6 years in the 9Cr treatment (model 2). The soil-solution ph was only signifiantly affeted in the 9Cr treatment. Starting about 4 years after the appliation, the ph in the 9Cr treatment was lower than in the ontrol and remained lower for several years (Fig. 1). The ph in the 9Cr treatment was positively orrelated with K and SO 4 2 -S (Pearson s orrelation oeffiients.26 and.74, respetively) and negatively orrelated with Al (.39). The Al onentration was elevated in the 9Cr treatment ompared with the ontrol (models 1 and 2) (Fig. 1). The K and Ca onentrations showed a large rise initially followed by a gradual deline (Fig. 1). K was elevated in all treatments with rushed ash, whereas Ca was signifiantly elevated only in the 9Cr treatment (model 1). The inrease in K persisted for about 4 years in the 3Cr treatment, 7 years in the 6Cr treatment, and at least 9 years in the 9Cr treatment (model 2). The effet on K appeared at lower ash dosages and was more long-lasting than the effet on Ca. The effets on Na and Mg seemed to follow the same temporal patterns as for K and Ca, respetively. The Cd onentration was signifiantly elevated in the 9Cr treatment ompared with the ontrol (model 1), and in the 6Cr treatment, a rise in Cd onentration was indiated 26 NRC Canada

6 158 Can. J. For. Res. Vol. 36, 26 Fig. 1. Changes in soil solution hemial variables at 5 m depth in the pure ash treatments at experiment 249 over time (least-square means over blok after appliation): ontrol ( ), 3Cr ( ), 6Cr (S), 9Cr ( ), and 3Pel ( ). Note that the units and sales differ for the different elements ph Condutivity (ms m ) Al (mmol L ).2 K (mmol L ) Cd ( µ mol L ).2.1 Ca (mmol L ) TOC (mg L ) SO -S (mmol L ) Year Year (p =.1). Aording to model 2, the elevation in the 9Cr treatment was present from the seond to the sixth year after appliation. Zn showed an initial inrease followed by a deline, similar to the patterns for Na and Mg, all of whih were measured on five oasions after the appliation. The highest Zn onentration (least-square mean) in the treat- 26 NRC Canada

7 Ring et al. 159 Fig. 2. Changes in soil solution hemial variables at 5 m depth in the N and low ash treatments at experiment 249 over time (leastsquare means over blok after appliation): ontrol ( ),N( ),N+Cr( ),NbeforeCr (+), and 3Cr ( ). Note that the units and sales differ for the different elements ph Condutivity (ms m ) Al (mmol L ) K (mmol L ) Cd ( mol L ) µ NH -N (mmol L ) Ca (mmol L ) NO -N (mmol L ) Year Year 26 NRC Canada

8 16 Can. J. For. Res. Vol. 36, 26 ments with rushed ash was 2.1 µmol L (in 3Cr). Signifiant effets were deteted on zero, two, and one oasions for Cu, Pb, and PO 3 4 -P, respetively (model 2), but a large proportion of the Cu, Pb, and PO 3 4 -P onentrations were below the detetion limits, thus adversely affeting the results of the statistial analyses. The highest onentrations (least-square mean) of these elements among all treatments were.47 µmol Cu L (in 6Cr),.26 µmol Pb L (in 3Pel (3 Mg ash pellets)), and 1.1 µmol PO 3 4 -P L (in 9Cr). Inorgani N ontents were not signifiantly affeted by the 3Cr, 6Cr, or 9Cr treatments. All measured onentrations of NH + 4 -N and NO 3 -N were below 1.2 and 21 µmol L, respetively. The SO 2 4 -S onentrations, on the other hand, were onsiderably inreased by these treatments ompared with the ontrol (Fig. 1). Although measured less frequently, Cl appeared to follow the same temporal pattern as SO 2 4 -S, although no signifiant hanges in Cl were deteted with model 2. Soil-solution TOC was inreased by the appliation of rushed ash. All the treatments with rushed ash, inluding N + Cr and NbeforeCr, resulted in elevated TOC onentrations a year after appliation (Fig. 1). In 9Cr plots, the TOC remained elevated for 2 years following the appliation. The data shown in Fig. 1 suggest that the effets of the rushed ash on several hemial variables were related to its dosage. However, models 1 and 2 revealed few suh differenes. Aording to model 1, κ and K were signifiantly elevated in 9Cr ompared with 3Cr, but no signifiant differenes in the Cd onentrations were found. Aording to model 2, κ, K, Cd, Pb, PO 3 4 -P, and SO 2 4 -S differed oasionally between two ash dosages. A prerequisite for suh omparisons was that least-square means ould be alulated for all treatments. With model 1, this ould only be done for κ, H +,K, Cu, Cd, Pb, and PO 3 4 -P. For Ca, Al, inorgani N, SO 2 4 -S, and TOC, least-square means ould not be alulated for some or all treatments, probably beause of missing data. However, some effets ould be related to the ash dosage by linear regression. When regressing the treatment means (µ + α i aording to model 1) of κ, H +, K, Cd, Pb, and PO 3 4 -P against the ash dosage, that is,, 3, 6, and 9 Mg ha, positive linear relationships (p <.5) were obtained with R 2 values ranging between.92 and.97. For Cu, no suh relationship was deteted. 3 Mg ash pellets (3Pel) The effets of the 3Pel treatment were not signifiantly different from the ontrol for any of the studied variables apart from SO 4 2 -S on one oasion and Al on the last two oasions (model 2) (Fig. 1). The data indiate that there were small inreases in some variables, for example, κ, K, Ca, Cd, and Zn. A omparison between the 3Pel and 3Cr treatments revealed few signifiant differenes. Only the effets on κ, Na, and SO 4 2 -S differed oasionally between the two ash produts, with 3Cr giving the highest values (model 2). Fig. 3. Average Al onentration from June 1997 onwards, that is, after the initial peak in the N treatment. The average values were alulated on the least-square means presented in Fig. 2. The bars show the standard deviation, whih in this ase reflets the temporal variation. Al (mmol L ).2.1 CTRL N N+Cr NbeforeCr 3Cr NandN+3Mgrushed ash Following the N treatment, the NH 4 + -N and NO 3 -N onentrations were elevated for 1 and nearly 2 years, respetively (Fig. 2). Soon after fertilization, the ph fell by approximately.6 unit in the N treatment ompared with the ontrol. This drop was aompanied by inreases in Al, Cd, and Zn. The maximum onentrations of these elements were 2. mmol Al L,.34 µmol Cd L, and 11 µmol Zn L. A signifiant inrease in B was deteted in the N treatment 6 years after fertilization. On this oasion, the B onentration was 2 µg L in the ontrol and 42 µg L in the N treatment. All of these onentrations were above the detetion limit. Despite the simultaneous appliation of rushed ash, the phinn+crplots initially dropped by.4 unit ompared with the ontrol (Fig. 2). The ph in N+Crplots also tended to fall below the ontrol values in 1999 (p =.53 and.54) (Fig. 2). On these oasions, there was no signifiant differene in the ph values between the N and the N+Cr treatments. As in the N treatment, the Al onentration inreased initially in N+Crplots, and additional inreases were reorded at the end of the study period (Figs. 2 and 3). In ontrast with the effets of the N treatment, the Cd and Zn onentrations were not signifiantly inreased in N+Cr plots. Compared with the N treatment, the NH 4 + -N onentration was lower in NbeforeCr plots, but no signifiant differene was found for NO 3 -N. As shown in Fig. 2, the N + Cr treatment seemed to produe a smaller peak in NH 4 + -N than the N treatment, whereas the peaks in NO 3 -N were similar. Leahing The alulated annual perolation in varied between 165 and 628 mm, with an average of 33 mm, whih is of the same order as the long-term average in the area (Brandt et al. 1994). Leahing of K, Ca, Cd, and SO 4 2 -S was higher in the rushed ash-treated plots than in the ontrol plots (Table 3). The Cd leahing in the N-treated plots was also higher than in the ontrol plots. However, 64% of the Cd leahing in the N-treated plots ourred in the 1st year after fertilization, whereas the Cd leahing in the treatments with rushed ash ourred more evenly during the study period. In general, the leahing inreased with inreasing ash dosage. The alulations suggest that Ca was retained in the soil to a greater extent than K (Table 3). 26 NRC Canada

9 Ring et al. 161 Table 3. Estimated amounts of potassium (K), admium (Cd), and SO 4 2 -S leahed in and of alium (Ca) leahed in Treatment K (kg ha ) a Ca (kg ha ) Cd (g ha ) SO 4 2 -S (kg ha ) Control Cr 49 (22%) 27 (2%) 2.7 (5%) 28 (69%) 6Cr 12 (28%) 46 (3%) 3.5 (3%) 57 (79%) 9Cr 22 (37%) 56 (3%) 5.4 (4%) 75 (71%) N ne ne 3.6 ne Note: The additional leahing (X) ompared with the added amount is given in parentheses; X = (leahing treatment leahing ontrol )/added amount; ne, not estimated. a The leahing of K in was 5.5, 44, 11, and 19 kg from the ontrol, 3Cr, 6Cr, and 9Cr plots, respetively. Disussion A Swedish environmental impat assessment of forest biofuels and ash reyling to forests reommended that the maximum dosage of ash should be 3 Mg ha per rotation period (Egnell et al. 1998). The study presented here overed both this and substantially higher ash dosages (up to 9 Mg ha ). However, in pratial appliations of rushed ash, there is likely to be great spatial variability in dosages. Although a dosage of 3 Mg ha might be obtained on an average over a large area, spots with dosages up to 9 Mg ha do not seem unlikely. Effets related to the ash dosage were antiipated in the present study, sine the rushed ash was readily soluble and the experimental design enompassed a wide range in the dosages. Suh effets were revealed when treatment mean onentrations of H +, K, Cd, Pb, and PO 4 3 -P and the eletrial ondutivity were regressed against the ash dosage ( 9 Mg ha ). The average onentrations of these elements (and κ) after appliation inreased linearly with inreasing ash dosage. A study on the ammonium latate extratable nutrients in the soil at experiment 249 showed that the K store in the upper mineral soil 5 years after the appliation inreased with inreasing ash dosage (Jaobson et al. 24). The stores of Mg, Ca, and P in the mor layer and (or) the upper mineral soil were signifiantly inreased in the 6Cr and 9Cr treatments ompared with the ontrol but did not signifiantly differ from eah other. The high solubility and high ontents of Na and S in the rushed, self-hardened ash were refleted in the results. The ash applied at experiment 249 inluded readily soluble salts suh as Na 2 SO 4,K 3 Na(SO 4 ) 2, and Na 2 SO 4 1H 2 O, thus generating high onentrations of Na and SO 4 2 in the soil solution. In May 1996, for instane, Na aounted for about 9% of the positive harge in the soil solution in the 3Cr, 6Cr, and 9Cr treatments. SO 4 2 aounted for an even higher perentage of the negative harge. Given the very high SO 4 2 onentrations found and their temporal pattern, it seems likely that the SO 4 2 found in the soil solution of the treated plots largely originated from salts in the ash. Four years after the addition of 4.28 Mg wood ash ha, Lundström et al. (23a) measured elevated onentrations of SO 4 2 in the soil solution at 25 3 m depth at a site in southern Sweden. They suggested that the observed inrease in SO 4 2 onentration was aused by desorption. In a laboratory leahing test of self-hardened wood ashes, high onentrations of Na, K, Ca, Cl, and SO 4 2 were deteted in the initial leahates (Steenari et al. 1999). Our results demonstrate that similar effets may our initially in the deep soil solution. The treatments with rushed ash also inreased the onentrations of K, Mg, Ca, Al, and TOC. Despite the high solubility of the rushed ash, the effets on K (and SO 4 2 )of the 9Cr treatment persisted throughout the 9 years of the study. Lundström et al. (23a) measured inreased onentrations of dissolved organi C in the upper soil 4 years after the appliation of 4.28 Mg wood ash ha at a site in southern Sweden. At experiment 249, the effet on TOC disappeared after about 2 years. Leahing from damaged bryophytes in the treatments with rushed ash might have ontributed to the inreases in TOC at experiment 249, sine severe damage on some bryophytes was reported from the treatments with rushed ash in a nearby experiment omprising the same treatments and ashes (Jaobson and Gustafsson 21). In ontrast with two other studies on wood-ash appliation (Högbom et al. 21b; Ludwig et al. 22), the ash treatments at experiment 249 did not signifiantly affet the onentration of inorgani N in the soil solution. The low bakground leahing of inorgani N, <.1 kg N ha year, and the C/N ratio in the mor layer (31) suggest that the N limitation of the forest at experiment 249 is more pronouned, whih might explain why the ash treatments did not indue an exess leahing of N. One of the major environmental onerns regarding woodash appliation is the risk of the availability of heavy metals like Cd inreasing (e.g., Zhan et al. 1996). Few studies have onsidered the leahing of Cd in water after wood-ash appliations (f. Aronsson and Ekelund 24). However, no effets of wood-ash appliation on the Cd onentration in the soil solution were deteted at a Swedish site (Ring et al. 1999) and a Finnish site (Perkiömäki et al. 23). The present study showed that wood-ash appliation at high dosages might inrease the Cd onentration in the soil solution for several years. Aording to the leahing alulations, the aumulated leahing of Cd orresponded to about 5% of the amount added in the treatments with rushed ash. Note, however, that the leahed Cd might have originated from the soil as well as from the ash. In a nearby experiment, where the same treatments and ashes were tested, elevated onentrations of Cd (and Zn) were measured in the urrent-year pine needles 2 years after the ash appliations (data not shown). Perkiömäki et al. (23), on the other hand, found no signifiant elevation in the Cd onentration in pine needles or in the soil solution during 2 and 4 years, respetively, follow- 26 NRC Canada

10 162 Can. J. For. Res. Vol. 36, 26 ing ash addition, even after the addition of Cd-enrihed loose wood ash. The pelleted ash did not have statistially signifiant effets on the measured soil-solution hemistry parameters, but inreases in some variables are indiated in Fig. 1. Sine many of the monitored variables spanned a wide range in onentrations (Figs. 1 and 2), the variation of the residuals was probably quite large, thus reduing the likelihood of deteting weak effets suh as those indiated for the 3Pel treatment. We applied model 1 to the ontrol and the 3Pel treatment alone and found that the ph was signifiantly lower and both κ and K were signifiantly greater in 3Pel than in the ontrol. In addition, tendenies (p =.5.7) for the treatment to inrease Ca, Al, Cd, SO 4 2, and TOC were found. The effets of pelleted ash were ompared with those of the same dosage of rushed, self-hardened ash. The effets on the soil and soil-solution hemial variables of the two ash produts ould seldom be statistially separated from eah other. Jaobson et al. (24) found no signifiant differenes in ph, K, Mg, Ca, P, N, and C in the mor layer down to 15 m in the mineral soil 5 years after the ash appliation. The leahing test suggested that the solubility of the rushed ash was higher than that of the pelleted ash. The higher solubility of the rushed ash was further demonstrated by the fat that three frequently ourring bryophytes were severely damaged initially in all treatments with rushed ash at a nearby experiment with an idential experimental design (both regarding treatments and ashes) as at experiment 249, whereas no damage to them was observed in the treatment with pelleted ash (Jaobson and Gustafsson 21). In experiment 249, signifiant dereases in the ph of the soil solution, aompanied by inreases in Al, were found in 9Cr during the last 5 years of the study. Dereases in the soil-solution ph and inreased Al onentrations have also been observed in other studies following wood-ash appliation or liming (Lundström et al. 23a, 23b; Aronsson and Ekelund 24). Lundström et al. (23a) hypothesized that lowered ph values measured 4 years after the appliation of 4.28 Mg wood ash ha in southern Sweden were due to an inreased flow of H + from the upper soil as a result of displaement by Ca and Mg in the enlarged exhangeable pool. We do not know how the exhangeable pool at experiment 249 was affeted by the treatments. However, the stores of Ca and Mg were signifiantly elevated in the mor layer in the 6Cr and 9Cr treatments 5 years after the appliation (Jaobson et al. 24). The ph values of the mor layer and the upper 7.5 m of the mineral soil were signifiantly inreased only in the 6Cr treatment. The N treatment was inluded mainly as a referene for the ombined treatments with N and rushed ash, but the N treatment per se was also of interest. Effets on soil-solution hemistry of this type of N fertilizer (NH 4 NO 3 with dolomite), whih is the only N fertilizer used on a broad sale in Swedish forestry today, had not been tested, to our knowledge, in detail in the field at that point. Some unexpeted effets of the N treatment were observed, of whih the inreases in Al, Zn, and Cd were the most striking. Similar results were obtained in a followup study by Högbom et al. (21a) and were attributed to a temporal derease in the ph. Högbom et al. (21a) suggested that the mobilized Al, Zn, and Cd were insolubilized further down in the soil profile, sine no effets on the onentrations of these elements in the streams draining the fertilized areas were observed. The elevated Al, Zn, and Cd onentrations in experiment 249 originated from exhange reations in the soil rather than from the N fertilizer. A simple alulation showed that the Cd and Zn ontents of the fertilizer had a negligible impat on the soilsolution onentrations of these elements. The probable mehanisms ausing the inreases in Cd, Al, and Zn were ion exhange with NH 4 + from the fertilizer and (or) ion release from the soil matrix beause of the lowered soilsolution ph. The negative orrelation (.8, p <.1) between the ph in the N-treated plots and the soil-solution onentrations of both Cd and Al suggest that the derease in ph ontributed to the reported inreases in Al, Zn, and Cd. Thus, the present study supports the onlusions drawn by Högbom et al. (21a), inluding the finding that the dolomite supplied with the NH 4 NO 3 does not ounterat the redutions in ph in the soil solution, as found in an earlier study with pure NH 4 NO 3 (Nohrstedt 1992). Based on results from liming experiments (Lundström et al. 23b), the dolomite is probably largely retained in the mor layer, whereas NH 4 + and NO 3 move muh faster through the soil profile. All effets in the N treatment, apart from the effets on B, were onfined to the first 1 2 years with high onentrations of inorgani N. Unfortunately, the effets of the N+CrandNbeforeCr treatments ould not be validly ompared. Uptake and leahing of fertilizer N may have differed between the treatments, sine the appliations were made late in the growing season and 24 mm of rain fell between the first and seond appliations of N. Simultaneous appliation of NH 4 NO 3 and ash risks the formation of NH 3. Indeed, the NH 4 + -N onentration shortly after the appliations appeared to be lower in the N + Cr treated plots than in the N-treated plots, while the NO 3 -N onentrations were similar (Fig. 2). This implies that the simultaneous appliation of NH 4 NO 3 and ash in experiment 249 resulted in ammonia volatilization. Plotwise harge balanes for the soil-solution hemistry revealed a large surplus in negative harge in about half of the plots in May 1996 during the period of rapid inrease in ioni ontents. In a few plots, a large surplus in positive harge was found. Sine the soil was podzolized, a surplus in positive harge was expeted, sine organi anions were not monitored (Cronan and Aiken 1985; Lundström 1993). When the soil solution was unaffeted, or only slightly affeted, by the treatments, for instane before appliation of ashes and N and in November 1999, a surplus in positive harge (62 and 48 µmol L ) was found at experiment 249. We believe that the unertainties in the hemial analyses at high onentrations ontributed to the large surplus harges. For instane, the maximum unertainty in the SO 4 2 analysis (+15%) was often of the same magnitude as the surplus negative harge in the plots with rushed ash in May In priniple, the soil-solution onentrations of the various elements reflet their respetive leahing rates, sine most of the ash treatments probably had a negligible impat on tree growth and thus the water balane (Jaobson 23). However, sine the forest growth at experiment 249 is N limited, the treatments with N might have dereased the perolation, sine N fertilization at N-limited sites inreases the leaf area 26 NRC Canada

11 Ring et al. 163 index and hene evapotranspiration (Berdén et al. 1997; Alavi 22). No orresponding orretion for the potential derease in perolation was made for the N treatment. Only a few observations per year were available for alulating the annual mean onentrations. Although the alulations are rough, they suggest that Ca was retained in the soil to a greater extent than K, as expeted, and that the elevated leahing of K, Ca, Cd, and SO 4 2 in the treatments with ash might ontinue for several years to ome. The soil-solution hemistry at the study site did not display any obvious signs of severe aidifiation prior to treatment, but we found the studied forest realisti for pratial biofuel extrations and hene wood-ash appliation. Our overall onlusions of the rather short-term effets of the studied wood-ash appliations are that improvements of the soil-solution hemistry were laking and, at high dosages, adverse effets were obtained. Aknowledgements The large numbers of soil-solution samples were expertly olleted by Lars Åhlén. Britt-Marie Steenari at the Chalmers University of Tehnology performed the X-ray diffration analysis of the ashes. We thank Sveaskog for hosting the experiment on their property. The study was finaned by the forest ompanies Sveaskog (formerly AssiDomän Forestry), Holmen Skog, Korsnäs, SCA Forest Produts, Stora Enso Wood Supply Sweden, Skogforsk, and the Swedish Energy Ageny. Referenes Alavi, G. 22. The impat of soil moisture on stem growth of sprue forest during a 22-year period. For. Eol. Manage. 166: Alexandersson, H., Karlström, C., and Larsson-MCann, S Temperaturen oh nederbörden i Sverige : Referensnormaler [Temperature and preipitation in Sweden : referene normals]. Rep. Meteor. 81. Swedish Meteorologial and Hydrologial Institute, Norrköping. [In Swedish.] Aronsson, K.A., and Ekelund, N.G.A. 24. Biologial effets of wood ash appliation to forest and aquati eosystems. J. Environ. Qual. 33: Berdén, M., Nilsson, S.I., and Nyman, P Ion leahing before and after lear-utting in a Norway sprue stand effets of long-term appliation of ammonium nitrate and superphosphate. Water Air Soil Pollut. 93: Brandt, M., Jutman, T., and Alexandersson, H Sveriges vattenbalans, Årsmedelvärden av nederbörd, avdunstning oh avrinning. Rep. Hydrol. 49. Swedish Meteorologial and Hydrologial Institute, Norrköping. [In Swedish.] Cronan, C.S., and Aiken, G.R Chemistry and transport of soluble humi substanes in forested watersheds of the Adirondak Park, New York. Geohim. Cosmohim. Ata, 49: Egnell, G., Nohrstedt, H.-Ö., Weslien, J., Westling, O., and Örlander, G Miljökonsekvensbeskrivning (MKB) av skogsbränsleuttag, asktillförsel oh övrig näringskompensation. Rep National Board of Forestry, Jönköping, Sweden. [In Swedish.] Högbom, L., Nohrstedt, H.-Ö., and Nordlund, S. 21a. Nitrogen fertilization effets on stream water admium onentration. J. Environ. Qual. 3: Högbom, L., Nohrstedt, H.-Ö., and Nordlund, S. 21b. Effets of wood-ash addition on soil-solution hemistry and soil N dynamis at a Piea abies (L.) Karst. site in southwest Sweden. Skogforsk Rep. 4. Skogforsk, Uppsala. Jaobson, S. 23. Addition of stablized wood ashes to Swedish oniferous stands on mineral soils effets on stem growth and needle nutrient onentrations. Silva Fenn. 37: Jaobson, S., and Gustafsson, L. 21. Effets on ground vegetation of the apppliation of wood ash to a Swedish Sots pine stand. Basi Appl. Eol. 2: Jaobson, S., Högbom, L., Ring, E., and Nohrstedt, H.-Ö. 24. Effets of wood ash dose and formulation on soil hemistry at two oniferous forest sites. Water Air Soil Pollut. 158: Ludwig, B., Rumpf, S., Mindrup, M., Meiwes, K.-J., and Khanna, P.K. 22. Effets of lime and wood ash on soil-solution hemistry, soil hemistry and nutritional status of a pine stand in northern Germany. Sand. J. For. Res. 17: Lundström, U The role of organi aids in the soil solution hemistry of a podzolized soil. J. Soil Si. 44: Lundström, U., Bain, D.C., Taylor, A.F.S., van Hees, P.A.W., Geibe, C.E., Holmström, S.J.M. et al. 23a. Effets of aidifiation and its mitigation with lime and wood ash on forest soil proesses in southern Sweden. A joint multidisiplinary study. Water Air Soil Pollut. Fous, 3: Lundström, U.S., Bain, D.C., Taylor, A.F.S., and van Hees, P.A.W. 23b. Effets of aidifiation and its mitigation with lime and wood ash on forest soil proesses: a review. Water Air Soil Pollut. Fous, 3: Mälkönen, E Effets of whole-tree harvesting on soil fertility. Silva Fenn. 1: Nohrstedt, H.-Ö Soil water hemistry as affeted by liming and N fertilization at two Swedish oniferous forest sites. Sand. J. For. Res. 7: Nohrstedt, H.-Ö. 21. Response of oniferous eosystems on mineral soils to nutrient additions: a review of Swedish experienes. Sand. J. For. Res. 16: Perkiömäki, J., Kiikkilä, O., Moilanen, M., Issakainen, J., Tervahauta, A., and Fritze, H. 23. Cadmium-ontaining wood ash in a pine forest: effets on humus miroflora and admium onentrations in mushrooms, berries, and needles. Can. J. For. Res. 33: Ring, E., Lövgren, L., Nohrstedt, H.-Ö., and Jansson, G Ash fertilization in a learut and in a Sots pine stand in entral Sweden effets on soil-water and soil hemistry oupled to laboratory leahings of six ash produts. Skogforsk Rep. 2. Skogforsk, Uppsala. Saarsalmi, A., and Mälkönen, E. 21. Forest fertilization researh in Finland: a literature review. Sand. J. For. Res. 16: SAS Institute In SAS/STAT user s guide. Version 8. SAS Institute In., Cary, N.C. Steenari, B.-M., and Lindqvist, O Stabilisation of biofuel ashes for reyling to forest soil. Biomass Bioenergy, 13: Steenari, B.-M., Karlsson, L.G., and Lindqvist, O Evaluation of the leahing harateristis of wood ash and the influene of ash agglomeration. Biomass Bioenergy, 16: Weetman, G.F., and Webber, B The influene of wood harvesting on the nutrient status of two sprue stands. Can. J. For. Res. 2: Zhan, G., Erih, M.S., and Ohno, T Release of trae elements from wood ash by nitri aid. Water Air Soil Pollut. 88: NRC Canada