Export of Dissolved Organic Carbon , Nitrogen and Phosphorus Following Clear-Cutting of Three Norway Spruce Forests Growing on Drained Peatlands in Southern Finland

The effect of clear-cutting on the concentrations of dissolved organic carbon (DOC), organic nitrogen (DON), NH4, NO3, and P in outflow water from three productive, Norway spruce dominated drained peatlands (RCC, VCC-1, VCC-2) were studied. Changes in runoff and transport loads (concentration × runoff) at two of the catchments during the frost-free period are also presented. Approximately 40% of the area was cut at RCC and VCC-2, and 72% at VCC-1. The volume removed was 250 m3 ha–1 at RCC, 259 m3 ha–1 at VCC-1, and for VCC-2, 317 m3 ha–1. The mean annual increase in outflow concentrations of DOC during the first four years after clear-cutting was 9.0 mg l–1 at RCC, 22.8 mg l–1 at VCC-1 and 8.4 mg l–1 at VCC-2. Corresponding increases in the forms of nitrogen were: 0.23, 0.51 and 0.16 mg DON l–1; 0.06, 0.31 and 0.04 mg NH4-N l–1; and 0.05, 0.12 and 0.22 mg NO3-N l–1. Clear-cutting did not significantly (p > 0.05) increase P concentrations. The increase in non-frost season runoff over the first three years after clear-cutting was 107 mm at RCC and 207 mm at VCC-1. The export loads of DOC during the nonfrost season increased by 80 kg ha–1 at RCC and by 184 kg ha–1 at VCC-1 over the first three years. Corresponding increases for the other studied solutes were: 1.78 and 3.98 kg DON ha–1; 0.39 and 1.49 kg NH4N ha–1; 0.45 and 0.48 kg NO3-N ha–1, and 0.09 and 0.06 kg P ha–1. The study demonstrated that clear-cutting may significantly increase the export of DOC and different forms of nitrogen from drained productive peatlands while only small increases in phosphorus export may occur.


Introduction
Drainage of waterlogged sites has been part of the normal forestry practice in Fennoscandia, the Baltic countries, the British Isles and in some parts of Russia since the early 20th century (Päivänen and Paavilainen 1990).However, it was not until the 1950's and 1960's that ploughing and mechanized excavating techniques replaced manual ditching and forest drainage activity significantly increased.Currently, about 15 million ha of peatlands and wetlands have been drained for forestry in the boreal and temperate zones (Paavilainen and Päivänen 1995), and the growth of peatland forests has significantly increased.In the former Soviet Union, for example, the drainage of about 5.5 million ha has increased the wood resources by 110 million m 3 (Sabo 1988).The rate of forest clear-cutting on drained peatlands will undergo a rapid increase in the near future, when a large number of these forests approach their regeneration age.Increasing clear-cuttings on drained peatlands, the environmental consequences need to be clarified, however.Deterioration of downstream water quality due to enhanced nutrient leaching is one of the major concerns raised in connection with an increased clear-cutting of drained peatland forests.
Water quality effects of clear-cutting on mineral soil forests have long been of concern (Bormann et al. 1968, Verry 1972, Wiklander 1981, Grip 1982, Martin et al.1985, McClurkin et al. 1985, Martin and Harr 1988, Tiedemann et al. 1988, Kubin 1995).The effects of clear-cutting virgin peatland stands on water quality have also received attention (e.g.Knighton andStiegler 1980, Ahtiainen 1992).However, although many stands have already reached their regeneration age, the effects of clear-cutting of drained peatlands on nutrient leaching has only been studied at two locations in central Sweden (Lundin 1999(Lundin , 2000)), and one site in southern Finland (Nieminen 2003).
Outflows of many nutrients following clearcutting may be higher from drained peatlands than from mineral soils.This is particularly true for nitrogen, because nitrogen reserves in organic soils are much greater than in mineral soils.If mineralization of soil nitrogen increases after clear-cutting, the outflow loss of nitrogen may be high from drained peatlands.Indeed, Lundin (1999Lundin ( , 2000) ) found significant increases in nitrogen outflow following forest regeneration on drained and productive, Norway spruce dominated peatlands.However, a combination of clear-cutting and ditching decreased streamwater concentrations of nitrate and total nitrogen from one area.In a study by Nieminen (2003) on low productive, Scots pine dominated peatlands, leaching of nitrate and ammonium, in contrast, increased only at sites that were ditched after clear-cutting.
Export of phosphorus following clear-cutting may also increase more from drained peatlands than from mineral soils.This is due to the low phosphate adsorption capacity of peat (Kaila 1959, Cuttle 1983, Nieminen and Jarva 1996).Peat with low Al and Fe content, such as often occur in low productive peatlands, have particularly low phosphate adsorption capacities.Nieminen (2003) reported very high increases in outflow water phosphorus concentrations after clear-cutting of low productive, Scots pine dominated peatlands, whereas Lundin (1999Lundin ( , 2000) ) found only slight changes in productive peatlands.
Clear-cutting of drained peatlands can also increase the production and leaching of easily soluble organic substances.The processes controlling the leaching of dissolved organic carbon (DOC) from drained peatlands are poorly known, but the risk for enhanced leaching due to clearcutting may be high, as was shown by Lundin (1999Lundin ( , 2000)).
This paper presents the results from a study to investigate the effects of clear-cutting on the outflow of DOC, phosphorus, and different forms of nitrogen from drained and productive, Norway spruce dominated peatlands in southern Finland.

Site Description
The study was conducted at two locations in southern Finland; at Ruotsinkylä (60°21´N, 25°03´E, 49 m a.s.l.) and Vesijako (61°23´N, 25°03´E, 125 m a.s.l.).On the basis of measure- There was a control and treatment catchment at Ruotsinkylä (R UC and R CC ) and a control and two treatment catchments at Vesijako (V UC , V CC-1 , V CC-2 ) (Fig. 1).The sizes of catchments were: R CC 6.5 ha, R UC 3.7 ha, V CC-1 4.3 ha, V CC-2 7.8 ha, and V UC 6.9 ha.Productive peatland forest stands covered the lowest parts of the catchments and mineral soil forests the surrounding uplands.The peatland and upland forests were mature and dominated by Norway spruce (Picea abies Karst.)except for the upland forest at R UC , where Scots pine (Pinus sylvestris L.) was dominant.The proportion of peatlands from the total catchment area was: R CC 29%, R UC 38%, V CC-1 58%, V CC-2 39%, and V UC 89%.The peatland areas at Vesijako had been drained for forestry purposes in 1914.At Ruotsinkylä, the control catchment (R UC ) was drained in 1932 and the clear-cut area (R CC ) in 1927.According to the classification of old drainage areas used in Finland (Heikurainen and Pakarinen 1982), the peatland forests at R CC and V CC-2 were of the Herb-rich type and of the Vaccinium myrtillys type at the other catchments.The upland forests were Vaccinium myrtillus type (R CC , V CC-1 , V CC-2 , V UC ) or Vaccinium vitisidaea type (R UC ) according to the site type classification for upland forests (Cajander 1926).
On the basis of chemical analysis, the peatland area at the Ruotsinkylä control catchment (R UC ) was less fertile than the peat at the other catchments, particularly concerning nitrogen (Table 1).The depth of the peat layer at R UC and V CC-1 was about 0.5 m and over 1 m at the other catchments.The mineral soils were podzolic and at Ruotsinkylä mainly developed in sandy-silty till and in silty till at Vesijako.
Clear-cutting was carried out in January-February 1994 at R CC and V CC-1 and in January 1996 at V CC-2 .The areas were cut using conventional stem-only harvesting in which only stems down to a diameter limit of 7 cm were removed.Approximately 40% of the area of R CC and V CC- 2 was clear-cut and 72% of V CC-1 .The volume removed was 250 m 3 ha -1 at R CC , 259 m 3 ha -1 at V CC-1 and 317 m 3 ha -1 at V CC-2 .The ground was frozen and covered by snow during cuttings, and the heavy harvesting machinery therefore caused no significant damage to the soil.In November 1997, the clear-cut areas at Vesijako were prepared for planting using a method known as ditch-mounding.It is done in such a way that the excavator digs shallow ditches (40-50 cm deep) at 12-15 m spacings and creates small mounds from the soil removed from the ditches.Norway spruce seedlings were planted on mounds in May 1999

Sampling and Analyses
Outflow water sampling was started in November 1992 at catchments R CC , R UC , V CC-1 , and V UC and in June 1994 at V CC-2 .Outflow water was sampled manually from the overflow of the Vnotched weir in the outlet ditch of the catchment (Fig. 1).The sampling interval was 3-4 days during the snowmelt period in spring and weekly (1992)(1993)(1994)(1995)(1996)(1997) or fortnightly (1998)(1999) during other seasons.
Water sampling continued until the end of 1998 at R UC and R CC , and until the end of 1999 at the other catchments.
Runoff was monitored at all catchments except V CC-2 .Runoff was recorded for one year before clear-cutting and for the first three years after cutting using a continuous water-stage recorder.Due to freezing of the weirs and ditches, runoff was monitored only during the frost-free period (May-September or May-October).
The outflow water samples were analysed at the Finnish Forest Research Institute according to procedures described by Jarva and Tervahauta (1993).The samples were first filtered through 1.0 µm glass fibre filters (Schleicher & Schull Rundfilter 589 3 ).The filtrates were then analysed for dissolved organic carbon (DOC) with a Shimadsu carbon analyzer and for NO 3 -by ion chromatography.Total dissolved N (N tot ) and NH 4 + were determined with a Tecaton FIA-analyzer and total dissolved P with plasma emission spectrophotometry (ICP-AES, ARL 3580).Dissolved organic N (DON) was calculated as the difference between N tot and inorganic nitrogen.Results concerning nitrogen concentrations in water outflow from R CC and V CC-1 during the first growing season following clear-cutting have been presented previously (Nieminen 1998).
Changes in runoff and outflow concentrations of DOC, DON, NH 4 + -N, NO 3 --N, and P resulting from clear-cutting were studied using the pairedcatchment method.On the basis of measurements made during the pre-treatment calibration period, a linear regression equation was calculated for the relationship between measured values in the area to be cut and respective values from control area.After clear-cutting, values were predicted for the area as if it had not been cut using this equation and measured values from the control area.The effect of clear-cutting on runoff and outflow concentrations is the difference between measured values after cutting in the clear-cut area and the predicted values.Regression equations used for prediction are given in Table 2. Except for NH 4 + , the correlations between clear-cut areas and control areas during calibration period were high and statistically significant (p < 0.05).
As can be seen from the regression equation for Table 2. Regression equations and correlation coefficients for mean mothly runoff and outflow concentrations of DOC, DON, NH 4 + -N, NO 3 --N, and P between the areas to be clear-cut and control areas during the calibration period.
Ruotsinkylä Vesijako runoff between V CC-1 and V UC (Table 2), a negative value is predicted for V CC-1 with V UC mean monthly runoff values of < 0.002 l s -1 ha -1 .Negative values for NH 4 + at V CC-1 were also predicted.Such negative values were always substituted with zero in subsequent analysis of the data.Monthly export loads (kg ha -1 or g ha -1 ) were simply calculated as the product of the monthly runoff and monthly mean concentration, and the changes in export loads due to clear-cutting as the difference between measured and predicted loads.The non-parametric Wilcoxon signed-rank t-test was used to test if the measured and predicted uncut outflow concentrations after cutting were significantly different.The BMDP (1990) software package was used.

Changes in Concentrations
Outflow concentrations of DOC and DON following clear-cutting increased at all three clear-cut areas although measured and predicted concentrations from V CC-2 differed significantly (p < 0.05) only during the first two years after cutting (Tables 3-5).The average annual increase in DOC concentrations during the first four years after clear-cutting was 9.0 mg l -1 at R CC , 22.8 mg l -1 at V CC-1 and 8.4 mg l -1 at V CC-2 .Corresponding increases for DON were 0.23, 0.51 and 0.16 mg l -1 .
Outflow concentrations of NH 4 + and NO 3 -also increased at all three clear-cut areas, but there were large differences between areas.The average annual increase in NH 4 + -N concentrations during the first four years after clear-cutting was 0.06 mg l -1 at R CC , 0.31 mg l -1 at V CC-1 and 0.04 mg l -1 at V CC-2 .Corresponding increases for NO 3 --N were 0.05, 0.12, and 0.22 mg l -1 .There were no significant (p > 0.05) changes in P concentrations due to clear-cutting.

Changes in Runoff and Export Loads
The total increase in non-frost season runoff during the first year following clear-cutting was 52 mm at both R CC and V CC-1 and 49 mm and 85 mm during the second year.While the increase in runoff remained high at V CC-1 during the third year, it had almost disappeared at R CC .Over the entire three-year study period, the increase in runoff was 107 mm at R CC and 207 mm at V CC-1 (Table 6).
Clear-cutting increased the export loads of DOC and DON at both R CC and V CC-1 .The increases were similar during the first year fol-Table 3. Mean annual concentrations (mg l -1 ) of DOC, DON, NH 4 + -N, NO 3 --N, and P in outflow water from R CC 1-5 years after clear-cutting.An asterisk indicates significant (p < 0.05) difference between measured (M) and predicted (P) values.± = standard deviation.lowing clear-cutting, but the increases from V CC-1 were substantially higher during the second and third years.The total increase in the export of DOC over the first three years after clear-cutting was 80 kg ha -1 at R CC and 184 kg ha -1 at V CC-1 (Table 6).The corresponding increase in the export load of DON was 1.78 kg ha -1 at R CC and 3.98 kg ha -1 at V CC-1 .The export loads of NH 4 + and NO 3 -also significantly increased as a result of clear-cutting.The increases in export loads were particularly high in autumn 1994 and 1995, and spring 1995 and 1996, when runoff was high and the concentrations from the clear-cut areas significantly increased.The total increase in export of NH 4 + -N over the first three years after cutting was 0.39 kg ha -1 at R CC and 1.49 kg ha -1 at V CC-1 .The corresponding increase for NO 3 --N was 0.45 and 0.48 kg ha -1 .The changes in loads of P due to clear-cutting were small; the total increase at R CC was 0.09 kg ha -1 and 0.06 kg ha -1 at V CC-1 .

Discussion
The effect of clear-cutting on runoff and outflow concentrations and export loads of DOC, DON, NH 4 + , NO 3 -and P from drained and productive, Norway spruce dominated peatlands were studied using a paired-catchment approach.The correlations between clear-cut areas and control areas during the calibration period were usually high and statistically significant (Table 2).However, the correlations for NH 4 + concentrations between R CC and R UC and between V CC-2 and V UC were low.The predicted uncut concentrations for NH 4 + -N may therefore be less reliable than those for the other solutes.The predicted monthly runoff values for the Vesijako clear-cut area V CC-1 may also be somewhat less accurate than those for the Ruotsinkylä clear-cut area (R CC ).This is because the non-frost season monthly runoff at the Vesijako control area (V UC ) varied more during the post-treatment period in 1994-1996 (0-90 mm) than during the calibration period in 1993 (4-46 mm).Ideally, when using the paired-catchment approach, the calibration period and the posttreatment period should be hydrologically similar (Seuna 1988).Nevertheless, the results for both Vesijako and Ruotsinkylä concerning changes in runoff due to clear-cutting are in good agreement with previous studies from wetland forests (Seuna 1988, Lundin 1999, 2000).
The export loads and outflow concentrations of DOC and DON from this study significantly increased as a result of clear-cutting.Lundin (1999Lundin ( , 2000) ) also found high increases in their concentrations in water outflow after clear-cutting of drained and productive, Norway spruce dominated peatlands, and increased leaching also occurred after shelterwood-cutting.Thus, forest regeneration on drained productive peatlands appears to be an important source of DOC and DON to watercourses.However, ditching of peat soils has been shown to significantly decrease the leaching of soluble organic substances (Joensuu et al. 2001), and the overall effect of peatland forestry on DOC and DON may therefore be small.
In the study by Lundin (1999) on Norway spruce dominated peatlands such as in this study, the outflow concentrations of nitrate decreased from the area that was ditched after clear-cutting.This contrasts with the results of this study, where the highest nitrate concentrations at the two ditched (ditch-mounded) clear-cut areas (V CC-1 and V CC-2 ) occurred after ditching.In  -October (1994-October ( , 1995) ) or May-September (1996) Nieminen Export of Dissolved Organic Carbon, Nitrogen and Phosphorus Following Clear-Cutting … Scots pine dominated peatlands of low fertility, nitrate leaching only increased from areas that were ditched after clear-cutting (Nieminen 2003).Decreased leaching of nitrate due to clear-cutting and ditching is thus unlikely to be a general pattern.Indeed, although nitrate leaching decreased from one ditched clear-cut area in the previously mentioned study by Lundin (1999), there was also an area where nitrate leaching increased after clear-cutting and ditching.Ditching of pristine mires for forestry and maintenance of ditch networks in old drainage areas (ditch cleaning and/or complementary ditching) are generally significant sources of NH 4 + to water courses (Hynninen andSepponen 1983, Joensuu et al. 2002).The present study and the results reported by Lundin (1999) indicate that clear-cutting of drained productive peatlands stands may also enhance NH 4 + leaching.In drained unproductive, Scots pine dominated peatlands, minor NH 4 + leaching may occur unless ditching operations are performed in connection with forest regeneration (Nieminen 2003).Significantly higher NH 4 + and NO 3 -concentrations in percolating soil water following clear-cutting were observed under piles of cutting residues than respective residue-free areas (Rosén and Lundmark-Thelin 1987).Whole-tree harvesting might thus be an effective means of decreasing NH 4 + and NO 3 -leaching.
In the study by Lundin (2000), phosphorus concentrations in streamwater increased by 3-12 µg l -1 after clear-cutting.In this study, the outflow concentrations of phosphorus from the clearcut areas also increased slightly.The changes in concentrations were not significant (p > 0.05) however.The risk for enhanced leaching of phosphorus following clear-cutting is likely to be significantly higher from nutrient poor mires than from productive Norway spruce mires because of low phosphate adsorption capacities (Cuttle 1983, Nieminen andJarva 1996).Thus, Knighton and Stiegler (1980) reported higher phosphorus concentrations in outflow water following clearcutting of a nutrient poor, ombrotrophic peatland than a minerotrophic peatland area.Very high phosphorus concentrations (> 400 µg l -1 ) in outflow water following clear-cutting of drained, low productive Scots pine dominated peatlands were also shown by Nieminen (2003).
The study demonstrated that clear-cutting may significantly increase the export of DOC and different forms of nitrogen from Norway spruce forests growing on drained productive peatlands, but only small increases in phosphorus export may occur.The results from this investigation are generally similar to those reported from Sweden (Lundin 1999(Lundin , 2000)).Future research should be directed towards the development of methods to reduce leaching from drained peatlands.Uncut riparian buffer zones between clear-cut areas and water courses might be an effective means of decreasing leaching.The effect of alternative cutting methods, such as whole-tree harvesting, on nutrient leaching should also be investigated.

Table 1 .
Chemical characteristics of peat (0-20 cm peat layer) at each catchment area.
CC , where a dense cover of birch (Betula pendula Roth and B. pubescens Ehrh.) seedlings developed soon after clear-cutting.

Table 6 .
Measured (M) and predicted (P) non-frost season a) runoff and export load of DOC, DON, NH 4 + -N, NO 3 --N, and P from R CC and V CC-1 during the first three years after clearcutting.