Harvest residue management and fertilisation effects on soil carbon and nitrogen in a 15-year-old Pinus radiata plantation forest

Forest Ecology and Management 262 (2011) 339–347

Contents lists available at ScienceDirect

Forest Ecology and Management journal homepage: www.elsevier.com/locate/foreco

Harvest residue management and fertilisation effects on soil carbon and nitrogen in a 15-year-old Pinus radiata plantation forest Haydon S. Jones ⇑, Peter N. Beets, Mark O. Kimberley, Loretta G. Garrett Scion, Private Bag 3020, Rotorua 3046, New Zealand

a r t i c l e

i n f o

Article history: Received 3 June 2010 Received in revised form 25 March 2011 Accepted 26 March 2011 Available online 19 May 2011 Keywords: Soil carbon Soil nitrogen Harvest residue management Forest floor Fertilisation

a b s t r a c t Growing interest in the use of planted forests for bioenergy production could lead to an increase in the quantities of harvest residues extracted. We analysed the change in C and N stocks in the forest floor (LFH horizon) and C and N concentrations in the mineral soil (to a depth of 0.3 m) between pre-harvest and mid-rotation (stand age 15 years) measurements at a trial site situated in a Pinus radiata plantation forest in the central North Island, New Zealand. The impacts of three harvest residue management treatments: residue plus forest floor removal (FF), residue removal (whole-tree harvesting; WT), and residue retention (stem-only harvesting; SO) were investigated with and without the mean annual application of 190 kg N ha1 year1 of urea-N fertiliser (plus minor additions of P, B and Mg). Stocks of C and N in the forest floor were significantly decreased under FF and WT treatments whereas C stocks and mass of the forest floor were significantly increased under the SO treatment over the 15-year period. Averaged across all harvesting treatments, fertilisation prevented the significant declines in mass and C and N stocks of the forest floor which occurred in unfertilised plots. The C:N ratio of the top 0.1 m of mineral soil was significantly increased under the FF treatment corresponding to a significant reduction in N concentration over the period. However, averaged across all harvesting treatments, fertilisation prevented the significant increase in C:N ratio of the top 0.1 m of mineral soil and significantly decreased the C:N ratio of the 0–0.3 m depth range. Results indicate that residue extraction for bioenergy production is likely to reduce C and N stocks in the forest floor through to mid-rotation and possibly beyond unless fertiliser is applied. Forest floors should be retained to avoid adverse impacts on topsoil fertility (i.e., increased C:N ratio). Based on the rate of recovery of the forest floor under the FF treatment, stocks of C and N in the forest floor were projected to reach pre-harvest levels at stand age 18–20. While adverse effects of residue extraction may be mitigated by the application of urea-N fertiliser, it should be noted that, in this experiment, fertiliser was applied at a high rate. Assessment of the sustainability of harvest residue extraction over multiple rotations will require long-term monitoring. Ó 2011 Elsevier B.V. All rights reserved.

1. Introduction The importance of forest soils in the global cycling of carbon (C) and as a potential reservoir for the accumulation and storage of C is well recognised (Peng et al., 2008), as is the vital role of organic matter in sustaining forest productivity and environmental quality (McLaughlin and Phillips, 2006). The ability of forest soils to act as sinks for atmospheric C, sustain commercial biomass production, and provide other environmental services will likely depend on how well these soils are managed (Fox, 2000; Lal, 2005; Vanguelova et al., 2010). Forest management practices affect soil C and nitrogen (N) storage by changing the quantity or quality of organic matter inputs to the soil, causing physical disturbance of the soil profile, or by modifying the soil environment (temperature and ⇑ Corresponding author. Tel.: +64 7 343 5597; fax: +64 7 348 0952. E-mail address: [email protected] (H.S. Jones). 0378-1127/$ - see front matter Ó 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.foreco.2011.03.040

moisture regimes) and nutrient levels (Johnson, 1992; Johnson and Curtis, 2001; Jandl et al., 2007; Nave et al., 2009, 2010; Vanguelova et al., 2010). In-line with the increasing demand for renewable energy, interest in the use of planted forests for bioenergy production is growing in New Zealand (Hall et al., 2009) and elsewhere (Vanguelova et al., 2010). This demand could lead to the intensification of forest management with associated reductions in the quantities of residues returned to the site following harvesting and the need to apply fertilisers to maintain site productivity. Recent studies have suggested that harvest residue removal or forest floor disturbance could have implications for the long-term storage of C or N (or both) in plantation forest soils in New Zealand (Jones et al., 2008; Smaill et al., 2008a) and elsewhere (e.g., Laiho et al., 2003; Chen and Xu, 2005; Powers et al., 2005; Tan et al., 2005; Thiffault et al., 2006). In a recent review and meta-analysis of 75 published studies from around the world, Nave et al. (2010) found that forest harvesting (encompassing a range of

340

H.S. Jones et al. / Forest Ecology and Management 262 (2011) 339–347

intensities and residue management practices) resulted in a significant 8 ± 3% decrease in total soil (combined forest floor and mineral soil) C on average in temperate forest soils. Forest floors were found to be more susceptible to C loss than mineral soils with a reduction in forest floor storage of 30 ± 6% over all forest types whereas C storage in mineral soils was not significantly affected overall. Other work in New Zealand has indicated that the fertilisation of plantation forest soils may increase levels of C or N (or both) in forest floors (Smith et al., 1994a; Smaill et al., 2008b) and surface mineral soils (Watt et al., 2008; Huang et al., 2011). Similar findings have also been reported elsewhere (e.g., McFarlane et al., 2009; Rifai et al., 2010), although other studies have found little net effect of fertilisation on soil C and N (e.g., Van Miegroet and Jandl, 2007; Kim, 2008). In a review and meta-analysis of the impacts of elevated N inputs (including N fertilisation) on the storage of C in forest soils based on 72 experimental sites, Nave et al. (2009) found that N inputs increased total soil (combined forest floor and mineral soil) C by 7.7% and decreased the C:N ratio by 4.9%. Stocks of C were increased predominantly in the mineral soil (by 12.2%) whereas the C:N ratio was decreased predominantly in the forest floor (by 7.8%). Internationally, relatively few studies have reported on the combined effects of harvest residue management and fertilisation on C or N (or both) in plantation forest soils (e.g., Smith et al., 1994a,b, 2000; Vanguelova et al., 2010). Insufficient attention has been given to the change in soil C and N over time in response to forest harvest residue management and fertilisation in New Zealand. Although some studies have examined treatment difference at mid-rotation (e.g., Jones et al., 2008; Smaill et al., 2008a,b), changes relative to pre-treatment (preharvest) levels at mid-rotation have not been well established. Based on treatment effects at mid-rotation, some studies suggest that the effects of harvest residue manipulations and fertilisation on soil C and N may persist for up to 25–30 years (Smaill et al., 2008a,b). Schipper et al. (2011) have reported complex temporal changes (inter-annual trends) in soil C and N storage under longterm pastoral grazing of hill-country which highlights the likely difficulties in establishing long-term trends with infrequent remeasurement and the need for more long-term, and more frequent, monitoring. The primary objective of this study was to determine the change in forest floor and mineral soil C and N in response to harvest residue management and fertilisation between pre-harvest and mid-rotation (stand age 15 years) measurements in a Pinus radiata D. Don plantation forest. P. radiata is the dominant plantation forest species grown in New Zealand, comprising about 90% of the planted forest area. Treatment effects on tree growth (basal area) and the contribution of the mineral soil coarse fraction to total mineral soil C and N stocks were also examined. 2. Materials and methods 2.1. Site location and description The study was undertaken at the Kinleith Forest trial site which was part of the Long-Term Site Productivity (LTSP) Series I group of trial sites (Smaill et al., 2008b). The Kinleith Forest trial was located in a second rotation P. radiata (D. Don) plantation forest situated in the central North Island, New Zealand (latitude 38° 140 S and longitude 175° 580 E). The first rotation stand (P. radiata) was established in 1966 at a stocking of 1667 stems ha1. The stand was production thinned to 500 stems ha1 in 1978 and to a final stocking of 300 stems ha1 in 1985. The first rotation stand at the trial site was clear-fell harvested in 1991 at stand age 25 years. Harvesting of the stand was undertaken using chainsaws, and harvesting machinery was kept off the plots to avoid additional disturbance (Smith et al., 2000).

The trial site was initially planted in 1991 but, due to a high mortality rate, was completely replanted in September 1992. Seedlings were approximately 12 months old when planted and planting was undertaken without mechanical cultivation. Smith et al. (2000) gave a full description of the trial site, only the key points of which are summarised here. The trial was established on an elevated (490 m) and slightly hummocky surface within a deeply incised landscape (Smith et al., 2000). The annual rainfall at the site is 1764 mm year1 and the mean annual temperature is 11.5 °C (Leathwick et al., 2003). The trial was terminated shortly after the mid-rotation sampling (at stand age 15 years) in September 2007 due to the conversion of the forest to dairy-farm pastures. Smith et al. (2000) described the soil type at the site as a Taupo sandy loam to silt loam. These soils are classified as Typic Udivitrands (Soil Survey Staff, 2003), corresponding to Immature Orthic Pumice Soils of the New Zealand Soil Classification system (Hewitt, 1998). 2.2. Trial design The trial was a split-plot, randomized complete block design with four replicate blocks. Each block contained three main plots to which three harvest residue management treatments were randomly assigned. Main plots had dimensions of 80  40 m and each contained two split-plots, one of which was randomly assigned a fertiliser treatment and the other left unfertilised after replanting. Each split-plot consisted of a 20  20 m measurement area surrounded by a 10 m buffer. The measurement area of each split-plot initially contained about 100 trees (2500 stems ha1) and these were thinned-to-waste – meaning stems were retained on-site – at stand-age five (to 1250 stems ha1) and again at age 10 (to 625 stems ha1). The trees were not pruned. The main harvest residue management treatments were stemonly harvesting (SO), whole-tree harvesting (WT), and whole-tree harvesting plus forest floor removal (FF) (Smith et al., 2000). The FF treatment involved the careful removal of the forest floor (undertaken manually using rakes) in addition to the removal of harvest residues whereas the WT treatment involved the removal of harvest residues while minimising disruption of the forest floor. The WT treatment is effectively what would occur if harvest residues were removed for bioenergy production. With the SO treatment (the conventional approach to forest harvesting), branches and foliage were retained on-site after the removal of merchantable logs. It is estimated that approximately 50 Mg ha1 of residues were retained in SO plots based on the pre-harvest branch and foliage biomass estimates for the first rotation (described in Section 2.6). All treatments received weed control which involved periodic herbicide application through to crown closure. Urea-N fertiliser was applied annually (at a mean rate of 190 kg ha1 year1) to one split-plot within each main plot from stand age 1–6 years (totalling 950 kg ha1 of N). A high rate of Urea-N was applied with a view to ensure an ample supply of N for maximum tree growth. The high rate and timing of application (during early stand establishment) were not common practice in New Zealand forest operations. Total applications of other nutrients applied to each fertilised split-plot from stand age 1–6 years were 100 kg ha1 of P (LL Super/reactive phosphate rock), 12 kg ha1 of B (Ulxite), and 200 kg ha1 of Mg (calmag/MgSO4/dolomite). The other split-plot remained unfertilised (Smith et al., 2000). 2.3. Pre-treatment sample collection The forest floor and mineral soils at the trial site were sampled in 1991 prior to the harvesting of the previous stand. This sampling is referred to here as the pre-treatment sampling (PT). Forest floor material (combined LFH horizon) was sampled at five randomly selected points across the entire 80  40 m main plot area because

H.S. Jones et al. / Forest Ecology and Management 262 (2011) 339–347

the main plots had not been split at that time. Samples were collected from within a 0.25 m2 metal quadrat. Mineral soils were sampled at 10 points within each plot using a small-diameter tube sampler. Samples were collected from five depth intervals (0–0.1, 0.1–0.2, 0.2–0.3, 0.3–0.5, and 0.5–1 m) at each point but only data from the top three 0.1 m depth increments are presented here. 2.4. Mid-rotation (post-treatment) field data and sample collection Forest floor materials were sampled at 12 points situated on a regular grid pattern within each 20  20 m measurement plot (henceforth referred to as ‘plot’). Two forest floor horizons were identified and sampled separately: litter (L) and fermented/humic (FH) horizons. For the purposes of this study, the L horizon included needle, bark, cone, and branch (<100 mm diameter) materials. The L horizon was sampled from within a 0.25 m2 metal quadrat centred on the sample point. The FH material was collected from within a sampling ring (of 98 mm internal diameter) inserted in the centre of the same 0.25 m2 area. The FH samples were bulked by plot in the field. All L and FH materials contained within their respective sampling areas were collected. Slope measurements were made at each of the 12 sample points within each plot. The mineral soil was sampled at 25 locations on a regular grid pattern within each 20  20 m plot. At each location, two points, spaced approximately 1 m apart, were sampled separately, giving 50 sample points per plot in total. After carefully removing the forest floor, the mineral soil was collected from three 0.1 m depth ranges (0–0.1, 0.1–0.2, and 0.2–0.3 m) at each point using a small-diameter tube sampler. The samples from each plot were bulked by depth range in the field. Three bulk density samples were also collected at three randomly selected points within each unfertilised plot. Fertilised plots were not sampled for bulk density. The bulk density samples were collected using a 98 mm internal diameter (100 mm depth) stainless-steel soil sampling ring and the data were used to express mineral soil C and N stocks on a mass per unit area basis (Mg ha1). Mineral soil and forest floor sampling was undertaken in late August and early September 2007 when stand age was 15 years (i.e., at mid-rotation). 2.5. Sample preparation and analysis The forest floor (L and FH materials) samples were oven dried at 65–70 °C to constant weight, weighed, and ground to <1 mm prior to analysis for total C and N concentration. All forest floor samples were bulked by plot. Mineral soil samples were air dried at 640 °C to constant weight before being passed through a 2 mm sieve to separate the fine (<2 mm) and coarse (>2 mm) fractions. The coarse fraction of mid-rotation samples, which mainly consisted of fine pumice gravel, was finely ground and analysed separately to determine its contribution to total (fine plus coarse fraction) stocks. Each fine and coarse fraction sample was sub-sampled for C and N analysis. Mineral soil bulk density samples (unfertilised plots at mid-rotation only) were also air dried at 640 °C before being passed through a 2 mm sieve to separate the coarse and fine fractions. The samples were oven dried at 105 °C (to constant weight). Mineral soil and forest floor samples from the pre-treatment sampling and mid-rotation sampling were analysed for total C and N using a LECO FPS-2000 CNS thermal combustion furnace (LECO Corp., St Joseph, MI) at a similar point in time. Archived pre-treatment samples were reanalysed together with the midrotation samples to avoid the possibility of analytical techniques influencing the results. Where an insufficient quantity of archived sample remained for re-analysis (four mineral soil samples were affected), ratios of loss-on-ignition (LOI) to total C and total Kjeldahl nitrogen (TKN) to total N based on the data obtained from

341

the other samples were calculated for the affected depth ranges and applied to the LOI and TKN data from the samples that could not be re-measured to estimate the missing total C and N values. 2.6. Tree growth characteristics Diameter at breast height (1.4 m, DBH), diameter at base of green crown, and tree height were measured within the 80  40 m main plots prior to harvesting. Breast height and green crown basal areas were calculated on a plot basis. Individual component biomass was determined by the destructive sampling of nine trees in a stand adjacent to the trial site (due to the timing of harvesting at the trial site). A biomass measurement procedure similar to that described by Dyck et al. (1991) was used. The biomass data was used to estimate trial site biomass using a green crown basal area to basal area conversion approach. The mean total above-ground biomass in the first rotation stand was estimated to be 223.27 Mg ha1 with stems, branches, and foliage accounting for 173.73, 44.16, and 5.38 Mg ha1, respectively. Growth characteristics of the second rotation stand were remeasured at mid-rotation (stand age 15 years) as part of the normal measurement programme for the trial series. Tree heights and DBH (1.4 m) were recorded and used for the calculation of mean plot height, mean top height, basal area, and under-bark volume. A full description of the re-measurement is given by Oliver et al. (2011). They present growth characteristics data for harvest residue treatments in unfertilised plots only. The effects of harvest residue management and fertilisation on mean basal area, calculated using both fertilised and unfertilised plots, are examined here. 2.7. Data and statistical analyses Stocks (Mg ha1) of C and N in the forest floor horizons were calculated as the product of concentration and total oven-dry mass divided by total collection area on a plot basis, and slope corrections then applied. Stocks in L and FH horizons were calculated separately and summed to give stocks for the LFH horizon at mid-rotation (LFH material was collected together in the pre-treatment sampling). Mid-rotation mineral soil C and N stocks in unfertilised plots were calculated for both the fine and coarse fractions. For each mineral soil depth range (i.e., 0–0.1, 0.1–0.2, and 0.2–0.3 m), C and N stocks were calculated by multiplying the concentration data with the corresponding mean density and sample thickness on a plot basis. The stocks of each mineral soil depth range were summed to give fine and coarse fraction stocks over the 0–0.3 m depth range. Slope corrections were also applied to mineral soil stocks. Concentrations of C and N for the 0–0.3 m mineral soil depth range were calculated on a weighted-average basis. The oven-dry mass per unit area (Mg ha1) of the forest floor horizon was also calculated. Changes in forest floor and mineral soil variables over the 15 year time period were assessed by calculating the difference between the pre-treatment and mid-rotation values. All statistical analyses were carried out using SAS/STAT Version 9.1 (SAS Institute, Cary, NC). Analyses of variance (ANOVAs) were performed using the MIXED procedure to test for the effects of harvest residue management treatments and the effect of fertiliser. The interaction between these two experimental factors was also tested but found not to be significant for any variable analysed. The dependent variables for these analyses were the change in LFH C and N stocks, LFH oven-dry mass, mineral soil C and N concentrations, and mineral soil C:N ratio between the pre-harvest measurement and the mid-rotation measurement in the replanted stand. Based on these ANOVAs, two approaches were used to examine treatment effects on the changes in forest floor and

342

H.S. Jones et al. / Forest Ecology and Management 262 (2011) 339–347

mineral soil variables. For the first approach, the statistical significance of the change relative to zero was obtained for each treatment. For the second approach, pair-wise comparisons of treatment means were performed using the Tukey–Kramer test with significance level a = 0.05. All mineral soil variables were analysed over the entire 0–0.3 m depth and for the 0–0.1, 0.1–0.2, 0.2– 0.3 m depth ranges. Live-tree basal area at stand age 15 years was also analysed using stocking (stems per hectare) as a covariate. 3. Results 3.1. Effects of residue management on the forest floor pool The C and N stocks of the forest floor pool (LFH horizon) under the FF and WT treatments were significantly decreased over the 15-year period with reductions in C stocks of 3.07 and 3.73 Mg ha1 and reductions in N stocks of 0.10 and 0.15 Mg ha1, respectively (Table 1). Oven-dry mass of the forest floor under the FF and WT treatments was also decreased but in both cases the change was not significant (P-values were 0.084 and 0.088, respectively). In contrast to the effects of the FF and WT treatments, the C stocks and oven-dry mass of the forest floor under the SO treatment were significantly increased over the period by 2.71 Mg ha1 and 8.02 Mg ha1 (19% and 29%), respectively. The change in N stocks under the SO treatment was positive but did not represent a significant increase. Results of the pair-wise analysis that tested differences in mean change among harvest residue treatments were consistent with the results of the analysis testing mean change relative to zero within individual treatments as described above (Table 1). 3.2. Effects of fertilisation on the forest floor pool The C and N stocks and oven-dry mass of the forest floor were significantly decreased over the 15-year period in the absence of fertilisation with reductions of 2.98, 0.10, and 4.34 Mg ha1, respectively (Table 2). However, there were no significant changes observed in the fertilised plots. Pre-treatment stocks are represented by a single mean value because the main plots had not been split at the time of the pre-treatment sampling. For C stocks and oven-dry mass, results of the pair-wise analysis that tested differences in mean change among fertiliser treatments were generally consistent with the results of the analysis testing mean change relative to zero within individual treatments as described above (Table 2). Although, the change in N stocks in response to fertiliser application was marginally not significantly different to the change in N stocks observed in the absence of fertiliser (P = 0.054). 3.3. Effects of residue management on the mineral soil pool Concentrations of C and N and the C:N ratio of the 0–0.3 m mineral soil depth range did not significantly change in response to harvest residue management over the 15-year period (Table 3). However, it is worth noting that the changes observed in C and

N concentration under the FF treatment in the 0–0.3 m depth range were negative whereas the changes under the WT and SO treatments were positive. The N concentration in the 0–0.1 m depth range was significantly decreased under the FF treatment but did not change significantly under the WT and SO treatments over the period. A similar result was observed for C concentration but the change under the FF treatment was marginally not significant (P = 0.058). Corresponding to the reduction in N concentration and the proportionally smaller (non-significant) reduction in C concentration in the 0–0.1 m depth range under the FF treatment, the C:N ratio was significantly increased by 1.23 over the period. The C:N ratio was significantly decreased by between 3.00 and 3.92 under all residue management treatments in the 0.2–0.3 m depth range due to increased N concentrations and decreased C concentrations (non-significant in all cases). No other significant effects were observed within individual depth ranges over the period. Pair-wise analysis revealed that the changes in C and N concentrations in the 0–0.1 m depth range under the FF treatment were significantly different to those under the SO treatment over the 15-year period (Table 3). Changes in C and N concentrations under the WT treatment were not significantly different to those under the SO and FF treatments. 3.4. Effects of fertilisation on the mineral soil pool Fertilisation did not significantly change the C and N concentrations in the 0–0.3 m depth range over the 15-year period (Table 4). However, the C:N ratio was significantly decreased by 1.98 in fertilised plots. The C:N ratio in the 0–0.1 m depth range was significantly increased by 1.64 over the period in unfertilised plots but did not significantly change in the fertilised plots. In contrast, the C:N ratio in the 0.1–0.2 m depth range was significantly decreased by 2.94 under fertilisation and there was no significant change in the C:N ratio of the unfertilised plots. In the 0.2–0.3 m depth range the C:N ratio was significantly decreased in both fertilised and unfertilised plots. However, the decrease in the fertilised plots (4.60) was close to double that (2.56) in the unfertilised plots. The concentrations of N in the 0.2–0.3 m depth range were significantly increased in both fertilised and unfertilised plots over the period. Changes (non-significant) in C concentration below 0.1 m depth were negative in fertilised plots and positive in unfertilised over the period. Pair-wise analysis indicates that changes in N concentration and C:N ratio within the 0–0.3 m depth range under fertiliser application were significantly different to those observed in the absence of fertiliser over the 15-year period (Table 4). The positive change in N concentration within the 0–0.3 m depth range was around an order of magnitude greater under fertilisation than in the absence of fertiliser whereas the change in C:N ratio was negative under fertilisation and neutral without fertiliser. The changes in C:N ratio under fertilisation were significantly different to those without fertiliser in all individual depth ranges over the period whereas the change in N concentration was significantly different in the 0–0.1 m depth range alone. Consistent with the results of the

Table 1 Mean pre-treatment (PT) C and N stocks and oven-dry mass in the forest floor (LFH horizon) and mean change at mid-rotation (DMR) under forest floor removal (FF), whole-tree (WT), and stem-only (SO) harvest residue treatments. Treatment

FF WT SO a b

C stock (Mg ha1)

N stock (Mg ha1) a

b

Mass (Mg ha1) a

b

PT

DMR

P-value

PT

DMR

P-value

PT

DMRa

P-valueb

17.88 18.09 14.14

3.07b 3.73b 2.71a

0.040 0.019 0.047

0.46 0.50 0.37

0.10b 0.15b 0.06a

0.049 0.010 0.173

35.54 35.67 27.65

4.82b 4.73b 8.02a

0.084 0.088 0.010

Means within a column carrying the same letter do not differ significantly (Tukey–Kramer test, 6 d.f., a = 0.05). P-value for the comparison of mean change relative to zero change. P-values less than 0.050 indicate a statistically significant change.

343

H.S. Jones et al. / Forest Ecology and Management 262 (2011) 339–347

Table 2 Mean pre-treatment (PT) C and N stocks and oven-dry mass in the forest floor (LFH horizon) and mean change at mid-rotation (DMR) under fertilised (F) and unfertilised (UF) treatments. Treatment

UF F a b

C stock (Mg ha1)

N stock (Mg ha1) a

b

Mass (Mg ha1) a

b

PT

DMR

P-value

PT

DMR

P-value

PT

DMRa

P-valueb

16.70

2.98b 0.25a

0.012 0.805

0.44

0.10 0.03

0.029 0.497

32.96

4.34b 3.32a

0.043 0.131

Means within a column carrying the same letter do not differ significantly (Tukey-Kramer test, 6 d.f. for PT, 7 d.f. for DMR, a = 0.05). P-value for the comparison of mean change relative to zero change. P-values less than 0.050 indicate a statistically significant change.

Table 3 Mean pre-treatment (PT) C and N concentrations and C:N ratio in the mineral soil and mean change at mid-rotation (DMR) under forest floor removal (FF), whole-tree (WT), and stem-only (SO) harvest residue treatments. Depth (m)

a b

Treat.

C concentration (g 100 g1)

N concentration (g 100 g1)

C:N ratio

PT

DMRa

P-valueb

PT

DMRa

P-valueb

PT

DMRa

P-valueb

0–0.1

FF WT SO

6.49 6.28 6.37

0.74b 0.46ab 0.69a

0.058 0.201 0.073

0.34 0.32 0.32

0.06b 0.01ab 0.03a

0.025 0.470 0.127

18.82 19.85 19.75

1.23 0.60 0.004

0.020 0.173 0.992

0.1–0.2

FF WT SO

3.14 3.20 3.03

0.41 0.22 0.18

0.244 0.505 0.599

0.15 0.17 0.17

0.003 0.02 0.02

0.876 0.302 0.208

21.00 19.56 18.69

3.09 0.80 1.59

0.063 0.580 0.288

0.2–0.3

FF WT SO

1.54 2.01 1.73

0.03 0.13 0.13

0.869 0.408 0.399

0.08 0.09 0.09

0.01 0.01 0.01

0.142 0.115 0.125

19.35 21.11 19.70

3.00 3.82 3.92

0.002 0.001 0.0004

0–0.3

FF WT SO

3.43 3.65 3.51

0.35 0.16 0.20

0.136 0.454 0.348

0.18 0.18 0.18

0.01 0.02 0.02

0.474 0.243 0.115

19.56 20.01 19.41

0.95 0.72 1.22

0.156 0.264 0.084

Means within a column and depth increment carrying the same letter do not differ significantly (Tukey–Kramer test, 6 d.f., a = 0.05). P-value for the comparison of mean change relative to zero change. P-values less than 0.050 indicate a statistically significant change.

Table 4 Mean pre-treatment (PT) C and N concentration and C:N ratio in the mineral soil and mean change at mid-rotation (DMR) under fertilised (F) and unfertilised (UF) treatments. Depth (m)

Treat.

C concentration (g 100 g1) PT

a b

a

DMR

N concentration (g 100 g1) P-value

b

a

C:N ratio b

PT

DMR

P-value

PT

DMRa

P-valueb

0–0.1

UF F

6.38

0.11 0.16

0.629 0.474

0.33

0.02b 0.01a

0.182 0.317

19.47

1.64a 0.41b

0.001 0.247

0.1–0.2

UF F

3.12

0.11 0.11

0.606 0.587

0.16

0.01 0.02

0.423 0.153

19.75

0.71a 2.94b

0.463 0.012

0.2–0.3

UF F

1.76

0.02 0.21

0.893 0.114

0.09

0.01 0.01

0.041 0.042

20.05

2.56a 4.60b

0.001 <0.0001

0–0.3

UF F

3.53

0.08 0.06

0.582 0.646

0.18

0.003b 0.02a

0.730 0.109

19.66

0.05a 1.98b

0.926 0.003

Means within a column and depth increment carrying the same letter do not differ significantly (Tukey-Kramer test, 6 d.f. for PT, 9 d.f. for DMR, a = 0.05). P-value for the comparison of mean change relative to zero change. P-values less than 0.050 indicate a statistically significant change.

analysis of change relative to zero with treatments, change in C concentration was similar under both treatments in all depth ranges. In the 0–0.1 m depth range, the change in C:N ratio was neutral under fertiliser application and positive in the absence of fertiliser. Below 0.1 m depth, the change in C:N ratio was negative under both fertilised and unfertilised plots but the change was much greater under fertiliser application.

of N stocks within the top 0.3 m of mineral soil. The contribution of the coarse fraction was greatest in the top 0.1 m of mineral soil and decreased with depth down the profile as did the mass of the coarse fraction (in response to a decrease in the proportion of pumice gravel with depth). Fine fraction mineral soil density increased with depth down the profile. 3.6. Effects of residue management and fertilisation on tree growth

3.5. Coarse fraction stocks in the mineral soil pool The mineral soil coarse (>2 mm) fraction, which was predominantly comprised of fine pumice gravel, contributed 2.70 Mg ha1 to total (fine plus coarse fraction) C stocks and 0.09 Mg ha1 to total N stocks in the 0–0.3 m depth range (Table 5). Proportionally, the coarse fraction accounted for 4.0% of total C stocks and 2.8%

Basal area was analysed as an indicator of tree growth to determine the effects of harvest residue management and fertilisation on tree growth at mid-rotation. Residue management had no significant effect on basal area (Fig. 1a). However, a clear trend in mean values consistent with treatment intensity (SO > WT > FF) is apparent. Mean basal area ranged from 31.9 m2 ha1 under the

344

H.S. Jones et al. / Forest Ecology and Management 262 (2011) 339–347

Table 5 Mean C and N stocks and density of fine and coarse mineral soil fractions under unfertilised plots at mid-rotation with proportion that each fraction contributes to the total. Depth (m)

Fraction

C stocks

N stocks

Mean (Mg ha

a b

1

)

a

b

Density

S.E.

Prop. (%)

Mean (Mg ha

1

)

S.E.

a

b

Prop. (%)

Mean (g cm3)

S.E.a

Prop.b (%)

0–0.1

<2 mm >2 mm Total

32.93 1.62 34.54

1.52 0.19 1.50

95.3 4.7 100

1.58 0.05 1.63

0.09 0.004 0.09

96.7 3.3 100

0.50 0.03 0.53

0.01 0.002 0.01

94.7 5.3 100

0.1–0.2

<2 mm >2 mm Total

19.28 0.70 19.98

0.99 0.15 1.02

96.5 3.5 100

1.02 0.03 1.05

0.06 0.003 0.06

97.5 2.5 100

0.59 0.03 0.62

0.01 0.003 0.01

95.6 4.4 100

0.2–0.3

<2 mm >2 mm Total

11.88 0.38 12.25

0.74 0.07 0.78

96.9 3.1 100

0.68 0.01 0.69

0.03 0.003 0.03

97.9 2.1 100

0.67 0.03 0.70

0.02 0.003 0.02

95.7 4.3 100

0–0.3

<2 mm >2 mm Total

64.08 2.70 66.78

2.75 0.36 2.73

96.0 4.0 100

3.28 0.09 3.37

0.16 0.01 0.16

97.2 2.8 100

0.59 0.03 0.62

0.01 0.002 0.01

95.4 4.6 100

Standard error about the mean (6 d.f.). Proportion that each mineral soil fraction contributes to the total on average.

(a)

42

4. Discussion

Basal Area (m2 ha-1)

40

4.1. Effects of residue management on the forest floor pool

38 36 34 32 30 28

FF

WT

SO

Treatment

Basal Area (m2 ha-1)

(b) 38 36 34 32 30 28

UF

F

Treatment Fig. 1. Mean basal area of the second rotation stand at mid-rotation (a) under forest floor removal (FF), whole-tree (WT), and stem-only (SO) harvest residue treatments (6 d.f.) and (b) under fertilised (F) and unfertilised (UF) treatments (8 d.f.). Error bars represent the standard error about the mean.

FF treatment to 37.5 m2 ha1 under the SO treatment. The difference in mean basal area between the SO and WT treatments (4.2 m2 ha1) was almost three times greater than that between the FF and WT treatments (1.5 m2 ha1). Mean basal area in fertilised plots (35.0 m2 ha1) was greater than, but marginally (P = 0.057) not significantly different to, that in unfertilised plots (33.5 m2 ha1) at mid-rotation (Fig. 1b). Variation in tree growth across the trial at mid-rotation may be limiting the detection of statistically significant harvest residue and fertilisation effects.

The results indicate that C and N storage in the forest floor would be reduced through to mid-rotation and possibly beyond by harvest residue removal, either with or without forest floor removal, for bioenergy production. Furthermore, harvest residue manipulations may have implications for the productivity of the following rotation. Tree growth at stand age 15 years (as reflected by mean basal area) was greater under residue retention than under residue removal and residue plus forest floor removal by 4.2 and 5.6 m2 ha1, respectively (although the differences were not statistically significant). Consistent with this finding, Oliver et al. (2011) reported that the carbon stocks in the total above ground biomass at stand age 15 years in unfertilised plots were lowest under residue plus forest floor removal (71 Mg ha1) and greatest under residue retention (85 Mg ha1). Residue management-driven stand productivity differences may have contributed to the differences among treatments in terms of the change in C and N storage observed in the forest floor over the 15 year period via presumed differences in litter inputs. However, it is also possible that some material from the harvest residues retained on the SO plots was present in the forest floor at mid-rotation (stand age 15 years). The application of a decay rate for deadwood and branch material developed for P. radiata forests in New Zealand (Garrett et al., 2010), including data from the Kinleith site, suggests that approximately 8 Mg ha1 (18%) of the branch component (44 Mg ha1) of the harvest residues retained in the SO treatment (50 Mg ha1) were likely to have been present at mid-rotation. It was estimated that none of the foliage component of the residues would have persisted. Residue retention increased forest floor oven-dry mass by 8.02 Mg ha1 over the 15year period (Table 1) and it also offset a loss of 4.73 Mg ha1 that would have occurred if the residues had not be retained (WT treatment). Therefore, it is estimated that the residue retention treatment accounted for about 13 Mg ha1 of the forest floor mass at mid-rotation (36 Mg ha1) and that approximately 5 Mg ha1 of this may be attributed either to productivity gains or the prevention of forest floor decomposition – via physical protection provided by retained residues – (or both) as the remaining 8 Mg ha1 is directly attributed to remnant harvest residue material. It should be noted that the remnants of decomposing stems derived from the thinning at stand age 5 years may have contributed to the forest floor sampled at mid-rotation in the second rotation stand under all treatments whereas the first rotation stand

H.S. Jones et al. / Forest Ecology and Management 262 (2011) 339–347

was production thinned and so no thinned stems would have contributed to the forest floor sampled prior to harvest. Inputs from the thinning at stand age 5 years may have been slightly greater under SO treatment compared to WT and FF treatments (Oliver et al., 2011) and slightly greater in fertilised plots compared to unfertilised. The C and N stocks of the forest floor observed under the FF treatment at mid-rotation must necessarily reflect the extent of stock recovery over the 15-year period because the forest floor of the first rotation stand was completely removed at the time of treatment application (meaning that C and N stocks were reduced to zero). Therefore, the FF treatment provides an opportunity to examine the rate of C and N stock recovery at the Kineith site and to project the time required for the stocks to reach pre-treatment levels. Given that 14.70 Mg ha1 of C and 0.36 Mg ha1 of N accumulated during the reformation of the forest floor over the period, the rates of stock recovery were approximately 1 Mg ha1 year1 (5.5% of original stock per year) for C and 0.02 Mg ha1 year1 (5.2% of original stock per year) for N assuming linear change. Approximately 3 Mg ha1 of C and 0.10 Mg ha1 of N were still to be gained before the stocks under the FF treatment returned to pre-harvest levels. On this basis, it is estimated that full recovery would have occurred at about stand age 18 years for C and 20 years for N assuming the same linear rate of recovery applied. Smaill et al. (2008b) measured the rate of litter-fall mass accumulation at all six LTSP Series I trial sites which ranged in stand age from 8 to 16 years at the time of sampling. On average, the daily rates they presented equate to about 2.8 Mg ha1 year1 which, allowing for some decomposition of the fresh litter, is fairly consistent with the rate of forest floor mass recovery calculated here. Full recovery of C and N stocks in the forest floor (from either complete removal under the FF treatment or from the period of limited inputs between harvest and crown closure in the absence of slash addition under the WT and FF treatments) to the levels observed under the mature, first-rotation forest clearly takes longer than 15 years. However, the 15-year period was sufficiently long for the C and N stocks in forest floor pool under the FF treatment to recover to the levels they would have been at had the forest floor not been removed (i.e., similar to the levels observed under the WT treatment). Furthermore, the effects of residue retention were found to persist for at least 15 years. These findings, together with the above estimate of forest floor C and N stock recovery time under the FF treatment, agree with those of Smaill et al. (2008a), who reported that significant effects of harvest residue management on FH horizon C and N stocks across four P. radiata LTSP Series I sites (including Kinleith) throughout New Zealand could, in some cases, persist through to mid-rotation and asserted that the effects may even persist through to the end of the rotation. The two studies are not directly comparable however, as Smaill et al. (2008a) did not report stocks for the combined LFH horizon and the effects reported were observed at earlier stages of stand development at some sites. Crown closure occurred about stand age 6 years at the site. Foliage biomass at stand age 5 years (Oliver et al., 2011) suggests that crown closure in SO plots may have occurred slightly sooner than in WT or FF plots. Crown closure is also expected to have been more rapid in fertilised plots than in unfertilised. The similarity in C and N stocks (and oven-dry mass) of the FF and WT treatments at mid-rotation may be explained by initial net decomposition of the forest floor under the WT treatment during the early stages of stand development – due to the combined effect of limited litter inputs and exposure – while net accumulation of litter occurred under the FF treatment over the same period. As litter inputs increased and the amount of exposure decreased with stand development, net accumulation of material in the forest floor would have begun under the WT treatment and continued

345

under the FF treatment. At some time prior to the mid-rotation sampling, the trajectories of forest floor accumulation under the two treatments appear to have coalesced. 4.2. Effects of fertilisation on the forest floor pool The results tend to suggest that fertiliser application improved tree growth which, in turn, increased litter inputs to the forest floor and thus offset reductions in oven-dry mass and C and N stocks associated with harvest residue manipulations. Although the difference in mean basal area between the fertilised and unfertilised plots was not quite significant (P = 0.057) at mid-rotation, a statistically significant growth difference was observed earlier in the rotation. Smith et al. (2000) found that fertilisation significantly increased tree diameter at breast height at stand age 4 and 5 years. More recently, Huang et al. (2011) reported that combined estimates of litter-fall and thinning inputs at the Kinleith trial were greater in fertilised plots than in unfertilised at mid-rotation. The findings of Smith et al. (2000) and Huang et al. (2011) support our assertion that a growth response to fertilisation occurred. Furthermore, the significant decline in C:N ratio of the top 0.3 m of mineral soil observed under fertilised plots in the present study (Table 4) indicates that fertiliser application improved soil fertility. Fertiliser application ceased at around stand age 6 years, but the effect on the forest floor was still observable at mid-rotation. While it should be noted that the rate of N application (averaging 190 kg ha1 year1) in the trial was much higher than what would normally be applied in New Zealand plantation forests, the results show that the adverse effects of harvest residue manipulations on C and N storage in the forest floor can be successfully ameliorated by fertilisation. Fertiliser applications should always be managed carefully, with consideration of the sensitivity of the receiving environment, and in-line with relevant environment regulations or guidelines. 4.3. Effects of residue management on the mineral soil pool The significant decrease in the concentration of N and increase in the C:N ratio in the 0–0.1 m depth range that occurred under harvest residue plus forest floor removal (FF treatment) over the 15-year period can be explained by the complete removal of the forest floor which would have considerably reduced organic matter and nutrient inputs to the surface mineral soil. The reduced inputs would have persisted until the forest floor began to reform as the second-rotation stand developed. The temperature of the surface mineral soil exposed by forest floor removal may have been increased and soil moisture content decreased. The latter is supported by Smaill et al. (2008a) who found that the surface soil moisture content was decreased by forest floor removal across four LTSP Series I sites. A decrease in soil moisture content probably subdued microbial activity and so the reduction in organic matter inputs is likely to be the main cause of the effect on N concentration observed in the present study. Surface erosion was not observed to have occurred at the trial site during the period the soil in the FF treatment plots was exposed (or at any other time) ruling-out topsoil loss as a possible mechanism for the observed effect under residue plus forest floor removal. Modelling undertaken by Kirschbaum et al. (2008) has shown that accumulation of N in a developing forest floor is associated with reductions of N in the mineral soil. The increase in C:N ratio (suggesting a decline in fertility) in the top 0.1 m of mineral soil under the harvest residue plus forest floor removal treatment is likely to have contributed to the mean basal area under the same treatment being the lowest of the three treatments (although the differences were not statistically significant).

346

H.S. Jones et al. / Forest Ecology and Management 262 (2011) 339–347

The results highlight the importance of forest floor retention in limiting the risk of C and N loss from the surface mineral soil and suggest that forest floor disturbance should be avoided during harvesting and biofuel extraction operations. The lack of any significant effect of residue removal without forest floor removal (WT treatment) indicates that the extraction of residues for bioenergy production would not adversely impact mineral soil C and N provided the extraction was undertaken with care to avoid forest floor disturbance. Significant effects of harvest residue management on the C and N concentrations of the mineral soil over the 15-year period were restricted to the 0–0.1 m depth range (Table 3). This result confirms the surficial nature of harvest residue management effects and is consistent with findings of a similar study (Jones et al., 2008), although Powers et al. (2005) reported that effects of forest floor removal extended to 0.2 m depth at sites across the United States. The C:N ratio was significantly reduced under all treatments in the 0.2–0.3 m depth range but this probably reflects a general downward movement of N within the profile after harvest. The lack of a significant increase in the C concentration of the top 0.1 m of mineral soil – despite a significant increase in C stocks in the forest floor – under the SO plots suggests that soil mixing processes at the interface between the mineral soil and forest floor were not particularly active, or at least unable to readily respond to the increased supply of organic substrate, at the Kinleith site. This could perhaps be a function of the relatively moderate mean annual temperature (11.5 °C) at the site. Powers et al. (2005) postulated that under moderate or warmer temperatures only a minor proportion of decomposing residues is incorporated into the mineral soil with the remainder respired to the atmosphere. Scott et al. (2006) reported that only about 10% of C accumulated through P. radiata litter-fall actually entered the mineral soil over the course of a 26-year rotation at an agro-forestry trial site in the North Island. 4.4. Effects of fertilisation on the mineral soil pool The application of about 190 kg ha1 year1 of urea-N fertiliser between second-rotation stand ages 1 and 6 years caused a divergence in C and N concentrations in the top 0.3 m of mineral soil by marginally boosting levels of N and reducing levels of C, thus decreasing the C:N ratio over the 15-year period. The significant increase in the C:N ratio of the top 0.1 m of mineral soil in unfertilised plots – presumably a reflection the net effects of the residue management treatments in the absence of fertiliser – was effectively mitigated by fertiliser application via the manipulation of the N concentration. Huang et al. (2011) reported that N fertilisation at the Kinleith site significantly increased the concentration of C in the top 0.05 m of mineral soil at mid-rotation (comparison of means mid-rotation). They suggested that the increase in C concentration was due to N fertilisation driving increased stand productivity which, in turn, led to increased organic matter inputs to the mineral soil. Results of the present study (Table 4) indicate that this effect did not extend much below 0.05 m and may reflect a depth limitation to the incorporation of organic matter inputs. Huang et al. (2011) also found, via 14C and lignin oxidation analyses, that N fertilisation increased the rate of decomposition of both recalcitrant and labile soil C in the 0– 0.05 m depth range. The detection of significant effects of fertiliser application nine years after the cessation of applications indicates that fertilisation early in the rotation – at least at the high rate of N applied in this study – provides benefits in terms of improved fertility (decreased C:N ratio) and prevention of fertility decline through to at least mid-rotation. The decrease in C:N ratio observed in response to fertilisation in the top 0.3 m of mineral soil is most likely responsible

for the greater (close to significant, P = 0.057) mean basal area in fertilised plots compared to unfertilised. This result suggests that a proportion of the N applied as fertiliser was incorporated into the standing biomass via uptake from the mineral soil. The significant decrease in the C:N ratio of the 0.2–0.3 m depth ranges and the significant increase in N concentration in the 0.2– 0.3 m range over the 15-year period in both fertilised and unfertilised plots suggests that there has been a general migration of N down the profile over the period. Downward movement of N after harvest and prior to crown closure may be expected due to increased infiltration of rainfall and generally subdued plant uptake following weed control. 4.5. Coarse fraction stocks in the mineral soil pool The contribution of the coarse fraction to total mineral soil C and N stocks of the 0–0.3 m depth range was relatively small at the Kinleith site. Organic and gravel components of the coarse fraction were not analysed separately. However, the similarity of coarse and fine fraction C concentrations (a difference in mean concentrations of 0.56 g 100 g1 in the 0–0.3 m depth range) suggests that small quantities of organic matter fragments with relatively high C concentration probably accounted for the majority of the coarse fraction C stock. Jones et al. (2008) reported that the coarse fraction accounted for almost 25% of total mineral soil C stocks in the top 0.3 m at the Tarawera LTSP Series 1 site. However, much of the C stock in the coarse fraction was found to be in fine fraction material contained within the pores of the highly vesicular scoria gravels at Tarawera. The results of the present study indicate that the pumice gravels at Kinleith – which were generally less vesicular, finer, and less abundant than the scoria gravels at Tarawera – entrained less fine fraction material. Avoidance of soil physical disturbance during the application of treatments at the LTSP Series 1 trial sites may explain the relatively small amounts of coarse organic material observed at both Kinleith and Tarawera. The source of coarse organic material contained in samples is likely to be the decaying roots of the harvested trees (Powers et al., 2005) which fragmented in the process of sample collection. Based on the results of the present study, further decomposition of dead roots from the first rotation stand over time could be expected to add up to 2.7 Mg ha1 of C to the fine fraction mineral soil C stocks in the top 0.3 m at Kinleith. Although the magnitude of the addition is relatively small, it does highlight the importance of including an assessment of coarse fraction stocks in the monitoring of soil C stock changes over time. 5. Conclusions The removal of harvest residues (either with or without forest floor removal) for bioenergy production can be expected to reduce C and N stocks in the forest floor through to, and perhaps beyond, mid-rotation based on the findings of this study. It was estimated that full recovery of the forest floor pool C and N stocks following complete forest floor removal would have occurred at about stand age 18 years for C and 20 years for N assuming a linear rate of recovery. Removal of the forest floor decreased organic matter inputs to the mineral soil which, in turn, decreased the N concentrations and increased the C:N ratio in the top 0.1 m of mineral soil over the 15-year period. However, the results suggest that harvest residue removal for bioenergy production could proceed without decreasing mineral soil C and N concentrations provided the forest floor is not disturbed in the process. The application of urea-N fertiliser can mitigate reductions in C and N stocks of the forest floor via improved fertility of the top 0.3 m of mineral soil and enhanced tree growth. However, the rate at which urea-N fertiliser was

H.S. Jones et al. / Forest Ecology and Management 262 (2011) 339–347

applied in this study was high (190 kg ha1 year1). Measurement and monitoring of changes in soil C and N and other properties under managed plantation forests over time, including full and multiple rotations, are required in New Zealand across a range of site conditions to improve our understanding of the rates of change and longer-term impacts of management practices. Role of the funding source The research and preparation of the manuscript was funded by the New Zealand Ministry for the Environment. The Ministry had no role in the experimental design; in data collection, analysis, or interpretation; or in the preparation of the manuscript. The Ministry supported the decision to submit the manuscript for publication. Acknowledgements The authors thank all those who assisted with the field sampling, and sample preparation and analysis (G. Oliver, S. Pearce, D. Graham, A. Leckie, G. Coker, S. Smaill, R. Bagnal, and K. Eason). Thanks also to Karen Chambers for assembling the dataset. Funding was provided by the New Zealand Ministry for the Environment. References Chen, C.R., Xu, Z.H., 2005. Soil carbon and nitrogen pools and microbial properties in a 6-year-old slash pine plantation of subtropical Australia: impacts of harvest residue management. Forest Ecology and Management 206, 237–247. Dyck, W.J., Hodgkiss, P.D., Oliver, G.R., Mees C.A., 1991. Harvesting sand-dune forests: impacts on second-rotation productivity. In: Dyck, W.J., Mees, C.A. (Ed.), Long-term Field Trials to Assess Environmental Impacts of Harvesting. In: Proceedings, IEA/BE T6/A6 Workshop, Florida, USA, February 190. IEA/BE T6/A6 Report No. 5. Forest Research Institute, Rotorua, New Zealand, NZ FRI Bulletin No. 161, pp. 163–176. Fox, T.R., 2000. Sustained productivity in intensively managed forest plantations. Forest Ecology and Management 138, 187–202. Garrett, L.G., Kimberley, M.O., Oliver, G.R., Pearce, S.H., Paul, T.S.H., 2010. Decomposition of wood debris in managed Pinus radiata plantations in New Zealand. Forest Ecology and Management 260, 1389–1398. Hall, P.W., Hock, B.K., Palmer, D.J., Kimberley, MO., Walter, C., Wilcox, P.L., Jack, M., Pawson, S.M., Giltrap, D.J., Newsome, P.F., Dymond, J., Aussile, A.G., Ekanayake, J., Todd, M., Zhang, W., Kerr, S., Stroombergen, A., 2009. Bioenergy options for New Zealand – analysis of large-scale bioenergy from forestry. Scion. ISBN 9780-478-11026-X. Hewitt A.E., 1998. New Zealand Soil Classification. Landcare Research Science Series No. 1. Manaaki Whenua — Landcare Research, Lincoln, New Zealand. Huang, Z., Clinton, P.W., Baisden, W.T., Davis, M.R., 2011. Long-term nitrogen additions increased surface soil carbon concentration in a forest plantation despite elevated decomposition. Soil Biology and Biochemistry 43, 302–307. Jandl, R., Lindner, M., Vesterdal, L., Bauwens, B., Baritz, R., Hagedorn, F., Johnson, D.W., Minkkinen, K., Byrne, K.A., 2007. How strongly can forest management influence soil carbon sequestration? Geoderma 137, 253–268. Johnson, D.W., 1992. Effects of forest management on soil carbon storage. Water, Air, and Soil Pollution 64, 83–120. Johnson, D.W., Curtis, P.S., 2001. Effects of forest management on soil C and N storage: meta analysis. Forest Ecology and Management 140, 227–238. Jones, H.S., Garrett, L.G., Beets, P.N., Kimberley, M.O., Oliver, G.R., 2008. Impacts of harvest residue management on soil carbon stocks in a plantation forest. Soil Science Society of America Journal 72, 1621–1627. Kim, C., 2008. Soil carbon storage, litterfall and CO2 efflux in fertilized and unfertilized larch (Larix leptolepis) plantations. Ecological Research 23, 757–763. Kirschbaum, M.U.F., Guo, L.B., Gifford, R.M., 2008. Why does rainfall affect the trend in soil carbon after converting pasture to forests? A possible explanation based on nitrogen dynamics. Forest Ecology and Management 255, 2990–3000. Laiho, R., Sanchez, F., Tiarks, A., Dougherty, P.M., Trettin, C.C., 2003. Impacts of intensive forestry on early rotation trends in site carbon pools in the southeastern US. Forest Ecology and Management 174, 177–189.

347

Lal, R., 2005. Forest soils and carbon sequestration. Forest Ecology and Management 220, 242–258. Leathwick, J., Wilson, G., Rutledge, D., Wardle, P., Morgan, F., Johnston, K., McLeod, M., Kirkpatrick, R., 2003. Land Environments of New Zealand. Ministry for the Environment/Manaaki Whenua Landcare Research, Wellington/Hamilton, 184 pp. McFarlane, K.J., Schoenholtz, S.H., Powers, R.F., 2009. Plantation management intensity affects belowground carbon and nitrogen storage in northern California. Soil Science Society of America Journal 73, 1020–1032. McLaughlin, J.W., Phillips, S.A., 2006. Soil carbon, nitrogen, and base cation cycling 17 years after whole-tree harvesting in a low-elevation red spruce (Picea rubens)-balsam fir (Abies balsamea) forested watershed in central Maine, USA. Forest Ecology and Management 222, 234–253. Nave, L.E., Vance, E.D., Swanston, C.W., Curtis, P.S., 2009. Impacts of elevated N inputs on north temperate forest soil C storage, C/N, and net N-mineralization. Geoderma 153, 231–240. Nave, L.E., Vance, E.D., Swanston, C.W., Curtis, P.S., 2010. Harvest impacts on soil carbon storage in temperate forests. Forest Ecology and Management 259, 857– 866. Oliver, G.R., Beets, P.N., Pearce, S.H., Graham, J.D., Garrett, L.G., 2011. Carbon Accumulation in two Pinus radiata stands in the North Island of New Zealand. New Zealand Journal of Forestry Science 41, 71–86. Peng, Y., Thomas, S.C., Tian, D., 2008. Forest management and soil respiration: implications for carbon sequestration. Environmental Reviews 16, 93–111. Powers, R.F., Scott, D.A., Sanchez, F.G., Voldseth, R.A., Page-Dumroese, D., Elioff, J.D., Stone, D.M., 2005. The North American long-term soil productivity experiment: findings from the first decade of research. Forest Ecology and Management 220, 31–50. Rifai, S.W., Markewitz, D., Borders, B., 2010. Twenty years of intensive fertilization and competing vegetation suppression in loblolly pine plantations: impacts on soil C, N, and microbial biomass. Soil Biology and Biochemistry 42, 713–723. Schipper, L.A., Dodd, M.B., Fisk, L.M., Power, I.L., Parenzee, J., Arnold, G., 2011. Trends in soil carbon and nutrients of hill-country pastures receiving different phosphorus fertilizer loadings for 20 years. Biogeochemistry, doi:10.1007/ s10533-009-9353-5. Scott, N.A., Tate, K.R., Ross, D.J., Parshotam, A., 2006. Processes influencing soil carbon storage following afforestation of pasture with Pinus radiata at different stocking densities in New Zealand. Australian Journal of Soil Research 44, 85– 96. Smaill, S.J., Clinton, P.W., Greenfield, L.G., 2008a. Postharvest organic matter removal effects on FH layer and mineral soil characteristics in four New Zealand Pinus radiata plantations. Forest Ecology and Management 256, 558– 563. Smaill, S.J., Clinton, P.W., Greenfield, L.G., 2008b. Nitrogen fertiliser effects on litter fall, FH layer and mineral soil characteristics in New Zealand Pinus radiata plantations. Forest Ecology and Management 256, 564–569. Smith, C.T., Lowe, A.T., Beets, P.N., Dyck, W.J., 1994a. Nutrient accumulation in second-rotation Pinus radiata after harvest residue management and fertiliser treatment of coastal sand dunes. New Zealand Journal of Forestry Science 24, 362–389. Smith, C.T., Dyck, W.J., Beets, P.N., Hodgkiss, P.D., Lowe, A.T., 1994b. Nutrition and productivity of Pinus radiata following harvest disturbance and fertilization of coastal sand dunes. Forest Ecology and Management 66, 5–38. Smith, C.T., Lowe, A.T., Skinner, M.F., Beets, P.N., Schoenholtz, S.H., Fang, S., 2000. Response of radiata pine forests to residue management and fertilisation across a fertility gradient in New Zealand. Forest Ecology and Management 138, 203– 223. Soil Survey Staff, 2003. Keys to Soil Taxonomy, Ninth ed. Natural Resources Conservation Service, U.S. Department of Agriculture, Washington, DC. Tan, X., Chang, S.X., Kabzems, R., 2005. Effects of soil compaction and forest floor removal on soil microbial properties and N transformations in a boreal forest long-term soil productivity study. Forest Ecology and Management 217, 158– 170. Thiffault, E., Paré, D., Bélanger, N., Munson, A., Marquis, F., 2006. Harvesting intensity at clear-felling in the boreal forest: impact on soil and foliar nutrient status. Soil Science Society of America Journal 70, 691–701. Van Miegroet, H., Jandl, R., 2007. Are nitrogen-fertilized forest soils sinks or sources of carbon? Environmental Monitoring and Assessment 128, 121–131. Vanguelova, E., Pitman, R., Luiro, J., Helmisaari, H.-S., 2010. Long term effects of whole tree harvesting on soil carbon an nutrient sustainability in the UK. Biogeochemistry 101, 43–59. Watt, M.S., Davis, M.R., Clinton, P.W., Coker, G., Ross, C., Dando, J., Parfitt, R.L., Simcock, R., 2008. Identification of key soil indicators influencing plantation productivity and sustainability across a national trial series in New Zealand. Forest Ecology and Management 256, 180–190.