16
Decomposition of oak leaf litter is related to initial litter Mn concentrations Matthew P. Davey, Bjo¨rn Berg, Bridget A. Emmett, and Phil Rowland
Abstract: The factors determining the quantity of litter being incorporated into stable organic matter were examined as part of a broader study investigating carbon (C) sequestration in forest ecosystems. Litter was collected from 20 common oak (Quercus robur L.) stands in Wales (UK) and placed in litter-decomposition bags. These bags were installed in an oak stand for 3, 6, 12, 21, and 31 months to study the effect of litter quality on decomposition (mass loss) rates and the limit value for a broad-leaf species. Results indicate that the initial decomposition rate is highly correlated with the manganese content of the litter (P = 0.007, R2 = 0.34). In the final stages of decomposition, limit values ranged between 57% and 95% of initial litter mass. These estimated limit values were not significantly correlated with initial concentrations of other nutrients. However, Ca concentrations gave a significance level of P = 0.067. Estimated rates of C sequestration in soil ranged from 0.93 to 80.22 g Cm–2year–1. Key words: decomposition, limit values, manganese, mass loss, oak, Quercus robur, C sequestration. Re´sume´ : Dans le cadre d’une e´tude plus large sur la se´questration du carbone dans les e´cosyste`mes forestiers, les auteurs ont examine´ les facteurs implique´s dans la quantite´ de litie`re incorpore´e dans la matie`re organique stable. Ils ont re´colte´ de la litie`re dans 20 peuplements de cheˆne commun (Quercus robur L.) en Wales (UK) et l’ont place´e dans des sacs pour la de´composition de la litie`re. Ces sacs ont e´te´ installe´s dans un peuplement de cheˆne pendant 3, 6, 12, 21 et 31 mois, afin d’e´tudier les effets de la qualite´ de la litie`re sur les taux de de´composition (perte de masse), ainsi que la valeur limite pour cette espe`ce a` larges feuilles. Les re´sultats indiquent que le taux initial de de´composition est fortement corre´le´ avec la teneur en mangane`se de la litie`re (P = 0.007, R2 = 0.34). Aux stades finaux de la de´composition, les valeurs limitent vont entre 57 % et 95 % de la masse de litie`re originale. Ces valeurs limites estime´es ne sont pas significativement corre´le´es aux teneurs initiales en nutriments. Cependant, la teneur en Ca montre un degre´ significatif de P = 0.067. Les taux de se´questration estime´s de C dans le sol varient de 0.93 a` 80.22 g Cm–2a–1. Mots cle´s : de´composition, valeurs limites, mangane`se, perte de masse, cheˆne, Quercus robur, se´questration du C. [Traduit par la Re´daction]
Introduction For a given site and climate, one would expect the litter mass-loss rate to be related primarily to the chemical and physical properties of the litter. Such relationships have been demonstrated in many studies (e.g., Fogel and Cromack 1977; McClaugherty et al. 1985; Upadhyay and Singh 1985; Dyer 1986). The litter properties that affect the decomposition rate are often called ‘‘substrate quality’’ withReceived 2 August 2006. Published on the NRC Research Press Web site at http://canjbot.nrc.ca on 27 March 2007. M.P. Davey.1,2 Centre for Ecology and Hydrology, Deiniol Road, Bangor, Gwynedd, LL57 2UP, UK. B. Berg.3 Danish Center for Forest, Landscape, and Planning, Hørsholm Kongevej 11, DK-2970 Hørsholm, Denmark. B.A. Emmett. Centre for Ecology and Hydrology, Deiniol Road, Bangor, Gwynedd, LL57 2UP, UK. P. Rowland. Centre for Ecology and Hydrology, Lancaster Environment Centre, Library Avenue, Bailrigg, Lancaster, LA1 4AP, UK. 1Corresponding
author (e-mail:
[email protected]). address: Department of Animal and Plant Sciences, Alfred Denny Building, University of Sheffield, Western Bank, Sheffield, S10 2TN, UK. 3Present address: Department of Forest Ecology, P.O. Box 27, University of Helsinki, FIN-00014 Helsinki, Finland. 2Present
Can. J. Bot. 85: 16–24 (2007)
out identifying the chemical components. However, this concept (substrate quality) may be subdivided into at least three main groups, (i) the carbon (C) source and (ii) the nutrients, as well as (iii) "modifiers" including, for example, phenols and heavy metals (Swift et al. 1979). The initial decomposition rate has typically been related to the main nutrients such as nitrogen (N) and phosphorus (P), both considered limiting to the microbial decomposition of litter (e.g., Berg and McClaugherty 2003). In the late stage of decomposition, the litter C-source may change as the more easily degraded carbohydrates are depleted and the degradation rate of lignin dominates the decomposition process. Consequently, limiting factors may also change with higher N concentrations suppressing the formation of the ligninase system in numerous white-rot fungi (Eriksson et al. 1990), whilst manganese (Mn) can enhance decomposition rates owing to its role as a cofactor to Mn-peroxidase, an enzyme in the ligninase system (e.g., Hatakka 2001). As the litter-degradation rate decreases with accumulated litter mass loss the rate may even approach zero as the substrate becomes increasingly resistant to decomposition (Berg and Ekbohm 1991; Couˆteaux et al. 1998). Normally a limit value for decomposition can be calculated and recalcitrant remains estimated, which can provide an estimate of the amount of litter material that is incorporated into the stable soil organic matter pool, that is, the C sequestration rate.
doi:10.1139/B06-150
#
2007 NRC Canada
Davey et al.
Thus, by using an asymptotic function, Howard and Howard (1974) calculated limit values, and later Berg and Ekbohm (1991) used a simplified function that allowed the calculation of both initial rate and limit value for litter degradation. Berg and Ekbohm’s (1991) calculations encompassed several boreal and temperate litter species. For some tree species, for example, Scots pine (Pinus silvestris L.) a three-stage conceptual model is applicable (Berg and Matzner 1997), whereas for others, such as needle litter of Norway spruce (Picea abies (L.) Karst.), a different pattern is seen. For the latter species the initial decomposition rate was related to litter-Mn concentration, indicating possible participation of the lignin-degrading microflora already from litterfall. However, there are relatively few rates and limit values for foliar litter species available in the literature. Therefore, the aim of the present study was to examine the effect of litter quality on the controls of decomposition (mass loss) rates and limit values for leaf litter of common oak (Quercus robur L.). Additionally, based on the limit values and litterfall data, C sequestration of the soils can be estimated to identify conditions of high and low C sequestration rates for such forest ecosystems. To this purpose, we collected leaf litter of common oak from 20 woodland stands around Wales, thus obtaining samples that varied in chemical composition, and incubated all litter types at the same site.
Materials and methods Leaf-litter collection Oak litter was collected monthly from 20 woodland stands (see Table 1 for sites and codes) around northwest and east Wales from October to December 1998. Litter was collected using six collectors per stand installed 1 m above ground level. Woodland sites were selected based on the presence of areas dominated by oak, acid podzolic soils, and underlying Lower Palaeozoic rocks (Table 1). Apart from two coniferous sites, all sites were predominantly oak woodlands. Site characteristics are shown in Table 1 with more information available in Williams et al. (2000). Litter decomposition site location Litterbags were placed at the Abergwyngregyn oak woodland, Wales, UK (230 m a.s.l. 53822’N, 3899’W). The northfacing slope, of about 408 gradient, had canopy closure of approximately 70%–80%. The understory shrubs included hazel (Corylus avellana L.), holly (Ilex aquifolium L.), and Sitka spruce (Picea sitchensis (Bong.) Carrie`re), and ground cover was composed of tree litter, bramble (Rubus saxatilis L), grasses, moss, bluebell (Scilla nonscripta (L.) Hoffmanns. & Link), and ferns. Annual mean air temperature for the site was 11.7 8C, mean soil temperature was 10.6 8C, total annual rainfall was 1470 mm, and evapotranspiration, estimated using the chloride method, was 975 mm (Williams et al. 2000). The limitations of litterbag experiments are acknowledged, such as potential underestimation of decay rates due to the absence of mechanical and macrodetritivore damage and a potential increase in humidity within the bag (Yates and Day 1983). However, methods of estimating rates of decomposition of unconfined litter can overestimate decay rates because of harsh mechanical dam-
17
age. The litterbag technique was considered to be a more reliable, controlled, and reproducible technique for this study. The mesh size used was intended to allow fungal, moss, and various detritivorous species to come in contact with the decaying leaves and once the bags were opened, these types of species were found in the litter. Litter-incubation study Litterbags were constructed from nylon mesh fabric measuring 30 cm 8 cm, with 1 mm mesh on the upper surface to allow access by some soil-dwelling invertebrates, and smaller mesh on the lower surface to enable penetration by fungal hyphae but also to prevent loss of material. Each litterbag was filled with 3.000 ± 0.001 g air-dried (room temperature) whole oak leaf litter. To standardize the air-dried measurements to later decomposition weights, a correction factor was determined by oven-drying (80 8C) 3.000 g of air-dried oak litter from each of the 20 sites and calculating the average percent loss of moisture. Fifteen bags of leaf litter were assembled for each site, totalling 300 litterbags for the 20 sites. The 300 bags were divided into three replicate plots in three separate but similar areas of the woodland. Each plot of 100 bags was divided into four smaller groups of 25 bags. Each group of 25 bags contained five replicate litterbags from five different sites. The bags were laid out in a grid 5 bags across and 5 bags down, with the 5 down (one bag from each site) being strung together for ease of collection. The area covered by each of the four collections (25 bags) was ca. 3 m 5.5 m. The litterbags were deployed on 20 February 2002. Five litterbags per site per replicate plot were collected after 3 months (89 d; 20 May 2002), 6 months (181 d; 20 August 2002), 12 months (379 d; 6 March 2003), 21 months (636 d; 18 November 2003), and 31 months (936 d; 13 September 2004). Processing litterbags Litterbags were taken directly to the laboratory and the exteriors immediately cleared of moss and other large material. The bags were placed in an oven (40 8C) for 1–4 d. The litterbags were then carefully cut open and the remaining oak litter removed and placed onto a clean surface. Any invading vegetation or insects were carefully removed from the sample, using tweezers. The remaining oak litter was placed in an oven (80 8C) for between 12 h and 2 d. Litter samples were then cooled in a desiccator and weighed to ±0.001 g. Chemical analysis Bulked initial leaf litter from each site was finely ground, homogenized, and dried (105 8C). C and N concentrations were analysed on an Elementar Vario EL elemental analyser (Hanau, Germany). Samples were then digested in sulphuric acid / hydrogen peroxide (Allen 1989); thereafter Ca, Mn, and P concentrations were analysed on a Perkin Elmer 4300DV ICP-OES (Milan, Italy). Lignin concentration was determined by acid–detergent fibre digestion (cetyltrimethyl ammonium bromide (CTAB) detergent digest followed by 72% sulphuric acid digestion) (Rowland and Roberts 1994). Mass loss and limit value curve-fitting To calculate the initial rate of decomposition and the limit #
2007 NRC Canada
18
Table 1. Location of sites and basic site characteristics (from Williams et al. (2000)). Site name Abergwyngregyn oak 1 Aberwyngregyn oak 2 Bryn Brethynau Bryn Engan Camlyn Crowthers Coppice Cynfal Cymerau Dolgarrog Ganllwyd Coed Hafod Hafod Garregog, lower Hafod Garregog, upper Lletywalter Maentwrog Oak plantation Pendugwm Rheidol (grass) Rheidol (Vaccinium) Rhygen
Site code ABO1 ABO2 BB BE CAM CC CF CYM DG GL HF HGL HGU LL MT OP PEN RDG RDV RG
Northing 53.22 53.22 53.10 53.10 52.94 52.69 52.96 52.97 53.19 52.80 53.10 52.98 52.98 52.83 52.95 52.67 52.72 52.39 52.39 52.91
Westing 3.40 3.40 3.89 3.91 3.99 3.12 3.94 3.95 3.85 3.90 3.78 4.09 4.08 4.08 3.99 3.08 3.32 3.89 3.89 3.96
Altitude (m) 180 180 220 220 50 140 100 130 230 110 100 30 40 80 120 330 150 250 250 210
Soila BP BP BP BP BP BP I I I BP I I I I BP I I BP I I
Exchangeable acidity (meq. 100g–1 soil)b 3.26 2.87 1.65 1.34 2.29 2.38 0.60 0.55 0.28 2.50 2.19 2.09 0.24 1.20 2.16 1.21 3.6 2.83 3.28 1.58
Average tree height (m) 9 9 13 13 12 21 17 11 12 20 13 11 11 12 11 13 24 11 7 14
Tree density (no.ha–1) 667 667 320 453 550 489 250 689 433 333 533 350 400 1033 244 578 400 467 750 625
Annual litterfall (g m–2) c 10 10 17 24 143 328 204 94 157 319 247 53 189 253 43 437 441 38 70 305
a
BP, brown podzolic; I, integrade between brown podzolic and stagnopodzol (Avery 1990). Main mineral soil horizon (B). c Includes leaf litter and twigs but not large woody debris. b
2007 NRC Canada
Can. J. Bot. Vol. 85, 2007
#
Davey et al.
19
Table 2. Initial chemical composition of leaf litter (mgg–1 dry mass (DM)) for each oak woodland site. Concentration (mgg–1 DM) SITE AB01 AB02 BB BE CC CAM CF CYM DG GL HF HGL HGU LL MT OP PEN RDG RDV RG
Ca 6.1 6.2 5.3 6.6 9.5 8.5 6.0 7.2 11.0 9.3 10.0 5.5 7.1 8.8 4.1 8.9 6.7 6.9 7.2 7.0
Mn 2.0 2.4 1.8 1.6 1.5 0.81 0.36 0.93 0.87 2.2 1.4 2.3 1.6 3.1 0.4 3.1 1.1 1.4 2.3 1.6
N 10.6 12.0 12.2 10.3 10.8 8.7 9.8 12.7 11.2 11.1 9.3 10.5 10.3 11.6 9.5 17.8 13.0 12.8 10.0 12.0
P 0.73 0.87 0.43 0.58 0.44 0.46 0.52 1.2 0.65 0.39 0.66 0.61 0.46 0.68 0.41 1.1 0.53 0.49 0.53 0.64
C 520 517 525 519 510 517 525 520 516 519 514 519 519 514 526 519 518 525 525 523
Lignin 280 290 360 300 250 310 330 300 320 310 310 310 290 280 340 320 290 340 350 300
Note: Values are from bulked litter from each site.
values, we used the two-parameter single-exponential rise to maximum equation in Sigmaplot (version 8.0, Systat Software Inc., Calif.)). This is the same as the model for decomposition as described in Berg and Ekbohm (1991) ½1
ML ¼ mð1 ekt=m Þ
where ML is the accumulated mass loss (%) and t is the time in days. The maximum accumulated mass loss (limit value; %) is represented by m and the initial rate of decomposition is represented by k. The standard error and t and P values were calculated in Sigmaplot1 to indicate the goodness of fit for calculated k and m values. Soil C sequestration rate The soil C sequestration rates (g Cm–2year–1) were calculated using eq. 2: ½2
Soil C sequestration ¼ ð100 limit valueÞ=100 litterfall C fraction
where litterfall is given as grams per square metre and C fraction refers to initial leaf litter. Statistical analysis Pearson’s correlation coefficients (r) and linear regressions (r, R2 and R2adj) were calculated in SPSS version 12.0.1 (SPSS Inc., Chicago, Ill.).
Results Initial chemical composition of litter types Ca concentrations ranged between 4.1 and 11 mgg–1, with site MT having the lowest (4.1 mgg–1) and DG having
the highest (11 mgg–1) (Table 2). There was a wide range for the concentration of Mn (0.36–3.1 mgg–1) in the litter, with sites CF and MT having the lowest (0.36 and 0.4 mgg–1, respectively). The highest Mn concentration (3.1 mgg–1) was found in litter from sites LL and OP. N concentrations were rather uniform among sites, with most concentrations ranging between 10 and 12 mgg–1. Sites with low N concentrations were CAM, HF, and MT, with 8.7, 9.3, and 9.5 mgg–1, respectively. Litter from site OP possessed a very high N concentration at 17.8 mgg–1, about twice that of litter from the sites with the lowest N concentration. P concentrations, in most samples, ranged between 0.45 and 0.65 mgg–1. The litter from the GL and MT sites possessed the lowest concentrations of P (0.39 and 0.41 mgg–1, respectively). However, as with N, the P concentration of litter from sites OP and CYM (1.1 and 1.2 mgg–1, respectively) was almost three times the higher than that at sites with the lowest P concentration. C concentrations ranged between 510 and 526 mgg–1. The litter with the lowest concentration of C came from site CC (510 mgg–1) and the highest from MT (526 mgg–1). Lignin concentration for litter varied relatively little among sites, ranging from 250 mgg–1 at site CC to 360 mgg–1 at site BB. There were clear relationships among nutrient and lignin concentrations when values from all sites were crosscorrelated (Table 3). C concentrations covaried negatively with Ca but positively with lignin. N covaried positively with Mn and P concentrations. Initial mass-loss rates and limit values We analysed the initial decomposition rate using the estimated k value. This k value is specific for the used equation and cannot be compared to that estimated with, for example, the commonly used first-order kinetics function. Its value may be seen as an index for the initial decomposition rate. The equation performed well in providing the initial rate of mass loss and the estimated limit value (Table 4). The curve-fitting equation produced significant fits for initial mass-loss rates and limit values for all sites. The majority of sites had decomposition rates (k) between 0.0015 and 0.0020. Litter from sites CF and MT yielded the lowest decomposition rates, with k values at 0.0007 and 0.0010. Litter from sites OP and LL had the highest initial decomposition rates, with k values at 0.0029 and 0.0031. Limit values ranged from 56.9% to 95.1% loss of initial litter mass. The litter with the lowest limit values came from sites HGU and LL at 56.9% and 58.4%, respectively. The sites with the highest estimated limit value were MT, GL, and HGL at 91.5%, 94.4%, and 95.1%, respectively. There was one anomaly in that the limit value for litter from site CF was estimated to be >100%. This value was, therefore, omitted when limit values were correlated with initial nutrient and lignin concentrations and soil Csequestration rates. Correlations between initial nutrient and lignin concentrations with initial mass loss rate, and limit values The results indicate that the most significant correlation between initial nutrient values and initial mass-loss rates (k) was the initial Mn concentration of the litter, although rela#
2007 NRC Canada
20
Can. J. Bot. Vol. 85, 2007 Table 3. Pearson’s correlation coefficient (r) between the initial chemical components of leaf litter (mgg–1 DM) from each oak woodland site.
Ca Mn N C P Lignin
Ca Mn N C P — 0.160 — 0.106 0.468 * — –0.693 ** –0.238 0.024 — 0.128 0.303 0.599 ** –0.156 — –0.310 –0.179 0.022 0.780 ** –0.182
Lignin
—
Note: Values are from bulked litter of each site. n = 20. *, P ‡ 0.05, **, P ‡ 0.01, ***, P ‡ 0.001.
Table 4. Initial decomposition rate (k) and decomposition limit value (%) for leaf litter originating from 20 different oak woodland sites. Site AB01 AB02 BB BE CC CAM CF CYM DG GL HF HGL HGU LL MT OP PEN RDG RDV RG
Initial decomposition rate (k) 0.0016 0.0016 0.0015 0.0019 0.0024 0.0017 0.0007 0.0019 0.0016 0.0011 0.0017 0.0012 0.0023 0.0031 0.0010 0.0029 0.0015 0.0017 0.0016 0.0018
± SE 0.0003 0.0007 0.0004 0.0003 0.0005 0.0003 0.0003 0.0004 0.0004 0.0005 0.0005 0.0003 0.0009 0.0011 0.0004 0.0012 0.0004 0.0006 0.0006 0.0005
t 5.6181 2.2769 3.3198 6.5906 4.7020 4.9172 2.3733 4.4317 4.2684 2.4541 3.0465 3.6092 2.5798 2.7882 2.6368 2.4142 3.7308 2.6951 2.7914 3.9833
P
