Influence of soil type on mass loss and nutrient release from decomposing foliage litter of beech and Norway spruce

Influence of soil type on mass loss and nutrientrelease from decomposing foliage litter of beechand Norway spruce

Lars Vesterdal

Abstract: Mass loss and nutrient release from decomposing foliage litter of beech (Fagus sylvaticaL.) and Norwayspruce (Picea abies(L.) Karst.) were studied at three sites along a soil fertility gradient. The influence of soil type oninitial litter quality and on decomposition was separated by reciprocal transplantation of litter among soil types usingthe litterbag technique. Decomposition of beech litter was influenced by both initial litter quality and incubation site.Mass loss in beech litter was positively influenced by soil nutrient status. Decomposition of Norway spruce litter wasnot affected by initial litter quality, and the positive influence of a nutrient-rich soil environment on decomposition wasweak. Nutrient release in litters of both tree species was greatly affected by soil type through its influence on initiallitter quality, as nutrient release was positively related to initial nutrient concentrations. Nutrient release was lessaffected through the soil environment, as it only influenced release of some nutrients, and the differences were notconsistently related to soil nutrient status or mass loss. The influence of soil type on decomposition differed among thetwo tree species, suggesting that it may be more significant in species that produce relatively higher quality litter.

Résumé: La perte de masse et la libération des nutriments de litières de feuilles de hêtre commun (FagussylvaticaL.) et d’épicéa commun (Picea abies(L.) Karst.) en décomposition ont été étudiées à trois sites le long d’ungradient de fertilité du sol. L’influence du type de sol sur la qualité initiale de la litière et sur la décomposition a étéséparée par transplantation réciproque de litière entre les types de sol avec la technique des sachets de litière. Ladécomposition de la litière de hêtre était influencée par la qualité initiale de la litière et par le site d’incubation. Laperte de masse de la litière de hêtre était positivement influencée par le statut nutritif du sol. La décomposition de lalitière d’épicéa commun n’était pas affectée par la qualité initiale de la litière, et l’influence positive d’un sol à statutnutritif élevé sur la décomposition était faible. La libération des nutriments des litières des deux espèces étaitgrandement affectée par le type de sol via son influence sur la qualité initiale de la litière, étant donné que lalibération des nutriments était positivement reliée aux concentrations initiales en nutriments. La libération desnutriments était moins affectée par l’environnement sol, étant donné que ce dernier n’a affecté la libération que decertains nutriments et que les différences n’étaient pas reliées de façon cohérente au statut nutritif ou à la perte demasse du sol. L’influence du type de sol sur la décomposition différait selon l’espèce d’arbre, ce qui suggère qu’ilpourrait être plus significatif pour les espèces qui produisent des litières dont la qualité est relativement meilleure.

[Traduit par la rédaction] Vesterdal 105

Decomposition is an important process for cycling of nu-trients in forest ecosystems, and yet is itself partially con-trolled by site nutrient availability. Decomposition processesare influenced by macro- and micro-climate, litter quality,and activity of decomposing organisms. Differences in soilproperties were long ago noted as being important in deter-mining the accumulation and morphology of forest floors.Müller (1879) described the occurrence of mull and mor for-est floors at nutrient-rich and nutrient-poor soils, respec-tively, and attributed these features primarily to differences

in soil fauna communities. Later studies have suggested thatsoil nutrient status may also affect decomposition rates(Howard and Howard 1980; Boerner 1984; Staaf 1987). Soilproperties can affect forest floor decomposition ratesthrough influencing (i) initial litter quality, notably the con-centration of nutrients (Lukumbuzya et al. 1994; Nordén1994) and recalcitrant organic compounds like lignin andtannins (Flanagan and Van Cleve 1983; Sanger et al. 1996),and (ii ) the microenvironment in which litter decompositiontakes place, including the communities of soil fauna and mi-croorganisms. Soil fertility influences the species composi-tion and the biomass and activity of microflora andmicrofauna (Schaefer and Schauermann 1990; Raubuch andBeese 1995), and microbial activity may also be higher atnutrient-rich than at more nutrient-poor sites (Rastin 1994).

It is difficult to separate effects of soil properties in thefield, where the influence of soil environment can often beconfounded with effects of soil-induced litter quality, effectsof different tree species, or effects of different climate. Theextent to which soil properties influence litter decompositionhas therefore not been studied thoroughly. This study was

Can. J. For. Res.29: 95–105 (1999) © 1999 NRC Canada

95

Received December 5, 1997. Accepted November 10, 1998.

L. Vesterdal.1 Unit of Forestry, The Royal Veterinary andAgricultural University, Hørsholm Kongevej 11, DK-2970Hørsholm, Denmark.

1Present address: Department of Forest Ecology, DanishForest and Landscape Research Institute, Hørsholm Kongevej11, DK-2970 Hørsholm, Denmark. e-mail: [email protected]

I:\cjfr\cjfr29\cjfr-01\X98-182.vpTuesday, February 16, 1999 2:54:25 PM

Color profile: DisabledComposite Default screen

designed to separate the relative effects of soil properties onlitter decomposition rates and nutrient release into (i) influ-ence of soil-induced litter quality (litter origin) and (ii ) influ-ence of the soil environment (incubation site) within asimilar macroclimatic regime. This was done by reciprocaltransplantation and field incubation of foliage litter collectedat three different sites using the litterbag technique. It washypothesized that litter originating from a nutrient-rich sitewould decompose and release nutrients faster than litteroriginating from a nutrient-poor site and that litter incubatedat a nutrient-rich site would decompose and release nutrientsfaster than litter incubated at a nutrient-poor site.

SitesThe study was carried out in even-aged, 30-year-old adjacent

stands of beech (Fagus sylvaticaL.) and Norway spruce (Piceaabies(L.) Karst.) at three sites in Denmark with different soil nu-trient status (Table 1). The stands (approximately 0.25 ha) are partof a tree species experiment planted in 1964–1965 (Holmsgaardand Bang 1977). Stand densities differed slightly among sites, butthe effect of differences is expected to be negligible compared withthe effects of initial litter quality and soil type (Vesterdal et al.1995). The Christianssæde site was the most nutrient-rich with aMollic Hapludalf soil (Soil Survey Staff 1992) developed fromloamy and calcareous weichselian till, whereas the Ulborg site wasthe most nutrient-poor with a Typic Haplohumod soil developedfrom sandy till deposited during the Saale glaciation. TheLøvenholm site was considered of intermediate nutrient status witha Typic Haplumbrept soil developed from loamy sandy to sandyweichselian till. Forest floors were mor-like at Ulborg and mull-like at Christianssæde and Løvenholm, and data on forest floor nu-trient content at the three sites may be found in Vesterdal and

Raulund-Rasmussen (1998). Herb layers were absent in all sixstands, but mosses were present in the Norway spruce stands atLøvenholm and Ulborg. The 30-year mean annual temperatures are8.4°C at Christianssæde, 7.6°C at Løvenholm, and 7.5°C atUlborg. The mean annual precipitation during the study period was476 mm at Christianssæde, 498 mm at Løvenholm, and 715 mm atUlborg. The study period was much drier than indicated by the 30-year mean annual precipitation (Christianssæde, 614 mm;Løvenholm, 614 mm; Ulborg, 890 mm) (all climate data from TheDanish Meteorological Institute).

Litterbag experimentsFoliage litter was collected in October 1994 from each of the six

stands. Beech leaf litter was collected during litterfall by placing atarpaulin on the ground and shaking branches and trees. Norwayspruce neeedle litter was sampled from tarpaulins placed on theground for 3 weeks. Needles were sorted and green needles un-marked by senescence prior to abscission were rejected. Twograms dry weight (60°C for 48 h) of foliage litter was enclosed inpolyester litterbags measuring 13 × 15 cm and with a mesh size of1 × 0.75 mm. For each tree species, there were three qualities oflitter based on site of origin, and litter from all three sites of originwere incubated at each site. The litter from the three beech andNorway spruce stands were only incubated in stands of their re-spective tree species (i.e., litter of one species was not incubated instands of the other species). In early December 1994, litterbagswere placed on the litter layer in a randomized block design withten 1 × 1 m blocks randomly distributed in each stand. Each blockcontained five sets of the three litter qualities for sampling at fivedifferent dates. A total of 900 litterbags were used for the entirestudy (2 tree species × 3 incubation sites × 3 litter origins × 5 sam-pling dates × 10 blocks). Litterbags were fastened to the forestfloor by 10 cm long pins of high carbon steel and were collectedtwice a year in June and December for a period of 2.5 years.Litterbags were brought directly to the laboratory, where moss and

© 1999 NRC Canada

96 Can. J. For. Res. Vol. 29, 1999

Depth(cm)

Clay(%)

Silt(%)

Sand(%)

TotalC (%)

Total N(mg·g–1)

ExtractableP (mg·kg–1)*

Exchangeable (cmol+·kg–1)

Site pH Ca2+ Mg2+ K+

Christianssæde (Norway spruce)A1 0–5 3.8 12 10 78 2.8 2.1 110 4.6 0.6 0.12A2 5–25 5.2 12 11 77 1.5 1.7 110 9.3 0.4 0.10Bt 25–50 6.2 15 18 67 0.4 0.5 240 12.2 0.6 0.18Btg 50–73 7.5 18 20 62 380 21.7 0.5 0.12Ckg 73–110 7.7 9 16 75 0.3 0.07

Løvenholm (Norway spruce)A 0–27 4.1 3.5 12.0 85 0.9 0.7 119 0.47 0.06 0.06Bw 27–55 4.8 2.5 6.5 91 0.4 0.3 161 0.48 0.05 0.04BC 55–70 4.7 4.0 9.0 87 0.1 0.2 223 0.21 0.03 0.04Cg >70 4.5 3.5 10.5 86 0.1 123 0.78 0.07 0.07

Ulborg (adjacent Scots pine of the same age)A 0–18 2.7 2.1 3.0 95 10.7 3.3 22 0.39 0.67 0.23E 18–30 3.4 0.5 2.2 97 0.4 0.1 6 0.03 0.01 nd†Bh 30–34 3.5 10.9 3.8 85 7.0 2.4 24 0.20 0.13 0.09Bhs 34–40 4.1 6.2 3.6 90 3.6 1.3 20 0.06 0.04 0.04Bs 40–60 4.4 2.9 3.0 94 0.1 0.1 16 0.02 nd 0.01BC 60–100 4.5 1.3 2.1 97 10 0.01 nd ndC >100 4.6 0.9 2.5 97 11 0.01 nd nd

*Extracted with 0.1 M H2SO4.†Not detected (below detection limit).

Table 1. Physical and chemical characteristics of soils at the three sites (from Raulund-Rasmussen (1993), Raulund-Rasmussen andVejre (1995), and Vesterdal and Raulund-Rasmussen (1998)).



mineral soil were removed. The contents were dried at 60°C for48 h and weighed, and the 10 block replicates were subsequentlypooled to give one sample and were ground for chemical analysis.

Chemical analysesTotal C and N were measured by dry combustion in a Leco

CHN 1000 Analyzer. Total P, Ca, Mg, K, and Mn concentrationswere analysed following digestion of samples with concentratednitric acid in a microwave oven. Phosphorus in the diluted solutionwas measured by flow injection analysis, and the remaining nutri-ents were measured by atomic absorption spectroscopy. Klasonlignin, which also may include some tannin and cutin (Preston etal. 1997), was determined by first extracting water- and ethanol-extractable substances (Johansson et al. 1986) followed by hydro-lysis of the residue (12 M H2SO4 at room temperature for 2 h andrefluxing for 6 h after dilution to 0.358 M) and weighing of the re-sidual solid material. All chemical analyses were carried out in du-plicate or triplicate.

Nutrient release from litter was calculated as

[1] N N M N= − −in dl[( ) ]100

where N represents the amount of a nutrient released during de-composition (mg·g–1 incubated litter),Nin is the initial nutrient con-centration (mg·g–1), M is the mass loss (%), andNdl is the nutrientconcentration in decomposed litter (Entry et al. 1991).

StatisticsLitter mass loss within each tree species was analysed by three-

way ANOVA to determine significance of main effects (litter ori-gin, incubation site, and sampling date) and interaction effects overthe entire study period (Wieder and Lang 1982). Blocks werenested within incubation site and considered to be a random effect,whereas other effects were considered to be fixed. Percent massloss data were log transformed to normalize and homogenize vari-ances. Subsequently, the effect of litter origin and incubation sitewas analysed by sampling date using two-way ANOVA. In casesof significant main effects, comparisons between means were per-formed by Bonferroni’s multiple range test. All analyses were car-ried out with the procedure MIXED in SAS (SAS Institute Inc.1993). Differences between the two tree species in mass loss at thefive sampling dates were tested within incubation site and litterorigin with Student’st test (PROC TTEST) as there was no treespecies replication within sites. The effects on net nutrient releaseof different initial nutrient concentrations in litter (according to siteof origin) and of incubation site were tested using a general linearmodel (PROC GLM). The linear model included incubation site(class variable), initial nutrient concentration (quantitative vari-able), and the interaction term between these two variables. Incases of significant effects of incubation site, comparisons betweenmeans were performed by Bonferroni’s multiple range test.

Initial litter qualityThe initial quality of both beech and Norway spruce litter

differed among the three sites of origin (Table 2). Beech fo-liage litter from Christianssæde and Løvenholm was gener-ally more nutrient-rich than beech litter from Ulborg. Theonly exception was Mg, which was found in highest concen-trations in the litter from Ulborg. Beech litter from Løven-holm was considerably higher in P, K, and Mn than litterfrom Christianssæde. The amounts of water- and ethanol-extractable substances only differed marginally among thesites, but the Klason lignin concentration was lowest in thelitter from Løvenholm. The initial litter quality differences

© 1999 NRC Canada

Vesterdal 97

Sp

eci

es

an

dsi

teN (m

g·g

–1 )

P (mg

·g–

1 )C

a(m

g·g

–1 )

K (mg

·g–

1 )M

g(m

g·g

–1 )

Mn

(mg

·g–

1 )

Wa

ter-

ext

ract

ab

le(%

)

Eth

an

ol-

ext

ract

ab

le(%

)

Kla

son

lign

in(%

)

Bee

chC

hri

stia

nss

æd

e1

1.7

(0.2

)1

.19

(0.0

1)

17

.3(0

.3)

4.9

7(0

.04

)2

.09

(0.0

2)

0.4

1(0

.02

)1

0.2

(0.5

)5

.4(0

.4)

34

.3(0

.1)

Lø

ven

ho

lm1

2.7

(0.4

)2

.36

(0.0

7)

14

.2(0

.7)

7.0

5(0

.24

)1

.75

(0.0

5)

1.6

7(0

.01

)1

1.1

(0.9

)4

.7(0

.4)

31

.7(0

.1)

Ulb

org

10

.0(0

.2)

0.4

5(0

.02

)1

3.0

(0.7

)3

.02

(0.1

2)

2.2

5(0

.01

)0

.38

(0.0

0)

10

.3(0

.1)

4.5

(0.2

)3

5.5

(0.3

)

Nor

way

spru

ceC

hri

stia

nss

æd

e1

0.5

(0.2

)0

.93

(0.0

5)

16

.6(0

.2)

2.2

1(0

.02

)0

.95

(0.0

0)

0.8

9(0

.03

)7

.1(0

.4)

5.1

(0.2

)3

6.8

(0.3

)L

øve

nh

olm

9.4

(0.2

)1

.10

(0.0

3)

16

.8(0

.3)

3.1

0(0

.01

)1

.04

(0.0

0)

1.9

9(0

.08

)7

.8(0

.5)

4.7

(0.3

)3

5.1

(0.1

)U

lbo

rg1

2.1

(0.2

)0

.52

(0.0

0)

11

.8(0

.1)

2.8

2(0

.01

)1

.75

(0.0

1)

0.7

4(0

.04

)7

.2(0

.4)

4.5

(0.3

)3

7.6

(0.1

)

Not

e:V

alue

sar

em

eans

,w

ithS

Dgi

ven

inpa

rent

hese

sfo

rnu

trie

ntco

ncen

trat

ions

(n

=3)

and

for

extr

acta

ble

orga

nic

frac

tions

(n

=2)

.

Tabl

e2.

Initi

al

che

mis

try

of

be

ech

an

dsp

ruce

litte

rfr

om

the

thre

esi

tes.



were not quite as large in the Norway spruce litter, whichwas more nutrient-poor than beech litter except for Mn.Spruce litter from Ulborg was lower in P, Ca, and Mn, buthigher in N and Mg than litter from the other sites, and theconcentration of K was also higher in litter from Ulborgthan in litter from Christianssæde. Like the beech litter,spruce litter from Løvenholm was higher in P, K, and Mnthan litter from Christianssæde. The amounts of extractableorganic substances and lignin were similar for spruce litterfrom all three sites.

Mass lossBeech litter mass loss over the entire study period was

significantly affected by both incubation site and site of lit-ter origin, and there was no interaction effect (Table 3).There was a relatively large, significant interaction betweensampling date and incubation site compared with the maineffect of incubation site. The interaction between samplingdate and site of litter origin was relatively small comparedwith the main effect of site of litter origin. Norway sprucelitter mass loss was only weakly affected by incubation siteover the entire study (Table 3). Mass loss differences at eachsampling date are shown for both tree species by site of lit-ter origin and by incubation site in Fig. 1. As indicated by

the interaction effects for beech litter in Table 3, thedifferences in mass loss between litter origins were rela-tively consistent over the study, while the differences be-tween incubation sites were not. Mass loss was consistentlygreater for beech litter originating from Løvenholm than forlitter originating from Christianssæde, and except for the lastsampling date, beech litter originating from Christianssædedecomposed faster than litter originating from Ulborg

© 1999 NRC Canada

98 Can. J. For. Res. Vol. 29, 1999

Fig. 1. Remaining mass of beech and Norway spruce foliage litter of different site of origin (a and b) and incubated at different sites(c and d). Error bars are 1SE. At each sampling date, litter origins or incubation sites are ranked according to significant differences(p < 0.05) based on two-way analysis of variance and Bonferroni’s multiple range test (C, Christianssæde; L, Løvenholm; U, Ulborg).

Beech Norway spruce

df F p F p

Incubation site (I) 2 19.8 <0.0001 3.8 0.0346Litter origin (L) 2 137.5 <0.0001 1.7 0.1925Sampling date (S) 4 1282.0 <0.0001 1300.0 <0.0001I × L 4 1.3 0.2861 1.5 0.1937I × S 8 11.5 <0.0001 1.9 0.0556L × S 8 3.1 0.0023 0.5 0.8800I × L × S 16 0.9 0.6281 1.1 0.3426

Note: The F and p values are from a three-way analysis of variance.

Table 3. Effects of incubation site, litter origin, and samplingdate on mass loss in beech and Norway spruce litter through theentire period.



(Fig. 1a). Regarding incubation sites, beech litter decom-posed faster at Løvenholm than at Ulborg during the lastyear of incubation (Fig. 1c). At Christianssæde, beech littermass loss was intermediate during the last year. However, atthe end of the study, beech litter mass loss at Christianssædewas as great as at Løvenholm while decomposition atUlborg was slowing down. Norway spruce litter mass losswas completely unaffected by litter origin (Fig. 1b), and theeffect of incubation site was only significant during the firsthalf of the study (Fig. 1d) with the greatest mass loss re-corded at Christianssæde.

Mass loss differed between litters of the two tree specieswhen comparisons were made by the same site of litter ori-gin or by the same incubation site (Fig. 2). Litter originatingfrom Løvenholm and Ulborg showed the most consistenttree species differences in mass loss, whereas beech and

Norway spruce litter originating from Christianssædedecomposed at a very similar rate. Beech litter from Løven-holm had greater mass loss than Norway spruce litter fromthe same site, whereas the opposite result was found for lit-ter from Ulborg. Tree species differences by incubation siteshowed a similar picture. At Christianssæde the species dif-ference was inconsistent over the study period, but atLøvenholm, beech litter had greater mass loss than Norwayspruce after the first year. At Ulborg, Norway spruce litterhad greater mass loss than beech at the first sampling dateand after 2 years.

Nutrient releaseNet release of nutrients from beech and Norway spruce

litter was mainly influenced by the soil type through differ-ences in the initial nutrient concentrations (from Table 2) as

© 1999 NRC Canada

Vesterdal 99

Fig. 2. Remaining mass of beech and Norway spruce foliage litter of the same site of origin (left column) and incubated at the samesite (right column). Error bars are 1SE. Significant differences (p < 0.05) between tree species at the individual sampling dates basedon Student’st test are shown by a plus sign.



shown by the significant linear correlations in Figs. 3 and 4.For beech litter, net release of N, Mg, and Mn was also af-fected by incubation site (Fig. 3). There was no significantinteraction between initial concentrations of these nutrientsand incubation site; therefore, the best linear models had thesame slopes for the three incubation sites. Less N and Mgwere released at Ulborg than at the other two sites, whereasMn release was greatest at Christianssæde. Release of Mn,Mg, and Ca from spruce litter was also affected by incuba-tion site (Fig. 4). Manganese release from spruce litter wasgreatest at Ulborg, and Mg release was greater at Christians-sæde than at Ulborg. Calcium release from spruce litter was

affected by interaction between initial Ca concentration andincubation site (p < 0.01) as indicated by the slightly differ-ent slopes of regression lines. Calcium release increased inthe order Løvenholm < Ulborg < Christianssæde among in-cubation sites.

There was net immobilization of P in beech litter originat-ing from Ulborg (Fig. 3, initially 0.45 mg P·g–1 litter), andbeech litter incubated at Ulborg exhibited net immobiliza-tion of N (Fig. 3). Manganese was immobilized in bothbeech and spruce litter when the litter was initially low inMn (Figs. 3 and 4). C/N ratios for both beech and spruce lit-ter were characterized by a general decrease regardless of

© 1999 NRC Canada

100 Can. J. For. Res. Vol. 29, 1999

Fig. 3. Relationships between initial nutrient concentrations (mg·g–1; from Table 2) and net nutrient release (mg·g–1 incubated litter)from beech leaf litter during 2.5 years incubation. Differences in nutrient release among incubation sites are indicated by individualregression lines, which did not have significantly (p < 0.05) different slopes. Lines with the same letter are not significantly different(p > 0.05) based on a general linear model and Bonferroni’s multiple range test (incubation sites:d, Christianssæde;j, Løvenholm;m, Ulborg).



litter origin and incubation site (Figs. 5a and 5b). C/P ratiostended to converge toward 400–500 regardless of litter ori-gin or incubation site (Figs. 5c and 5d).

Site of litter origin and resultant litter qualityThe site of origin of the beech and spruce litter (Table 1)

determined initial litter quality (Table 2). Initial concentra-tions of P, Ca, K, and Mn in the litter appeared to be af-fected by soil nutrient concentrations (Table 1) as reportedelsewhere (Nicolai 1988; Nordén 1994). The influence of

specific soil properties on litter quality is as yet not wellunderstood, and the attributes quantified by soil pits are arelatively crude measure of soil nutrient status. However,beech and spruce litter from Ulborg had lower concentra-tions of P, Ca, and K (in beech only) than litter from theother two sites, and the soil at Ulborg has lower concentra-tions of these nutrients. The different Mn concentrations inlitter of both species at Løvenholm and Christianssæde mayreflect higher Mn availability to plants in acid soils as atLøvenholm (Tyler 1976). Higher P concentrations in bothlitters from Løvenholm than from Christianssæde despitecomparable amounts of soil P may also be a result of

© 1999 NRC Canada

Vesterdal 101

Fig. 4. Relationships between initial nutrient concentrations (mg·g–1; from Table 2) and net nutrient release (mg·g–1 litter) from Norwayspruce needle litter during 2.5 years incubation. Differences in nutrient release among incubation sites are indicated by individualregression lines. Only regression lines for Ca release had significantly different (p < 0.01) slopes. Lines with the same letter are notsignificantly different (p > 0.05) based on a general linear model and Bonferroni’s multiple range test (incubation sites:d,Christianssæde;j, Løvenholm;m, Ulborg).



different soil pH. Soil P might be more strongly bound atChristianssæde (higher pH) and may consequently be lessavailable to plants (Nihlgård and Lindgren 1977). The effectof soil Mg status on Mg concentrations in beech and sprucelitter was probably mediated by higher deposition of Mg atUlborg than at the other sites because of the proximity of theNorth Sea (Pedersen 1993). Klason lignin concentrationswere lowest in the beech litter from Løvenholm and highestin the beech litter from Ulborg. Sanger et al. (1996) reporteda similar but more pronounced effect of soil fertility onlignin concentrations in Scots pine (Pinus sylvestrisL.) nee-dle litter. The Norway spruce litters had very similar initialKlason lignin concentrations despite the differences in soilfertility, but the ranking of sites by lignin concentrations wasthe same as for beech litter.

Effect of site of litter origin on mass loss and nutrientrelease

The mass loss pattern reflected initial quality differencesin beech litter (Fig. 1a). Other studies have found clear(Staaf 1987), weak (Howard and Howard 1980) or nonexis-tent (Bocock and Gilbert 1957) effects of litter originatingfrom different soil types. Results may vary, depending on

comparability of soil gradients and the resultant litter qualitydifferences. Initial concentrations of N, P, K, and Mn in thebeech litters appeared to be important determinants of massloss over the study period. The first phases of decompositionhave been reported to be mainly controlled by concentra-tions of nutrients or readily available carbohydrates, whereaslater stages are controlled by concentrations of lignin (Berg1986; Taylor et al. 1989). Therefore, the effect of beechlitter origin on decomposition during 2.5 years might be par-tially attributed to differences in initial nutrient concentra-tions, but initial Klason lignin concentrations may also havebeen partly responsible for differences in later beech littermass loss (Rutigliano et al. 1996). Initial concentrations ofKlason lignin corresponded to the observed differences inmass loss, and a high initial lignin concentration can some-times retard decomposition (Melillo et al. 1982).

Norway spruce mass loss was remarkably unaffected bythe variation in initial nutrient concentrations (Fig. 1b). Itcould be that other litter properties controlling the initialstages of decomposition (e.g., solid carbohydrates and cellu-lose) did not differ among spruce litter originating from thethree sites. Alternatively, it could be that an even greatervariation in initial litter quality would be required to affect

© 1999 NRC Canada

102 Can. J. For. Res. Vol. 29, 1999

Fig. 5. C/N ratios for beech (a) and Norway spruce (b) and C/P ratios for beech (c) and Norway spruce (d) through 2.5 years ofdecomposition. Litter of three origins:d, Christianssæde;j, Løvenholm; andm, Ulborg and incubated at three sites: Christianssæde(solid line), Løvenholm (broken line), and Ulborg (dotted line). Open symbols indicate forest floor ratios at the three sites (Vesterdaland Raulund-Rasmussen (1998), Appendix 1).



mass loss in Norway spruce litter. For example, Lukum-buzya et al. (1994) also found that sugar maple leaf litter ofdifferent origin and with large differences in nutrient con-centrations unexpectedly had similar mass loss. This ledthem to question the magnitude of litter quality differencesrequired to result in mass loss differences within the samespecies and whether litters of the same species always de-compose at similar rates in the same environment. Mass lossresults in the present study suggest that the required magni-tude of initial litter quality difference varies between treespecies, and the results for beech indicate that initial litterquality differences may affect decomposition rates regard-less of the environment (i.e., the incubation site).

Net nutrient release over 2.5 years from beech and Nor-way spruce litter was strongly affected by initial nutrientconcentrations, which were determined by site of litter ori-gin (Figs. 3 and 4). For Norway spruce this occurred in spiteof similar mass loss for litters of different origin (Fig. 1), in-dicating that nutrient cycling in spruce is not solely deter-mined by physical breakdown. For beech litter, however,nutrient release and mass loss differences were more closelyrelated. There were large differences in initial P concentra-tions in beech litter originating from different sites, and Pwas immobilized on a net basis in the P-poor beech litterfrom Ulborg (Fig. 3, initial P concentration of 0.45 mg·g–1),probably because of selective import from the surroundings(Staaf and Berg 1982). The differences in P release amonglitter originating from different sites resulted in a character-istic converging trend in C/P ratios for both beech and Nor-way spruce litter during decomposition (Figs. 5c and 5d).Depending on the initial P concentration, P was either im-mobilized or released over time relative to C, resulting inmore similar C/P ratios for all three litter origins after 2.5years. Similar patterns of P immobilization or release havealso been reported from studies with other species (Blair1988; Prescott et al. 1993). Gosz et al. (1973) suggested acritical C/P ratio for decomposer organisms within the range360–480, which might be close to the convergence value inthis study. C/N ratios decreased for litter originating from allthree sites, suggesting that N was limiting decomposition re-gardless of initial concentrations. Both initial C/N and C/Pratios are important determinants of decomposition rates(Enriquez et al. 1993), but the microbial demand for addi-tional P to decompose beech litter originating from Ulborgmay partly explain why litter from this site also had thesmallest mass loss.

Effect of incubation site on mass loss and nutrientrelease

Mass loss was affected by incubation site in both tree spe-cies, but the effect was less consistent and less demonstrableover the incubation period for Norway spruce litter than forbeech litter (Figs. 1c and 1d). As climatic differences amongsites were small, these results indicate that decompositionwas mainly affected through the environment as defined bysite and soil properties. As expected, beech litter mass losswas lowest at the most nutrient-poor Ulborg site, but massloss was highest at Løvenholm and only intermediate at thesite most rich in exchangeable base cations, Christianssæde.For Norway spruce litter (Fig. 1d), mass loss tended to begreater at the rich Christianssæde site than at the poor

Ulborg site. A greater effect on Norway spruce mass lossmight have been found if the study had included the laterstages of decomposition. Much greater forest floor carboncontents in both beech and Norway spruce stands at Ulborgthan at the other sites (Vesterdal and Raulund-Rasmussen1998) indicate that differences in decomposition rate forNorway spruce would develop at some point. Few litterbagstudies have related mass loss within stands of the same spe-cies and in the same climatic region to the nutrient status ofsoils. Staaf (1987) found that mass loss in beech leaf litterwas positively related to nutrient status of soils and attrib-uted this to different conditions for microbial and micro-faunal turnover. The incubation sites in this study are likelyto differ in communities of soil fauna and microorganisms.Decomposer biomass and species diversity have been re-ported to be highest in nutrient-rich soil environments withhigh pH and large amounts of exchangeable bases (Schaeferand Schauermann 1990; Raubuch and Beese 1995), and alow species diversity may result in slow decomposition(Setälä et al. 1988). These effects of soil type on litter massloss may partially explain the slower decomposition atUlborg. Based on soil properties, the highest mass loss wasexpected at Christianssæde, and incubation site differencesfor Norway spruce to some extent supported the hypothesisthat mass loss would be greatest at the most nutrient-richsite (Fig. 1d). However, beech incubated at Løvenholm (anintermediate site) decomposed faster than at Christianssæde.The fast beech litter mass loss at Løvenholm could be due toa positive effect of the nutrient-rich native litter on microbialactivity in the incubated litter.

Effects of incubation site on mass loss did not interactwith effects of litter origin on mass loss as found in anotherstudy (Nicolai 1988). An incubation site × litter origin effectmay have been prevented by the exclusion of macrofaunaspecies from litter at the nutrient-rich sites. It must be em-phasized that results are relative estimates of decompositionunder standardized conditions, i.e., under influence of thesame size fraction of decomposer communities.

The differences in nutrient release among incubation siteswere not just a product of mass loss, and other factors thansoil nutrient status may be more involved in the release ofsome nutrients from litter. For instance, Mn release was sig-nificantly highest at Christianssæde for beech and signifi-cantly highest at Ulborg for Norway spruce. These releasedifferences could be due to a greater selective demand forMn among soil microorganisms such as lignin-degradingwhite-rot fungi (Perez and Jeffries 1992) at the two siteswith lower Mn availability than Løvenholm. Regarding theapparently lower Mg release at Ulborg, a higher input of Mgfrom the sea (Pedersen 1993) could be maintaining Mg con-centrations in decomposing litter at a higher level. Nitrogenrelease from litter at Christianssæde and Løvenholm wasgenerally low, and N was immobilized in beech litter atUlborg. This was presumably because of a higher microbialdemand for N in beech at Ulborg, as beech forest floors alsohad a lower C/N ratio at Ulborg (22) than at Christianssæde(28) and Løvenholm (27) (Vesterdal and Raulund-Rasmussen 1998, Appendix 1). Prescott et al. (1993) foundno influence of N and P availability at incubation sites on Nand P release. This agrees with the results for P release fromthis study. Although initial P concentrations in litter indicate

© 1999 NRC Canada

Vesterdal 103



great differences in P availability among sites, the releasedamounts of P were similar at all sites (Figs. 3 and 4). Thechanges in C/N and C/P ratios in decomposing litter overtime also suggest that incubation site did not influence Nand P dynamics as much as the initial N and P status of thelitter. However, C/P ratios for beech and especially for Nor-way spruce tended to converge more slowly in the P-poorlitter from Ulborg when it was incubated at its site of origin(Figs. 5c and 5d). C/P ratios in forest floor bulk materialfrom Ulborg were still lower, indicating further narrowing inC/P ratios. The slower convergence of C/P ratios at Ulborgmay also point at low P availability as a possible reason forslower mass loss at this site.

Nutrient release was less affected by soil type through in-cubation site than through litter origin in both tree species,i.e., initial nutrients in the litter were more intimately in-volved in nutrient release than nutrients in forest floor orsoil. This is not surprising, given the variation in initial nu-trient concentrations among litter origins. However, it is pos-sible that the influence of soil-induced initial litter quality onnutrient release decreases during later stages of decomposi-tion, whereas the importance of the soil environment mayincrease (McClaugherty et al. 1985).

Differences in soil type effect between tree speciesThe experiment with beech leaf litter supported the

hypothesis that soil type influences decomposition boththrough litter origin and incubation site, while decomposi-tion of Norway spruce needle litter only supported this hy-pothesis to some extent as regards effect of incubation site(Figs. 1a–1d). Effects of litter origin and incubation sitewere not equally pronounced for beech and Norway spruce,suggesting that soil type influences decomposition in differ-ent ways and to a different extent among tree species. As aresult of the fairly constant Norway spruce mass loss, beechdecomposed faster than Norway spruce when the litter wascollected or incubated at Løvenholm, whereas Norwayspruce decomposed fastest when litter was collected or incu-bated at Ulborg (Fig. 2). This result must be tempered by thelimitation that tree species plots were not replicated withinsites, but interactions between effects of tree species and ef-fects of soil type have also been reported from other studies(Bocock and Gilbert 1957; Bocock et al. 1960). These inter-actions all indicate that decomposition is controlled by what-ever factor is most critical for decomposer organisms. If thelitter is of low quality because of its tree species origin(e.g., Norway spruce), decomposition is less affected by soilfertility than in more favourable litter types (e.g., beech).Analogous to temperature and moisture interactions, thesoil-induced nutrient status of litter or the soil micro-environment may be of little significance if specific tree spe-cies related litter properties set the quality low for de-composer organisms.

The influence of soil type on decomposition differed be-tween the two tree species. The experiment with beech leaflitter supported the hypothesis that soil type influences de-composition both through litter quality and the incubationenvironment, while the experiment with Norway spruce nee-

dle litter only supported the hypothesis as regards effect ofincubation environment. Beech litter mass loss was posi-tively affected by nutrient status of the soil, whereas the ef-fect on mass loss was weak for Norway spruce litter.Nutrient release from litter of both species was greatly af-fected by soil type through the resultant litter quality, as thereleased amounts were positively related to initial nutrientconcentrations. Nutrient release was less affected by the soilenvironment, and the differences were not clearly related tonutrient status of soils or mass loss. The study indicated thatsoil types influence decomposition in different ways and to adifferent extent among tree species. The influence of soiltype may be more significant in species that produce rela-tively higher quality litter.

The Danish Forest and Landscape Research Institutekindly allowed me to use the tree species trials for the ex-periment. I thank Lena B. Troelsen for carrying out the nu-trient analyses, and Peter Theis Hansen and Jens Holst-Nielsen for assistance with preparation of litterbags.

Berg, B. 1986. Nutrient release from litter and humus in coniferousforest soils—a mini review. Scand. J. For. Res.1: 359–369.

Blair, J.M. 1988. Nitrogen, sulfur and phosphorus dynamics in de-composing deciduous leaf litter in the Southern Appalachians.Soil Biol. Biochem.20: 693–701.

Bocock, K.L., and Gilbert, O.J.W. 1957. The disappearance of leaflitter under different woodland conditions. Plant Soil,9:179–185.

Bocock, K.L., Gilbert, O., Capstick, C.K., Twinn, D.C., Waid, J.S.,and Woodman, M.J. 1960. Changes in leaf litter when placed onthe surface of soils with contrasting humus types. I. Losses indry weight of oak and ash leaf litter. J. Soil Sci.11: 1–9.

Boerner, R.E.J. 1984. Nutrient fluxes in litterfall and decomposi-tion in four forests along a gradient of soil fertility in southernOhio. Can. J. For. Res.14: 794–802.

Enriquez, S., Duarte, C.M., and SandJensen, K. 1993. Patterns indecomposition rates among photosynthetic organisms: the im-portance of detritus C:N:P content. Oecologia,94: 457–471.

Entry, J.A., Rose, C.L., and Cromack, K. 1991. Litter decomposi-tion and nutrient release in ectomycorrhizal mat soils of aDouglas fir ecosystem. Soil Biol. Biochem.23: 285–290.

Flanagan, P.W., and Van Cleve, K. 1983. Nutrient cycling in rela-tion to decomposition and organicmatter quality in taiga ecosys-tems. Can. J. For. Res.13: 795–817.

Gosz, J.R., Likens, G.E., and Bormann, F.H. 1973. Nutrient releasefrom decomposing leaf and branch litter in the Hubbard BrookForest, New Hampshire. Ecol. Monogr.43: 173–191.

Holmsgaard, E., and Bang, C. 1977. Et træartsforsøg mednåletræer, bøg og eg. De første 10 år. [In Danish.] Forstl.Forsøgsvæ. Dan.35: 159–196.

Howard, D.M., and Howard, P.J.A. 1980. Effect of species, sourceof litter, type of soil, and climate on litter decomposition. Mi-crobial decomposition of tree and shrub leaf litter 3. Oikos,34:115–124.

Johansson, M.B., Kögel, I., and Zech, W. 1986. Changes in thelignin fraction of spruce and pine needle litter during decompo-sition as studied by some chemical methods. Soil Biol.Biochem.18: 611–619.

© 1999 NRC Canada

104 Can. J. For. Res. Vol. 29, 1999



© 1999 NRC Canada

Vesterdal 105

Lukumbuzya, T.K., Fyles, J.W., and Côté, B. 1994. Effects of base-cation fertilization on litter decomposition in a sugar maple for-est in southern Quebec. Can. J. For. Res.24: 447–452.

McClaugherty, C.A., Pastor, J., Aber, J.D., and Melillo, J.M. 1985.Forest litter decomposition in relation to soil nitrogen dynamicsand litter quality. Ecology,66: 266–275.

Melillo, J.M., Aber, J.D., and Muratore, J.F. 1982. Nitrogen andlignin control of hardwood leaf litter decomposition dynamics.Ecology,63: 621–626.

Müller, P.E. 1879. Studier over Skovjord som Bidrag tilSkovdyrkningens Theori. I. Om Bøgemuld og Bøgemor paaSand og Ler. [In Danish.] Tidsskr. Skovbrug,3: 1–125.

Nicolai, V. 1988. Phenolic and mineral content of leaves influencesdecomposition in European forest ecosystems. Oecologia,75:575–579.

Nihlgård, B., and Lindgren, L. 1977. Plant biomass, primary pro-duction and bioelements of three mature beech forests in SouthSweden. Oikos,28: 95–104.

Nordén, U. 1994. Leaf litterfall concentrations and fluxes of ele-ments in deciduous tree species. Scand. J. For. Res.9: 9–16.

Pedersen, L.B. 1993. Nutrient cycling in Sitka spruce, Norwayspruce and beech stands in Denmark. [In Danish with Englishabstract and summary.] Forskningsserien nr.1-1993. Danish For-est and Landscape Research Institute, Lyngby.

Perez, J., and Jeffries, T.W. 1992. Roles of manganese and organicacid chelators in regulating lignin degradation and biosynthesisof peroxidases byPhanerochaete chrysosporium.Appl. Environ.Microbiol. 58: 2402–2409.

Prescott, C.E., Taylor, B.R., Parsons, W.F.J., Durall, D.M., andParkinson, D. 1993. Nutrient release from decomposing litter inRocky Mountain coniferous forests: influence of nutrient avail-ability. Can. J. For. Res.23: 1576–1586.

Preston, C.M., Trofymow, J.A., Sayer, B.G., and Niu, J. 1997.13C nuclear magnetic resonance spectroscopy with cross-polarization and magic-angle spinning investigation of theproximate-analysis fractions used to assess litter quality in de-composition studies. Can. J. Bot.75: 1601–1613.

Rastin, N. 1994. Biochemischer und mikrobiologischer Zustandverschiedener Waldböden. J.D. Sauerländer’s Verlag, Frankfurtam Main.

Raubuch, M., and Beese, F. 1995. Pattern of microbial indicators inforest soils along an European transect. Biol. Fertil. Soils,19:362–368.

Raulund-Rasmussen, K. 1993. Pedological and mineralogical char-

acterization of five Danish forest soils. For. Landscape Res.1:9–33.

Raulund-Rasmussen, K., and Vejre, H. 1995. Effect of tree speciesand soil properties on nutrientimmobilization in the forest floor.Plant Soil,168–169: 345–352.

Rutigliano, F.A., Virzo De Santo, A., Berg, B., Alfani, A., andFioretto, A. 1996. Lignin decomposition in decaying leaves ofFagus sylvaticaL., and needles ofAbies albaMill. Soil Biol.Biochem.28: 101–106.

Sanger, L.J., Cox, P., Splatt, P., Whelan, M.J., and Anderson, J.M.1996. Variability in the quality ofPinus sylvestrisneedles andlitter from sites with different soil characteristics: lignin andphenylpropanoid signature. Soil Biol. Biochem.28: 829–835.

SAS Institute Inc. 1993. SAS/STAT user’s guide, release 6.07 edi-tion. SAS Institute Inc., Cary, N.C.

Schaefer, M., and Schauermann, J. 1990. The soil fauna of beechforests: comparison between a mull and a moder soil.Pedobiologia,34: 299–314.

Setälä, H., Haimi, J., and Huhta, V. 1988. A microcosm study onthe respiration and weight loss in birch litter and raw humus asinfluenced by soil fauna. Biol. Fertil. Soils5: 282–287.

Soil Survey Staff. 1992. Keys to soil taxonomy. 5th ed. SMSSTech. Monogr. No. 19. Pocahontas Press, Blacksburg, Va.

Staaf, H. 1987. Foliage litter turnover and earthworm populationsin three beech forests of contrasting soil and vegetation types.Oecologia,72: 58–64.

Staaf, H., and Berg, B. 1982. Accumulation and release of plantnutrients in decomposing Scots pine needle litter. Long-term de-composition in a Scots pine forest II. Can. J. Bot.60:1561–1568.

Taylor, B.R., Parkinson, D., and Parsons, W.F.J. 1989. Nitrogenand lignin content as predictors of litter decay rates: a micro-cosm test. Ecology,70: 97–104.

Tyler, G. 1976. Soil factors controlling the metal ion absorption inthe wood anemoneAnemone nemorosa. Oikos, 27: 71–80.

Vesterdal, L., and Raulund-Rasmussen, K. 1998. Forest floorchemistry under seven tree species along a soil fertility gradient.Can. J. For. Res.28: 1636–1647.

Vesterdal, L., Dalsgaard, M., Felby, C., Raulund-Rasmussen, K.,and Jørgensen, B.B. 1995. Effects of thinning and soil propertieson accumulation of carbon, nitrogen and phosphorus in the for-est floor of Norway spruce stands. For. Ecol. Manage.77: 1–10.

Wieder, R.K., and Lang, G.E. 1982. A critique of the analyticalmethods used in examining decomposition data obtained fromlitter bags. Ecology,63: 1636–1642.



Documents

Influence of soil type on mass loss and nutrient release from decomposing foliage litter of beech and Norway spruce