lable at ScienceDirect

Soil Biology & Biochemistry 43 (2011) 1300e1307

Contents lists avai

Soil Biology & Biochemistry

journal homepage: www.elsevier .com/locate/soi lbio

No UV enhancement of litter decomposition observed on dry samples undercontrolled laboratory conditions

Miko U.F. Kirschbaum a,*, Suzanne M. Lambie a, Hui Zhou b

a Landcare Research, Private Bag 11052, Palmerston North 4442, New Zealandb State Key Laboratory of Environmental Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences, Guiyang 550002, China

a r t i c l e i n f o

Article history:Received 21 December 2010Received in revised form28 February 2011Accepted 2 March 2011

Keywords:BleachingDecompositionEnvironmental driversMicrobial degradationPhotooxidationUV radiation

* Corresponding author. Tel.: þ64 6 353 4902; fax:E-mail address: KirschbaumM@LandcareResearch.

0038-0717/$ e see front matter � 2011 Elsevier Ltd.doi:10.1016/j.soilbio.2011.03.001

a b s t r a c t

In field studies, various workers have observed a stimulation of organic matter breakdown by visiblelight and UV radiation. We aimed to confirm the involvement of UV radiation under controlled laboratoryconditions and quantify the magnitude of any stimulation. Grass and pine foliage samples were oven-dried and continuously exposed to UV radiation at room temperature for up to 60 days. A range of UVflux densities was established using shading to different levels. After UV exposure under air-dryconditions, samples were rewetted and incubated in the dark with microbial inoculums to investigatewhether UV exposure had rendered samples more susceptible to subsequent microbial decomposition.

However, we found no weight loss associated with different UV flux densities. The same finding heldtrue for grass and pine litter samples. Similarly, microbial decomposition of either grass or pine litter wasnot enhanced by prior UV exposure. These findings suggest that UV-induced photooxidation of drymaterials cannot be responsible for the observed apparent enhancement of weight loss of litter samplesunder UV exposure in the field.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

It has long been known that microbial decomposition of organicmatter responds strongly to temperature (e.g. Kirschbaum, 2000,2010) and soil or litter moisture contents (Borken and Matzner,2009). Global warming is likely to stimulate organic matterdecomposition and lead to a loss of soil carbon (e.g. Sitch et al.,2008), and the extent of that stimulation will critically affect thefuture natural biogenic contribution to net CO2 emissions to theatmosphere (Sitch et al., 2008; Kirschbaum, 2010).

Over recent years, a number of workers have shown, however,that organic matter decomposition can be affected not only by theknown biological drivers but can also be enhanced through expo-sure to visible and UV (UV-A and UV-B) radiation (Moorhead andReynolds, 1989; Anesio et al., 1999; Schade et al., 1999; Day et al.,2007; Austin and Vivanco, 2006; Rutledge et al., 2010; Brandtet al., 2010). To the extent that decomposition is controlled byabiotic processes such as photooxidation, it will reduce its depen-dence on biotic drivers. Systems would then be less responsive tochanges in the key controllers of microbial decomposition, such astemperature.

þ+64 6 353 4801.co.nz (M.U.F. Kirschbaum).

All rights reserved.

Most notably, Austin and Vivanco (2006) found that the stim-ulation of decomposition by radiation occurred in the absence ofmicrobial activity. Their observed increase of decomposition underradiation exposure must therefore have been due to direct photo-oxidation rather than through microbial facilitation, which is thebreakdown of complex organic compounds into simpler ones thatcan be degraded more easily by microbial enzymes at some timeafter UV or light exposure. They observed the strongest decompo-sition rates when they allowed all wavelengths to reach theirsamples, including UV and photosynthetically active radiation.

UV-B is believed to be particularly effective at breaking downlignins (Gehrke et al., 1995; Lanzalunga and Bietti, 2000; Henryet al., 2008), which are resistant to breakdown by most micro-organisms. Photochemical degradation of cellulose may also bepossible through visible light although photooxidation appears toincrease sharply with decreasing wavelength below about 500 nm(Schade et al., 1999; Brandt et al., 2009). We are not aware of anyother attempt at generating an action spectrum of litter decom-position effects.

Further compelling evidence for direct photooxidation to playa role in litter weight loss has come from a recent study by Rutledgeet al. (2010) who showed that CO2 emissions from peat samplesresponded almost instantaneously to changes in radiation. Theirexposed samples were air-dry during the experiment whicheffectively eliminated microbial CO2, and there was no CO2 release

280 300 320 340 360

280 300 320 340 360

0

20

40

60

80

100

120

Wavelength (nm)

Irra

dia

nc

e (m

W m

nm

)

1

10

100

1000

Solar

UV lamps

UV-B

a

b

Fig. 1. Irradiance received by unshaded samples in the experiment compared toirradiance received at noon on a typical New Zealand summer day. Data are shown oneither a linear (a) or logarithmic (b) scale. The shaded area shows the wavelengthrange usually designated as UV-B. Solar spectrum after McKenzie et al. (2009).

M.U.F. Kirschbaum et al. / Soil Biology & Biochemistry 43 (2011) 1300e1307 1301

when samples were darkened. This work indicated not only thatradiation played an important role for total CO2 release, but alsothat the mechanism at least included direct photooxidation ratherthan relying solely on microbial facilitation.

There are fewer reports of the effect of radiation on decomposi-tion under controlled laboratory conditions. Such studies undercontrolled conditions are important to not only confirm theapparent observations from the field, but to also better characterisethe relevant action spectra, determine dose responses and identifyto what extent an overall radiation effect is caused by directphotooxidation or microbial facilitation. Foereid et al. (2010) keptdried litter samples under broad-spectrum radiation sourcesincluding UV radiation for up to 289 days, and found no apparentweight loss with time. They did observe, however, that samplesexposed to radiation for longer periods showed faster subsequentmicrobial degradation when samples had been rewetted. Theyconcluded that microbial facilitation rather than direct photo-oxidative mass loss must have been responsible for any weight lossobserved in the field.

Brandt et al. (2009) exposed different litter samples to UV radi-ation in the laboratory and found a clear enhancement of CO2 effluxrates under UV exposure. However, their observed enhancementwas very small, amounting to aweight loss of less than 0.5% over 70days of continuous exposure, and there was no evidence formicrobial facilitation by UV exposure. While Brandt et al. (2009)showed that litter degradation can be enhanced by UV exposure,their observed rates were too small to account for the largeenhancement of decomposition observed in field experiments.

We conducted a laboratory experiment under controlledconditions to try and further quantify any effect of UV exposure onlitter decomposition and separately assess any effects on directphotooxidation and microbial facilitation.

2. Materials and methods

In our experiment, we investigated the effect of UV radiation onPinus radiata needles and perennial ryegrass (Lolium perennecv Nui). First, we investigated the effect of intensity of UV exposureon grass and pine needle degradation by observing any directweight loss. A range of UV intensities was generated either througha set of wire meshes or by using different amounts of grass litterthrough self-shading. After the end of UV exposure, samples weremoistened and incubated in the dark to assess the extent of anymicrobial facilitation by prior UV exposure.

2.1. Litter UV exposure

A metal frame was erected over a metal bench top to house 6fluorescent UV lamps (Phillips TL 40W/12RS). The bench top wascovered with black cloth to stop back-reflection of the UV radiationand ensure that samples received radiation only from the UVradiation sources. Radiation received by our samples was measuredwith a UV-B BiometerModel 501 Radiometer (Solar Light Company,Pennsylvania, USA). The biometer was calibrated, and the spectraloutput of the UV lamps was measured with a Bentham spectror-adiometer with DM150BC double monochromator, cosine diffuserand end window PMT detector (Bentham, Reading, UK).

Unshaded samples in our experiment received irradiancecomparable to that received at noon in summer under typical NewZealand conditions, especially in the biologically more active lowerwavelength range below about 310 nm (Fig. 1). The flux density ofsolar irradiance, on the other hand, was higher at wavelengthsgreater than 310 nm. Integrating over the whole UV-B range, solarnoon irradiance in New Zealand is about 2.0Wm�2 forwavelengthsfrom 290 to 315 nm and 3.7 Wm�2 from 290 to 320 nm (McKenzie

et al., 2009). This compares with irradiance received by our samplesof 1.9 Wm�2 up to 315 and 2.2 Wm�2 up to 320 nm. The UV lampsalso emitted about 0.27 W m�2 in the UV-C range below 290 nm(calculated from the data shown in Fig. 1).

As samples were exposed to that radiation continuously for 60days, the received UV radiation load was therefore comparable tothat received by litter under field conditions over a whole summer.In addition, our experimental lamps produced a much greaterproportion of shorter-wavelength radiation than solar radiation sothat the radiation under our experimental conditions was likely topromote litter breakdown even more than natural sunlight.

The top of the lighting frame was covered with fine-meshedmaterial to prevent dust collecting on top of the litter samples,which could have affected their weight during the experiment. Thesides of the lighting frame were covered with reflective aluminiumfoil to backscatter radiation from the walls and create a more evenradiation field for our samples. It also prevented UV light fromescaping the UV exposure area as a safety precaution for staffworking in the area.

Fresh pine needles were collected from 20-year-old P. radiatatrees (Old West Road, Palmerston North, New Zealand; Roger Par-fitt, pers. comm.) and oven-dried at 80 �C. Basal sheaths wereremoved from each fascicle and discarded, and needles were cutinto approximately 2 cm lengths. Nui Ryegrass was grown undercontrolled conditions in a shade house, and irrigated daily. Thegrass was harvested periodically to maintain a short and vigoroussward. The harvested grass was dried at 80 �C and stored at roomtemperature until the start of the experiment. The grass was sortedto remove dead material and cut into approximately 2 cm lengths.

Both pine and ryegrass litter were exposed to six UV radiationlevels of 1.4%, 18%, 41%, 60%, 73%, and 100% of incident UV radiation,with 5 replicates each. The level of UV exposure was controlled bya range of metal screens placed over individual litter samples.Frames were constructed of medium density fibre board withmetalscreens of the appropriate aperture size attached. Screens for the100% treatment consisted of a frame with no mesh attached.

M.U.F. Kirschbaum et al. / Soil Biology & Biochemistry 43 (2011) 1300e13071302

Pre-cut pine and grass material was heated to 80 �C, weighedafter cooling to room temperature and placed into 10 � 10 cmpolystyrene petri dishes (Labserv, BioLab, Auckland, New Zealand).We used approximately 2.5 g of pine litter and 1.0 g of grass litter.These amounts covered the area of the petri dishes with minimalself-shading. Accurate weights of each sample were recordedbefore the start of the experiment. Direct weight loss during theincubation was assessed by periodic weighing of samples. Acomplication arose in that litter samples and their trays (weighingabout 20 g) were found to rapidly adsorb atmospheric moistureafter having been dried in an oven (Fig. 2). To minimise moistureeffects, all weight measurements of litter material were madewithin 5 s of removal from the oven. Despite these precautions,measured weights appeared to change by about �1% betweenmeasurement periods in line with changes in atmospheric mois-ture (see also Dirks et al., 2010). Those differences did not affect therelativities between samples measured on the same day, as allsamples were treated in the same way so that relativities betweendifferent treatments should reflect true treatment differences, butthey added to the residual variability between measurements.

The effect of self-shading in a ryegrass litter layer was investi-gated by using a series of litter weights of 0.5, 1.0, 1.5, 2.0, 2.5, 3.0,4.0, and 5.0 g, with 5 replicates. The largest amount of litter filledthe petri dishes to their rim. 5 g of litter provided an effective screento material at the bottom of the petri dishes, as evidenced bymaterial at the bottom remaining slightly green even after 60 daysof exposure, whereas less shaded material was strongly bleached.

Litter samples from both the exposure and self-shading exper-iments were randomly assigned into 5 blocks of 20 samples each. Aseparate 6th block consisted of 10 pine and 10 grass samples, ofwhich 2 replicates were removed and weighed every 10 days tofollow weight changes over time. These samples showed noconsistent weight changes over time and are not further reported inthe following.

The litter samples were exposed to UV radiation continuouslyover 60 days. Samples were shifted to a different block every tendays, and the position of each sample was randomised within eachblock to ensure that each sample (subject to their individualshading treatment) received the same total UV exposure over thelength of the experiment. When samples were removed fromexposure, their air-dry weight was recorded, they were then driedto 80 �C to assess whether their moisture adsorption properties hadbeen affected by UV exposure. Also, their oven-dry weights wererecorded to assess whether there was any weight loss due to UVexposure. The percentage of moisture adsorption showed no trendswith UV exposure and is not further reported in the following.

0 60 120 180 240

0

10

20

30

Time (s)

Weig

ht g

ain

(m

g)

Trays with

samplesTrays only

Fig. 2. Example of observed weight change of litter samples after removal from theoven. This weight change was thought to reflect moisture adsorption in rapidly coolingsamples and their plastic trays. Trays weighed approximately 20 g and samples 1 g.

Results reported in the following are based on the weightmeasurements of oven-dried samples.

2.2. Chemical analysis

Samples were chemically analysed using the acid detergentfibreesulphuric acid procedure (Rowland and Roberts, 1994),involving sequential treatment of the sample with differentreagents to destroy various fractions, followed by gravimetricdetermination of the residues. A sub-sample of 0.5 g of the exposedlitter samples was used and ground to a fine powder. In a first step,protein was hydrolysed using boiling acid detergent (25 g cetyltrimethyl ammonium bromide dissolved in 2.5 L of 0.5 M H2SO4).This first fraction was designated as labile material. The fractionremaining was comprised of cellulose, lignin and ash. It was thentreated with 72% sulphuric acid to remove cellulose, leavinga fraction comprising lignin and ash. The residue was ignited at550 �C to combust all remaining organic matter, leaving the inor-ganic ash fraction. The amount lost on combustion was defined asacidedetergent lignin. For the present work, we used the initialdecanted fraction (after the initial acid detergent treatment) asa measure of the labile fraction and the lignin fraction as a measureof recalcitrant material.

2.3. Litter incubation

Following the UV radiation exposure described above, littersamples were then moistened and incubated in the dark. Sub-samples of the UV-exposed pine needle (2.0 g) and grass (0.5 g)material were weighed and placed into 1.8 L glass jars, inoculatedwith appropriate microbial extracts and incubated at 25 �C for 39days.

Microbial extracts were prepared from naturally decomposingmaterials, decomposing pine needles from the fresh humus layer ofa pine stand, and decomposing grass litter from recently cut lawngrass. Extracts were prepared from the media by shaking withdistilledwater in an orbital shaker (50 rpm) for 30min at 20 �C (DesRoss, pers. comm.). The decomposing grass was shaken at a 4:1water to grass ratio, and the pine-litter sample was shaken at a 10:1water to litter ratio both extracts were sieved through a 250 mmsieve to exclude larger litter fragments and undecomposed freshlitter material.

The concentrated stock solution was tested at a range of dilu-tions to ensure that sufficient inoculant was added to allow unre-strained decomposition of the litter but without adding excessiveextra carbon from the inoculants. The same CO2 efflux rates wereobtained across a wide range of inoculant dilutions (data notshown) so that the mid-range dilutions of 100:1 for pine-litterextracts and 1000:1 for grass-litter extracts could be used withconfidence. Between 5 and 7 mL of the microbial inoculant wassprayed onto the litter samples and the jars sealed with lids con-taining septa for gas sampling, and placed in a constant tempera-ture room in random order.

Gas samples were taken on days 3, 5, 7, 10, 12, 14, 17, 21, 28, and39 of the incubation. A 25 mL gas sample of the headspace of eachlitter sample was taken through the septum in the lid of the jar. Thegas sample was then pushed into an evacuated 12 mL glass vial(Labco Limited, Buckinghamshire, United Kingdom) and the carbondioxide (CO2) content of the gas sample measured with a gaschromatograph by flame ionisation after conversion of CO2 tomethane (Shimadzu 2010, Kyoto, Japan). Following each gassampling, the incubating jars were left open for 30 min in a fumecupboard to allow oxygen to be replenished and prevent any build-up of other gases. The moisture content of samples was then

-0.2

0.0

0 20 40 60 80 100

-2

-1

0

eig

ht c

ha

ng

e (%

)

a

Grass litter

M.U.F. Kirschbaum et al. / Soil Biology & Biochemistry 43 (2011) 1300e1307 1303

restored to their original weights by adding distilled water. Jarswere then resealed and incubated again at 25 �C.

After the end of the incubation period, the degraded littersamples, and as much as practical of the remaining moisture, wereremoved from the jars, dried at 80 �C, and the dry weight of theremaining organic material recorded. Decomposition activity wasassessed by both the cumulative amount of CO2 released and by theamount of organic matter remaining at the end of the incubation.Both those provided consistent results, thus confirming the meth-odology but providing no additional information. In the following,we therefore report only cumulative gas fluxes.

Statistical analyses were undertaken using StudenteNewmaneKuells ANOVA to determine the effects of each treatment on the littersamples. Treatment effects were considered to be significant ifp < 0.05.

-1.0

-0.8

-0.6

-0.4

0 20 40 60 80 100

UV Exposure (%)

W

b

Pine litter

Fig. 3. Weight change in samples incubated under UV radiation for 60 days. Data forgrass (a) and pine litter (b) exposed to different UV intensities. Data are expressed asa percentage of the UV irradiance received by unshaded samples, which received about11.6 MJ m�2 of UV-B radiation (290e320 nm) over the course of the experiment. Dataare means � 95% confidence intervals of 5 replicates. There were no significantdifferences with UV exposure for either pine or grass litter.

0 1 2 3 4 5

-1.5

-1.0

-0.5

0.0

0.5

Litter weight (g)

Weig

ht ch

an

ge (%

)

Grass litter

Fig. 4. Weight change in samples incubated under UV lights for 60 days with differentamounts of litter to create a range of self-shading. Data are means � 95% confidenceintervals for 5 replicates. There were no significant differences between samples withdifferent litter weights.

3. Results

Before UV exposure, over-dried litter material was green.Following UV exposure for 60 days, there was strong bleaching ofthe grass samples (data not shown). Only the most heavily shadedsamples (receiving only 1.4% of incident UV radiation) still retaineda slight tinge of green. In the self-shading treatment, it was alsoapparent that litter at the bottom of the most heavily packed trayshad also retained a slight green tinge, whereas blades higher up inthe tray were completely bleached (data not shown). Therewas notmuch bleaching of pine samples, with little difference in the colourof samples exposed to different intensities of radiation (data notshown).

Weight lost showed no significant relationship with UV expo-sure for either pine or grass litter samples (Fig. 3). Observed weightchanges were also quantitatively very small and amounted to onlya fraction of a percent over the full exposure period. Reversibleweight changes of up to a few percent were also observed whensamples were weighed throughout the exposure period (data notshown) and were likely to constituted variable water-vapouradsorption in response to changes in ambient water-vapour levels(see Fig. 2). Samples exposed to different UV exposure were,however, measured on the same days and should have beenaffected in the same way by atmospheric water vapour. Variablewater vapour should not have confounded the responses to UVexposure shown here.

In the self-shading treatment (Fig. 4), weight changes alsoshowed no statistical relationship with litter weight, especially nogreater weight loss for lower-weight samples. Instead, the leastapparentweight losswas observed for the lowest litterweights, andequal greatest weight loss for the greatest and second-lowest litterweights. As greater litter weights would have reduced the averageradiation received by samples, it would imply greater weight loss atreduced radiation exposure. Hence, this treatment also gave noindication of weight loss to increase with average UV exposure.

It is likely that the observed apparent slight weight loss ofsamples (Figs. 3 and 4) was due to different moisture adsorption onthe pre-incubation day relative to that on measurement days (seeFig. 2). Even though our experimental procedure had been designedto minimise any confounding effect of moisture adsorption, itappears that wewere unable to eliminate the problems completely.Hence, the recorded changes of up to 1.5% of initial dry weightprobably did not constitute a change in dry weight but a change inadsorbed moisture. If there had been changes in dry weight inresponse to UV exposure, it should have differentially affectedsamples exposed to different intensity (Fig. 3) or extent of self-shading (Fig. 4). Hence, it is apparent that UV exposure either didnot lead to dry-weight losses at all, or that any changes would have

been so small as not to be apparent within the variability due tomoisture interactions and other random effects.

Following UV exposure, a sub-sample of exposed litter sampleswas used for chemical analysis. Grass litter showed no change inchemical properties with UV exposure. Under all UV exposurelevels, about 68e70% of grass litter was classed as labile material(Fig. 5a) and only 2e3% as lignin (Fig. 5c). Pine litter had a lowerpercentage of only 45e55% labile material (Fig. 5b), whichdecreases slightly, but not significantly, with increasing UV expo-sure. Pine litter contained about 25e35% lignin which increasedslightly with increasing UV exposure (Fig. 5d). There was no indi-cation that lignin decreased with increasing UV exposure as hadbeen hypothesised if UV exposure had rendered litter samplesmore decomposable.

0 20 40 60 80 100

64

66

68

70

72

74

76

0 20 40 60 80 100

0

1

2

3

4

5

UV Exposure (%)

Pe

rc

en

ta

ge

o

f d

ry

w

eig

ht (%

)

Labile

Grass litter

0 20 40 60 80 100

30

40

50

60

70

0 20 40 60 80 100

10

20

30

40

50

60

Pine litter

Labile

Lignin Lignin

a b

c d

Fig. 5. Chemical analysis of litter samples, showing labile (a, b), and lignin (c, d) fractions for grass (a, c) and pine (b, d) litter. Data show means � 95% confidence intervals. Therewere no significant differences between individual samples at different UV exposure, but for lignin concentration of pine litters, a regression line had a positive slope with p < 0.05.

M.U.F. Kirschbaum et al. / Soil Biology & Biochemistry 43 (2011) 1300e13071304

The remaining material from the litter samples was inoculatedwith microbial extracts and incubated in the dark to assess anychanges in litter degradability. Grass litter decomposition pro-ceeded rapidly, with peak rates reached after about 6 days (Fig. 6a),when about 25 mg C g C�1 d�1 of the initially available carbon wasrespired each day. Rates then declined rapidly, and there was littleactivity remaining at the end of the 39-day incubation.

0

2

4

6

0 10 20 30 40

0 10 20 30 40

0

10

20

30

Incubation time (d)

CO

2 efflu

x rate (m

gC

g

C-1 d

-1)

b

a

Grass litter

Pine litter

Fig. 6. CO2 efflux rates as a function of time for grass (a) and pine (b) litter samplesfrom exposure and self-shading experiments. Data show means � 95% confidenceintervals of generally 5 samples per treatment, for a total of 70 grass (from exposureand self-shading treatments) and 30 pine samples. Some error margins were so smallto be obscured by the size of symbols. A small number of samples were lost during theincubation due to technical problems. Data are expressed in units of carbon loss perunit of original sample mass.

In contrast, it took about 12 days for peak CO2-loss rates to bereached for pine litter, and maximum daily carbon loss reachedonly about 5 mg C g C�1 d�1 (Fig. 6b). The subsequent decrease wasless sharp for pine than for grass litter, and rates at the end of theincubation were still about 1/3 of peak rates.

For grass litter, the cumulative amount of carbon lost to respi-ration over the incubation periodwas lowest for litter that had beenexposed to the highest UV intensity, and highest for litter exposedto the lowest intensity (Fig. 7a), but the differences was only about15% (between highest and lowest cumulative CO2 efflux). It wasalso in the opposite direction from what had been expected.

There was no apparent trend with UV exposure in carbon effluxof the pine samples (Fig. 7b). In particular, none of the samples ofeither grass or pine litter showed increasing carbon loss withincreasing UV exposure as had been hypothesised. In the self-shading experiment, there was also no statistical correlationbetween carbon efflux and the amount of litter weight used duringthe exposure phase (Fig. 8).

We also assessed whether microbial facilitation might havemanifested itself by samples reaching peak decomposition ratesearlier, but therewas no relationship between UV exposure and theperiod of peak CO2 rates (data not shown). The amount of residue ofsamples remaining at the end of the incubation was weighed andprovided an independent measure of the effect of UV incubation onmicrobial decomposition, with data being consistent with thoseobtained from measuring CO2 efflux (data not shown). Thisprovided an independent check of the accuracy of the method-ology, but provided no additional information and is therefore notreported here.

4. Discussion

Numerous workers have convincingly demonstrated in fieldexperiments that exposure to visible light and UV radiation couldenhance litter or organic matter decomposition (Moorhead andReynolds, 1989; Anesio et al., 1999; Schade et al., 1999; Day et al.,2007; Austin and Vivanco, 2006; Rutledge et al., 2010). We setout to confirm and quantify photooxidation through measuring the

0 20 40 60 80 100

0 20 40 60 80 100

300

350

400

450

80

100

120

140

UV Exposure (%)

Cu

mu

la

tiv

e C

O2 e

fflu

x (m

gC

g

C-1)

b

aGrass litter

Pine litter

b

aa

a,ba,b

a,b

Fig. 7. Cumulative CO2 efflux in samples incubated for 39 days at 25 �C in the darkafter having been exposed to UV radiation for 60 days. Data for grass (a) and pine litter(b) that had been exposed to different UV intensities (a) as indicated in the Figure. Datashow means � 95% confidence intervals. Points with the same letters in graph (a) arenot significantly different (p < 0.05). There were no significant differences betweenpine samples exposed to different UV intensities.

M.U.F. Kirschbaum et al. / Soil Biology & Biochemistry 43 (2011) 1300e1307 1305

effect of UV exposure under tightly controlled conditions in thelaboratory. We used two different litter samples, dried them andexposed them to UV radiation of an intensity (Fig. 1) and fora duration that tried to emulate the conditions that they mightexperience under typical field conditions over summer.

Exposure of our litter samples clearly bleached grass samples inall but the most deeply shaded treatments. However, it led to nodirect weight loss. Weight loss did not increase with radiationintensity, whether those different intensities were caused bydifferential shading through external screens (Fig. 3), or by self-shading in deeper litter samples (Fig. 4). The same conclusion wasreached for grass as for pine-litter samples.

Hence, despite the high and long exposure level, a direct weightloss due to photooxidation, was not discernable within themeasurement error of the experiment. Changes were thus either

0 1 2 3 4 5

300

350

400

Litter weight (g)

Cu

mu

lative C

O2

efflu

x (m

gC

g

C-1) Grass litter

Fig. 8. Cumulative CO2 efflux in samples incubated for 39 days at 25 �C in the darkafter having been exposed to UV radiation for 60 days. Data shown are for differentamounts of grass litter to induce self-shading during UV exposure. For the microbialincubations, the same quantity of litter material was used for all samples. Data showmeans � 95% confidence intervals. There were no significant differences betweensamples at different litter weights.

totally absent or much smaller than the changes observed byseveral authors in field experiments. Under our experimentalconditions, weight changes appeared to be dominated by apparentadsorption of atmospheric humidity (Fig. 2). While our experi-mental approach was designed to minimise any effects of water-vapour adsorption, we were not able to completely eliminate it asa confounding factor. However, it would not have affected therelativities between samples exposed to different UV intensities asthey were all measured under essentially the same atmosphericconditions.

We also investigated whether UV exposure might have merelysplit large and recalcitrant organic materials into smaller, morelabile compounds that would have been easier to decompose bysubsequent microbial action (microbial facilitation). Lignin, inparticular, is highly resistant to microbial attack but readily absorbsUV radiation (Gould, 1982; Lanzalunga and Bietti, 2000) so that itsbreakdown into more labile materials, evenwithout associated CO2loss, seemed possible (Henry et al., 2008). It is well known fromwood and paper processing that light exposure can lead tobleaching (Lanzalunga and Bietti, 2000), which was also observedin the present study, but the interaction between bleaching andmicrobial facilitation is less clear.

However, in the present study, we found no decrease in ligninconcentrations under UV exposure, nor an increase in theconcentration of labile material. If anything, therewas an indicationthat lignin in pine samples had increased with UV exposure. Wealso observed no UV facilitation of microbial breakdown. CO2 effluxfrom pine and grass samples exposed to different UV levels througheither shading by external screen or through self-shading showedno consistent relationship with prior UV exposure (Figs. 7 and 8).Hence, unlike Foereid et al. (2010), we found no consistent evidencefor UV exposure to render litter samples more susceptible tosubsequent microbial breakdown.

The results reported here appear to be very clear and robust, butare difficult to reconcile with the observations from field studiesthat have shown enhancement of decomposition rates by directvisible light and UV radiation. It is evenmore difficult to understandgiven that under field conditions UV has two potentially competingeffects. Direct photooxidation can enhance decomposition, but UVis also harmful to microbial populations and can thus reduce therate of microbial decomposition.

Gallo et al. (2006), for example, simultaneously manipulatedthe amount of UV radiation (A and B), litter type, and moistureavailability. While exposure to UV radiation did not affect overalldecomposition rates of the two litter types in their study, UVexposure interacted with moisture availability to affect thequantity and chemistry of dissolved organic matter leached fromlitter. In addition, Gallo et al. (2006) found litter that received highlevels of moisture and was not exposed to UV radiation decom-posed at the same rate as litter that received low moisture andwas exposed to UV radiation. The authors concluded thatdecomposition via UV radiation is just as effective as microbialdecomposition.

In systems that are typically dry and receive high levels of solarradiation, higher UV exposure was generally shown to be associ-ated with slightly higher decomposition rates (Verhoef et al., 2000;Austin and Vivanco, 2006; Gallo et al., 2006; Day et al., 2007).Others found increased UV-B radiation either did not affectdecomposition or decreased it because of detrimental effects on themicrobial community (Gehrke et al., 1995; Duguay and Klironomos,2000; Moody et al., 2001; Pancotto et al., 2003). These experimentswere generally conducted in high-latitude systems or systems withsubstantial canopy cover (e.g. forests), both of which normallyexperience relatively low levels of total solar radiation and thatexperience conditions that favour microbial decomposition.

M.U.F. Kirschbaum et al. / Soil Biology & Biochemistry 43 (2011) 1300e13071306

These studies indicate that the role of photooxidation isdependent on the balance between abiotic and biotic drivers in thedecomposition process. This balance can shift between arid andmore mesic ecosystems, but can also shift in the same ecosystembetween wet and dry periods. UV effects on decomposition may begreater in arid and semiarid systems where the faunal componentof decomposition is minimized by the harsh abiotic environment,where photooxidation is maximized by lack of canopy interceptionand where biotic decomposition is limited by water limitation sothat abiotic photooxidation can potentially play a bigger role(Kochy andWilson, 1997; Pancotto et al., 2005; Austin and Vivanco,2006; Zepp et al., 2007; Brandt et al., 2007; Smith et al., 2010).

Henry et al. (2008) found that grass litter exposed on the surfaceover summer lost approximately twice as much of its lignin fractionas its general mass. This proportionally high lignin loss is consistentwith photooxidation being the primary driver of lignin breakdownover summer. The large effect of summer sun exposure ondecomposition indicates that lignin degradation can make littermore susceptible to subsequent microbial degradation underwetter conditions. The effects of summer sun exposure on ensuingdecomposition were particularly strong for grass leaves. Ligninphotooxidation results in the formation of low molecular weightwater-soluble degradation products that can be easily washed away(Fiest and Hon, 1984). Given that a large fraction of cellulose can beprotected by lignin in lignocellulose complexes (Berg and Staaf,1980), it follows that photooxidation of lignin can providedecomposers greater access to relatively labile compounds such ascellulose throughout the wet season (Henry et al., 2008).

However, these various field observations are also at odds withthe results of our experiment. As we conducted our UV exposure onair-dry samples, microbial decomposition should have been insig-nificant so that the relative importance of non-biotic photooxida-tion should have been at a maximum. Similarly, while microbialfacilitation through the breakdown of lignin or other complex andrecalcitrant constituents could potentially have played an impor-tant role under our experimental conditions, we did, in fact, notobserve any enhancement of microbial decomposition at all.

If anything, our chemical analyses indicated (with marginalstatistical significance) that lignin levels in pine samples werehigher under greater UV exposure. The final step of lignin forma-tion involves the radical-mediated oxidative coupling of pre-cursormolecules (Hatfield and Vermerris, 2001), and radical formationcould have possibly been caused by UV. However, it seems unlikelythat it should have occurred to a significant extent in dry samplesunder our experimental conditions, or that there should have beena sizeable amount of pre-cursor molecules to lead to substantiallignin formation. We are thus not able to reconcile the apparentcontradiction between our findings and that of other researchreported in the literature.

It is possible that moisture status could possibly play a role(Dirks et al., 2010). Our samples were kept air-dry throughout theUV exposure, whereas samples under field conditions would bewetted at least occasionally by due or rainfall or high relativehumidity could at least allow some moisture adsorption (Dirkset al., 2010). As studies of aquatic ecosystems has shown that UVexposure of materials results in a reduction of the average molec-ular mass of organic compounds, alteration of the capacity toabsorb light both in the ultraviolet and visible spectrum, and theformation of novel photoproducts (Lanzalunga and Bietti, 2000).Photochemical reactions change the quality of dissolved organicmatter (DOM) and produce dissolved inorganic carbon and volatileCO2, CO and carbonyl sulphides. Photochemical reactions have beenwell-observed in aquatic systems (Kieber et al., 1989; Miller andZepp, 1995; Mopper et al., 1991; Tarr et al., 1995; Mayer et al.,2006, 2009) where it is both possible for photochemical reactions

to interact with chemical reactions (Lanzalunga and Bietti, 2000)and for any reaction products to be dissolved and washed awayfrom the original reaction source. The absence of that interactionbetween photochemical reactions and an aqueous medium mightpossibly be the factor that caused effects to be mooted, or absent,under our experimental conditions.

The fact that our experimental set-up allowed for no removal ofdissolved organic matter could possibly account for the absence ofany weight losses whereas they could possibly occur under fieldcondition in even dry environments. At the same time, if greaterproduction of dissolved organicmatter had occurred, it should haveled to some increase in the proportion of labile material in ourchemical analysis, or a stimulation of microbial decomposition,which we also did not observe.

We also used live plant material, which we oven-dried beforeUV exposure, whereas most other work, including previous labo-ratory incubation (Brandt et al., 2009; Foereid et al., 2010), usednaturally senesced material. It may be possible that live plantmaterial contains screening compounds that protects the live plantfrom UV damage (e.g. Kumari et al., 2009). If those compounds aredegraded as part of normal senescence processes, it might rendernaturally senesced plant material more susceptible to subsequentUV exposure than material taken from live plants. To account forthe large difference between some of the observations it would,however, require a substantial removal of these compounds duringsenescence, but this possibility cannot be discounted withoutfurther specific work to compare live and senesced material.

We conclude from our observations that under dry conditions, UVradiation does not lead to direct photooxidation of fresh litter samples,or only at such low rates that were no discernable with our experi-mental approach. Brandt et al. (2009) used a more sensitive method-ology to detect changes in carbon loss, andwhile they did observe CO2emissions, their measured rates were very small. Foereid et al. (2010)observed no direct weight loss as a result of radiation exposure, butsubstantial microbial facilitation. The key differences between theirwork and ours was that they exposed their samples to radiation con-taining both UV and visible radiation whereas we used only UV radi-ation, and we used fresh rather than senesced plant material.

In our work we found no direct weight loss after UV exposurenor subsequent microbial facilitation. These findings are thus indirect contradiction to a large number of field observations. At thisstage, we are not able to reconcile these different findings.

The question of the overall importance of photooxidation ormicrobial facilitation by UV exposure therefore still remainsunanswered. We had hoped to be able to quantify the effect ofphotooxidation as a step towards assessing its global importance.Instead, we found no effect of UV radiation under the conditionsthat had been designed to allow its direct quantification. This clearconflict between our findings and those of other studies calls foridentification of the key processes, or experimental conditions, thatare critically important in causing the difference in litter responseto UV exposure in field vs laboratory studies.

Acknowledgments

We would like to thank Rainer Hofmann and Richard McKenziefor useful background information about the conduct of UV exper-iments, Stephen Stilwell for assistancewith calibrating our radiationsources and UV sensor and RichardMcKenzie for provision of a solarspectrum of typical New Zealand solar radiation.Wewould also liketo thank AdrianWalcroft and Ted Pinkney for technical assistance inthe set-up of the experiment, Des Ross for advice on the extractionmethod for preparing microbial inoculant, Greg Arnold and GuyForrester for statistical advice, and Adrian Walcroft and LouisSchipper for useful comments and suggestions on the manuscript.

M.U.F. Kirschbaum et al. / Soil Biology & Biochemistry 43 (2011) 1300e1307 1307

References

Anesio, A.M., Tranvik, L.J., Graneli, W., 1999. Production of inorganic carbon fromaquatic macrophytes by solar radiation. Ecology 80, 1852e1859.

Austin, A.T., Vivanco, L., 2006. Plant litter decomposition in a semi-arid ecosystemcontrolled by photodegradation. Nature 442, 555e558.

Berg, B., Staaf, H., 1980. Decomposition rate and chemical changes of Scots pineneedle litter. II. Influence of chemical composition. In: Persson, T. (Ed.), Struc-ture and Function of Northern Coniferous Forests. Ecological Bulletin, Stock-holm, pp. 373e390.

Borken, W., Matzner, E., 2009. Reappraisal of drying and wetting effects on C and Nmineralization and fluxes in soils. Global Change Biology 15, 808e824.

Brandt, L.A., Bohnet, C., King, J.Y., 2009. Photochemically induced carbon dioxideproduction as a mechanism for carbon loss from plant litter in arid ecosystems.Journal of Geophysical Research 114, GO2004. doi:10.1029/2008JG000772.

Brandt, L.A., King, J.Y., Hobbie, S.E., Milchunas, D.G., Sinsabaugh, R.L., 2010. The roleof photodegradation in surface litter decomposition across a grasslandecosystem precipitation gradient. Ecosystems 13, 765e781.

Brandt, L.A., King, J.Y., Milchunas, D.G., 2007. Effects of ultraviolet radiation on litterdecomposition depend on precipitation and litter chemistry in a shortgrasssteppe ecosystem. Global Change Biology 13, 2193e2205.

Day, T.A., Zhang, E.T., Ruhland, C.T., 2007. Exposure to solar UV-B radiation accel-erates mass and lignin loss of Larrea tridentata in the Sonoran Desert. PlantEcology 193, 185e194.

Dirks, I., Navon, Y., Kanas, D., Dumbur, R., Grünzweig, J.M., 2010. Atmospheric watervapor as driver of litter decomposition in Mediterranean shrublands andgrasslands during rainless season. Global Change Biology 16, 2799e2812.

Duguay, K.J., Klironomos, J.N., 2000. Direct and indirect effects of enhanced UV-Bradiation on the decomposing and competitive abilities of saprobic fungi.Applied Soil Ecology 14, 157e164.

Fiest, W.C., Hon, D.N.S., 1984. In: Rowell, R.M. (Ed.), The Chemistry of Solid Wood.American Chemical Society, Washington, DC, pp. 401e451.

Foereid, B., Bellarby, J., Meier-Augenstein, W., Kemp, H., 2010. Does light exposuremake plant litter more degradable? Plant and Soil 333, 275e285.

Gallo, M.E., Sinsabaugh, R.L., Cabaniss, S.E., 2006. The role of ultraviolet radiation inlitter decomposition in arid ecosystems. Applied Soil Ecology 34, 82e91.

Gehrke, C.J.U., Callaghan, T.V., Chadwick, D., Robinson, C.H., 1995. The impact ofenhanced ultraviolet-B radiation on litter quality and decomposition processesin Vaccinium leaves from the Subarctic. Oikos 72, 213e222.

Gould, J.M., 1982. Characterization of lignin in situ by photoacoustic spectroscopy.Plant Physiology 70, 1521e1525.

Hatfield, R., Vermerris, W., 2001. Lignin formation in plants. The dilemma of linkagespecificity. Plant Physiology 126, 1351e1357.

Henry, H.A.L., Brizgys, K., Field, C.B., 2008. Litter decomposition in a Californiaannual grassland: interactions between photodegradation and litter layerthickness. Ecosystems 11, 545e554.

Kieber, D.J., McDaniel, J., Mopper, K., 1989. Photochemical source of biologicalsubstrates in sea water: implications for carbon cycling. Nature 341, 637e639.

Kirschbaum, M.U.F., 2000. Will changes in soil organic matter act as a positive ornegative feedback on global warming? Biogeochemistry 48, 21e51.

Kirschbaum, M.U.F., 2010. The temperature dependence of organic matter decom-position: seasonal temperature variations turn a sharp short-term temperatureresponse into a more moderate annually-averaged response. Global ChangeBiology 16, 2117e2119.

Kochy, M., Wilson, S.D., 1997. Litter decomposition and nitrogen dynamics in aspenforest and mixed-grass prairie. Ecology 78, 732e739.

Kumari, R., Singh, S., Agrawal, S.B., 2009. Combined effects of Psoralens and ultra-violet-B on growth, pigmentation and biochemical parameters of Abelmoschusesculentus L. Ecotoxicology and Environmental Safety 72, 1129e1136.

Lanzalunga, O., Bietti, M., 2000. Photo- and radiation chemical induced degradationof lignin model compounds. Journal of Photochemistry and Photobiology B:Biology 56, 85e108.

Mayer, L.M., Schick, L.L., Hardy, K.R., Estapa, M.L., 2009. Photodissolution and otherphotochemical changes upon irradiation of algal detritus. Limnology andOceanography 54, 1688e1698.

Mayer, L.M., Schick, L.L., Skorko, K., Boss, E., 2006. Photodissolution of particulateorganic matter from sediments. Limnology and Oceanography 51, 1064e1071.

McKenzie, R.L., Liley, J.B., Björn, L.O., 2009. UV radiation: balancing risks andbenefits. Photochemistry and Photobiology 85, 88e98.

Miller, W.L., Zepp, R.G., 1995. Photochemical production of dissolved inorganiccarbon from terrestrial organic matter: significance to the oceanic organiccarbon cycle. Geophysical Research Letters 22, 417e420.

Moody, S.A., Paul, N.G., Björn, L.O., Callaghan, T.V., Lee, J.A., Manetas, Y., Rozema, J.,Gwynn-Jones, D., Johanson, U., Kyparissis, A., Oudejans, A.M.C., 2001. The directeffects of UV-B radiation on Betula pubescens litter decomposing at four Euro-pean field sites. Plant Ecology 154, 29e36.

Moorhead, D.L., Reynolds, J.F., 1989. Mechanisms of surface litter mass-loss in thenorthern Chihuahuan Desert e a reinterpretation. Journal of Arid Environments16, 157e163.

Mopper, K., Zhou, X.L., Kieber, R.J., Kieber, D.J., Sikorski, R.J., Jones, R.D., 1991.Photochemical degradation of dissolved organic carbon and its impact on theoceanic carbon cycle. Nature 353, 60e62.

Pancotto, V.A., Sala, O.E., Cabello, M., Lopez, N.I., Robson, M., Ballare, C.L.,Caldwell, M.M., Scopel, A.L., 2003. Solar UV-B Decreases Decomposition inHerbaceous Plant Litter in Tierra del Fuego, Argentina: potential role of analtered decomposer community. Global Change Biology 9, 1465e1474.

Pancotto, V.A., Sala, O.E., Robson, M., Caldwell, M.M., Scopel, A.L., 2005. Direct andindirect effects of solar UV-B radiation on long-term decomposition. GlobalChange Biology 11, 1982e1989.

Rowland, A.P., Roberts, J.D., 1994. Lignin and cellulose fractions in decompositionstudies using acid detergent fibre methods. Communications in Soil Science andPlant Analysis 25, 269e277.

Rutledge, S., Campbell, D.I., Baldocchi, D., Schipper, L.A., 2010. Photodegradationleads to increased CO2 losses from terrestrial organic matter. Global ChangeBiology 16, 3065e3074.

Schade, G.W., Hofmann, R.R., Crutzen, P.J., 1999. CO emissions from degrading plantmatter. (1) Measurements. Tellus B 51, 889e908.

Sitch, S., Huntingford, C., Gedney, N., Levy, P.E., Lomas, M., Piao, S.L., Betts, R.,Ciais, P., Cox, P., Friedlingstein, P., Jones, C.D., Prentice, I.C., Woodward, F.I., 2008.Evaluation of the terrestrial carbon cycle, future plant geography and climate-carbon cycle feedbacks using five Dynamic Global Vegetation Models (DGVMs).Global Change Biology 14, 2015e2039.

Smith, W.K., Gao, W., Steltzer, H., Wallenstein, M.D., Tree, R., 2010. Moisture avail-ability influences the effect of ultraviolet-B radiation on leaf litter decomposi-tion. Global Change Biology 16, 484e495.

Tarr, M.A., Miller, W.L., Zepp, R.G., 1995. Direct carbon monoxide production fromplant matter. Journal of Geophysical Research 100, 11403e11413.

Verhoef, H.A., Verspagen, J.M.H., Zoomer, H.R., 2000. Direct and indirect effects ofultraviolet-B radiation on soil biota, decomposition and nutrient fluxes in dunegrassland soil systems. Biology and Fertility of Soils 31, 366e371.

Zepp, R.G., Erickson III, D.J., Paul, N.D., Sulzberger, B., 2007. Interactive effects ofsolar UV radiation and climate change on biogeochemical cycling. Photo-chemical and Photobiological Science 6, 286e300.