research article change in soil and forest floor carbon...

Research ArticleChange in Soil and Forest Floor Carbon after ShelterwoodHarvests in a New England Oak-Hardwood Forest USA

Kayanna L Warren and Mark S Ashton

School of Forestry and Environmental Studies Yale University 195 Prospect Street New Haven CT 06511 USA

Correspondence should be addressed to Mark S Ashton markashtonyaleedu

Received 7 December 2013 Revised 13 March 2014 Accepted 27 March 2014 Published 6 May 2014

Academic Editor Timothy Martin

Copyright copy 2014 K L Warren and M S Ashton This is an open access article distributed under the Creative CommonsAttribution License which permits unrestricted use distribution and reproduction in any medium provided the original work isproperly cited

There has been effort worldwide to quantify how much carbon forests contain in order to designate appropriate offset credits toforest carbon climatemitigation Carbon pools on or immediately below the soil surface are understood to be very active in responseto environmental change but are not well understood Our study focused on the effects of shelterwood regeneration harvests inNewEngland on the carbon stored in litter woody debris and surface soil carbon Results demonstrate significant difference in surface(0ndash10 cm) soil carbon between control (nonharvested) and harvested sites with higher carbon percentage on control sites Resultsshowed a significant difference in coarse woody debris with higher amounts of carbon per area on harvested sites No significantdifference in litter mass was recorded between harvested and control sites When coarse woody debris and litter are included withsoil carbon total carbon did not have a significant decline over 20 years following shelterwood treatment to the forest to secureregeneration but there was considerable variability among sites When taking all surface soil carbon measurements together ourresults suggest that for accounting purposes the measurement of below-ground carbon after shelterwood harvests is not necessaryfor the southern New England region

1 Introduction

The terrestrial carbon cycle takes up a considerable amountof atmospheric carbon especially CO

2 [1 2] the increase

of which has been cited as the main driver of global climatechange [3 4] Maintenance of forest cover and reduction ofdeforestation rates have been proposed by numerous studiesand policy processes as a means to reduce the magnitude ofclimate change [5ndash7] It is therefore important that below-ground carbon pools are well understood

Persistent lack of clarity surrounds much of the efforts totabulate quantify and account for the amount of carbon thatis stored by forest ecosystems and thus makes its inclusion inpolicy protocols and its accounting expensive and challeng-ing [8ndash11] There is ongoing debate about how managementaffects overall carbon with some studies suggesting that oldgrowth forests store more carbon [12ndash16] while other studiesfind that continuous timber harvests reduce atmosphericcarbon levels more [7 17] However the effects of anthro-pogenic disturbance such as harvest on belowground carbon

quantitymdashbelowground poolsmdashis still largely unknown andthinly researched [18ndash22] Only a handful of studies on soilcarbon pools represent all tropical forests combined withmost studies confined to temperate regions andmost of theseare from tree plantations that examine the surface soils only[22] For temperate forests primary obstacles to quantifyingsoil carbon include the slow rate of carbon accumulation [23]and the spatial variability of soil [7 24ndash26]

However even though little research has focused onbelow-ground carbon studies suggest forest soils litter andcoarse woody debris may comprise half of terrestrial carbonstorage and over two-thirds of forest carbon pools [1 23 2728] Temperate forests in the eastern United States may be asignificant carbon sink [29]

Studies on the effects of timber harvest on soil carbonhave yielded mixed results thus far [30ndash32] For example soilcarbon that is initially lostmay take a few years to hundreds ofyears to recover depending upon forest type climate and soilcondition [30 33ndash35] Other studies report that soil carbon isnot necessarily lost at all [27 36 37] In a recent study Johnson

Hindawi Publishing CorporationInternational Journal of Forestry ResearchVolume 2014 Article ID 527236 9 pageshttpdxdoiorg1011552014527236

2 International Journal of Forestry Research

et al [38] found a consistent drop in soil carbon after a timberharvest for a northern hardwood forest in New EnglandCarbon accretion in soils of regenerating forests may befurther limited by soil texture and other environmentalconditions [39] Information on second-growth hardwoodforests is particularly important because conclusions fromstudies on carbon sink quantification conflict [22]

Higher rates of carbon loss occur in the forest floor morethan in the soil for hardwood forests [40 41] The reasonsfor changes in soil carbon relate strongly to litter inputsdecomposition and respiration [42 43] Studies suggest thatnet soil carbon storage may be influenced by small changesin climate and soil condition [41 44] Over time and underthe colder wetter climate conditions the shift of live biomassinto woody debris litter and soil humus may build upstored carbon in soils [7] Studies conflict over whether forestfloor and soil carbon increase or decrease after harvest andit is unclear whether observed changes are more stronglycorrelated with the amount of debris left on-site or the changein efflux and respiration with disturbance of the soil [36 45]Other studies have shown weak ability to correlate litterfalland belowground carbon allocation in mature temperatehardwood forests [44] and weak conversion of litter carboninto long-term forest soil carbon storage [46] Not only arethe dynamics of decomposition from coarse woody debrisand litter to soil carbon quantity poorly understood butalso much remains to be researched about the long-termdynamics of soil carbon after forest harvest

There are few studies which have examined the effectsof silvicultural treatments (regeneration or thinning) on soilcarbon stocks in the northeastern region of the United States[30 47] Some of the limitations of these studies are thatthey looked only at differences between thinnings [47] weremodeled [38] or were focused on coniferous forests ratherthan the predominant hardwood type [48] Oak-hardwoodforests of the northeastern US have shown reduced soilcarbon stocks reduced carbon in litter and increased woodydebris with harvest [49] but this study was after clearcutharvests rather than shelterwood In another study coarsewoody debris was reduced by 17 in the long term by aselective timber harvest [50] but such harvests have showndramatic increases in coarse woody debris in the short term[51] Prescriptions that purposefully leave deadwood as snagsand slash have been shown to negate any loss of coarse woodydebris [52]

Our study had two objectivesWe wanted to first quantifythe change in carbon for the uppermost layers of soiland surface horizons (coarse woody debris and litter) aftershelterwood regeneration harvests in a second-growth oak-hardwood forest Second we wanted to document the changein soil carbon coarse woody debris and litter over timeafter shelterwood harvests We sought to understand howsoil carbon is affected by forest harvest and how surfacecarbon is partitioned between coarse woody debris and litterdecomposition over a 20-year time horizon after harvest Ourstudy focuses on the most common type of forest regenera-tion harvest for oak forests in the northeastern United Statesshelterwoods a method of harvest that retains a limitedcanopy of masting nut trees (oak hickory and beech) to

provide seed for dispersal and germination and shelter fornewly established seedlings [53 54]We expected that surfacecarbon dynamics will show significant declines similar tofindings by Covington [30] and to US Forest Service datafrom clearcuts We also expected rate and amount of declineto be lower than in a clearcut because of the moderatingshade from tree retention in shelterwood systems Lastly wepredicted that the surface carbon pools will decline with themost significant change in the surface litter

2 Methods

21 Site Description of the Study Area This study was con-ducted at the 3173-ha Yale-Myers Forest near Eastford CT inthe northeastern United States (41∘571015840N 72∘071015840W)The Yale-Myers Forest is a research and demonstration forest managedby the School of Forestry and Environmental Studies andowned by Yale University The forest overstory is predomi-nantly oak-maple-pine The climate is characterized as cooltemperate and humid Mean annual summer temperature is20∘C and mean annual winter temperature is ndash4∘C Precipi-tation is evenly distributed throughout the year with a meanof 110 cm [55 56] This is a moist temperate forest underlainby a metamorphic bedrock of schistgneiss and overlain withsoils that originated from till and fluvial sediments of the lastWisconsin glacial period 20000 bp [55]The soils aremostlycoarse-loamy mesic Typic Dystrudepts [57] The topographyis characterized by undulating hills with broad ridges andnarrow valleysThe elevation ranges froma lowpoint of 170mabove sea level to a high point of 300m above sea level [56]

Starting in the early 18th century the forest as with manyareas of New England had a long period of agriculturaluse with land used for various kinds of improved andunimproved pasture tilled for crops or relict forest patchesthat were used as woodlots Since the 1820s this land haslargely been reforested naturally [55]

22 Treatment Description All stands for our study wereon glacial till soils of ablation origin and comprising theCharlton-Chatfield series [57] This soil series is a dominantone for the oak-hardwood of southern New England makingour study results generalizable to a much larger regionthan the forest All study stands were originally cleared formarginal pasture but were never plowed After abandonmentin the 1850s these sites transformed to eastern white pine(Pinus strobus L) and were then cutover in the early 1900s forthe pine boxwood industry After cutting the pine understoryhardwood regeneration comprising oaks hickories birchesand maples was released Today this second-growth hard-wood forest varies in age between 90 and 110 years of ageThis age class and forest history are very representative of thesouthern New England area

Using available GIS maps of soil type and harvests withinYale-Myers a map of shelterwood harvests and reservesof similar land use history (as described above) overlainwith the Charlton-Chatfield soil type was created to identifysites that had suitable treated and control plots adjacent toeach other (Figure 1) Controls were therefore selected as

International Journal of Forestry Research 3

Control plotTreatment plot

0 05 1 2 (km)

Figure 1 Locations of each of the 15 shelterwood treatment andadjacent control sites in the Yale-Myers Forest

an uncut reference of similar cover soil land use history andtopographic position adjacent to each shelterwood harvesttreatment We did paired comparisons of treated versuscontrol plots to determine if there is a statistical difference (atthe 005 level or greater) between shelterwood harvest and noharvest Harvested sites were selected to represent a range ofages from 0 to 20 years since harvest and so can be regardedas surrogates to investigate effects of age

At theYale-Myers Forest shelterwoodharvests reduce thecanopy in a single cutting from a fully stocked basal area ofabout 37m2ha to about 11m2ha largely comprising aboutforty 50ndash60 cmdbh well-spaced trees per ha The aim is tomaintain species diversity among regenerating trees and seedsources by insuring the heavy seeded masting trees (oakshickories and beech) adequately disperse seeds and that thepartial shade of the canopy will moderate the harsh openconditions during germination and seedling establishmentSeven to eight years later half the standing trees (basal area =5m2ha) are removed in a second and final cut leaving the restas irregular structures to grow within the new regeneratingstand Harvest operations are generally undertaken duringwinter months when the ground is frozen and coveredwith snow Skid trails and landings are designated ahead ofoperations often utilizing previous trails and historic roads

23 Sampling Design We selected fifteen study sites com-prising pairs of treated and control plots Based on a priorstudy on soils within this forest the range was expected tolie between 10 and 75 kgm2 of carbon in the top 30 cm [58]with a world average around 117 kgm2 of carbon for all soilsto 100 cm depth [59] This study estimated that northeasterntemperate soils less carbon-rich than boreal soils in Canada

would have a spread toward the lower end of that rangeThe sample size for each treatment block was selected withthe University of Iowarsquos online sample size calculator [60]with settings for two-way ANOVA Using this estimation asample of five and three plots was determined for the treatedand control sites respectively Individual plots on each sitewere selected using a random number generator and a GPScoordinate grid

Plots were centered at the GPS coordinates Plot cornerswere defined as three meters north east south and west ofthe plot center Two 254 cmwide soil cores were taken at eachpoint (center north south east and west) and pooled intoone plot sample (Figure 2) Because the upper soil horizonhas more carbon than other increments of the mineral soil atdeeper levels [58] and because most of the observable change20 years after harvest was expected to be in the upper 10 cm[58 61] we cored to a depth of 10 cm at each core point fromthe mineral soil surface and below the litter and organic soilhorizon This meant that we carefully scraped of the organichorizon to carefully identify the mineral soil surface layerThe forest floor organic horizon was sampled using a 15 cm times15 cm square template that allowed removal of the completeincorporated and unincorporated litter from the center ofthe plots This included all detritus and dead wood less than25 cm in diameter This material was bagged and taken tothe lab for processing In addition all coarse woody debrisgreater than 25 cm in diameter was measured for length anddiameter but only for material that was within the 424m times424m plot Woody debris volume was calculated based onthe volume of a cylinder Lastly tree species was recorded andan estimation of basal area was taken using a variable radiusplot protocol and a 2 BAF metric angle gauge to characterizestand structure and composition and to insure there were nodifferences across sites and treatments [62]

24 Soil Measurements and Analysis In the field soil coredepth was measured to calculate total soil volume of thecollected cores In the lab the wet and dry weights thepercentmoisture andpercent carbonwere determinedWhilethere are difficulties with assessing total carbon losses byusing percent carbon other studies have demonstrated thatit can be used as a proxy [63] Soil was sieved in the labto 2mm using the protocol of Bradford et al [64] Thedried soil samples were sent to the Ecology Lab at theUniversity of Georgia where carbon content of each samplewas determined using a micro-Dumas flash combustionCHN Analyzer essentially combusting soil components andmeasuring carbon via gas (CO

2) release [64] Carbon per unit

forest area was calculated with the following equation

C(Mgha)

= [soil bulk density (gm3) times soil depth (m) times C]

times 100 times

10000m2

1 hatimes

1Mg1000000 g

(1)

See [65] After initial tests of bulk density in our plots ourmeasurements corroborated those of Kulmatiski et al [58]


Litter sample

3m

15 cm times 15 cm

Figure 2 Diagram of plot design Two soil cores were taken at thecenter and at each corner

Using both our initial tests and a penetrometer we foundno evidence of differences in bulk densities between ourtreatment and control for each our sites or across all oursites more generally We therefore used 081 from an averagesurface soil bulk density for the Yale-Myers Forest [58] Weused the Kulmatiski et al [58] measure of soil bulk densitybecause it was based on a much more substantial soil surveyof the forest soils at Yale-Myers than our study

Litter mass was measured wet and then dried in thelaboratory at 80∘C for 48 hours [66] Carbon content wasfound by converting a ratio of carbon to dried material forlitter and through additional steps for coarse woody debrisBiomass was assumed to comprise 498 of the wood volumefor coarse woody debris and 50 of the mass for dry litterfor oak-hickory stands in the northeastern US [67] Specificwood density was estimated at 0636 [67] in this study Thereare lower estimates from US Forest Inventory and Analysis(FIA) data but we selected a higher value given the largerproportion of slower growing denser woods within our forest[68]

Based on field observation decay classes were assigned anaverage decay class of two for harvested plots younger thansix years a decay class of three for harvested plots six yearsor greater and a decay class of three for control sites Ourapproachwas a simplified protocol using the Brown et al [65]decay classification and decay coefficients fromWaddell [69]The equation to convert volume of coarse woody debris perplot to carbon mass per area was

C (g)area (m2)

= [volume (m3) lowast specific gravity

lowastbiomass ratio lowast decay class]

(2)

25 Data Analysis The fifteen sites were treated as pairedreplicates (control and shelterwood harvest) using a two-wayANOVA using R version 2122 [70] The model comprisedtreatment (harvest control) sites over time (1ndash20 years) andan interaction term (treatment times site) Variables analyzedincluded (i) percent soil carbon in the top 10 cm (ii) car-bon in the litter and forest floor (iii) estimated carbon incoarse woody debris and (iv) all pools combined All soilcarbon and litter data was nonnormally distributed and waslog-transformed while coarse woody data was cube-root-transformed prior to analysis We used Tukeyrsquos studentizedt-test to compare levels of significance among treatments forall variables we tested In addition regression analysis wasused to explore change in carbon over time to determine ifa relationship exists in surface and soil carbon and time sinceharvest treatment

3 Results

31 Comparisons in Carbon Pools between Shelterwood Har-vest and Control Sites Harvested sites were shown to havesignificantly lower amounts of surface soil carbon (carbonin the top 30 cm of the soil) (601 kgm2) than control sites(836 kgm2) that were left intact and there was significantdifference among sites and in interaction between treatmentsand sites (Figure 3 Table 1)

Shelterwood harvest also had an impact on coarse woodydebris Harvest treatments had a significantly greater amountof carbon in coarse woody debris (316 kgm2) as comparedto unharvested controls (062 kgm2) (Figure 3 Table 1) butagain significant differences were shown among sites andin interactions between site and treatment Total carbon inlitter per unit area was found to have a significant differencebetween harvested (059 kgm2) and unharvested controls(047 kgm2) (Figure 3 Table 1) But no difference was shownamong sites and interactions between site and treatment

When surface soil carbon was combined with carbonin litter and coarse woody debris on a mass per unit areano significant difference between shelterwood harvestedtreatments and unharvested controls could be demonstrated(Figure 3 Table 1) However there was a significant differenceamong sites and in an interaction between treatment and siteindicating that sites (as surrogates for time since shelterwoodcut) not only had different starting carbon stocks but alsoresponded differently to harvests over time

32 Comparisons in Carbon Pools over Time for ShelterwoodHarvest and Control Sites Downward trends in carboncontent for sites over time since harvest were seen for allthree pools but with a weak 1198772 value when analyzed witha linear regression model (Figure 4) However 119875 valueswere significant for the harvested trend lines for both thecoarse woody debris and total pooled carbon indicatinga downward trend in both of these pools over time andsupporting the contention that time since harvest is theinteracting effect in the results for theANOVACoarsewoodydebris was the most significant with a decline from 45 kgm2to 040 kgm2 over the twenty-year period suggesting that


Control Harvested

Soil carbon mass per area

Treatment

0

2000

4000

6000

8000

10000

Carb

on (g

m2)

Pa1

b1

= 0000537lowastlowastlowast

(a)

Litter carbon mass per area

Control HarvestedTreatment

0

2000

4000

6000

8000

10000

Carb

on (g

m2)

P = 0225

a2 a2

(b)

DWD carbon mass per area


0

2000

4000

6000

8000

10000

Carb

on (g

m2)

P

a3

b3


(c)

Figure 3 Mass per area of each carbon pool by treatment Tukey paired t-tests between the variables indicate level of significance at 005level whereby a gt b Different letters indicate significant difference

it had reached control baseline levels of unharvested siteswithin this time frame Soil carbon and litter suggested aweakdownward trend over time relative to the control but bothwere not significant

4 Discussion

The results indicate that soil carbon pools significantlydeclined after shelterwood harvest but that total soil carbonmass per area in the top 10 cm showed no significant differ-ence as compared to uncut controls Carbon mass per area ofcoarse woody debris was found to increase significantly afterharvest as expected Furthermore the difference betweenmore recently cut sites and older sites appeared to narrow asexpected with the decomposition of woody material It couldbe concluded that over the 20-year period of shelterwoodharvests differences in amount of coarse woody debris could

be related to changes in site treatment (eg prescribed burn-ing whole-tree chipping etc) We can attest that no changesoccurred All site treatments were the same with all stemsand tops less than 20 cmdbh being left purposefully scatteredacross the harvest sites However this trend may have beenmore accurately depicted if woody debris volume had beenindividually classified to a decay class rather than pooled toparticular years since harvest Litter carbon was not shownto be significantly different between shelterwood harvest andcontrol plots although it was shown to be generally higherthan US Forest Service estimates for postharvest oak-hickoryforests in the northeast [49] Coarse woody debris howeverwas shown to be less for those forest service estimates [49]which in turn were less than those found by Covington [30]The difference between our results and those of Covington[30] and Smith et al [49] could be attributed to methodof harvest (shelterwood versus clearcut) or to differences inmeasurement protocols and sampling design


Table 1 ANOVA of soil litter and coarse woody debris (CWD) carbon mass per area by treatment

df Surface soil Litter CWD TotalF value P F value P F value P F value P

Treatment 1 28273 0000lowastlowastlowast 55742 0020lowast 25314 0000lowastlowastlowast 0024 0877Site 14 5188 0000lowastlowastlowast 15889 0099 3868 0000lowastlowastlowast 2977 0001lowastlowastlowast

Treatment times site 14 4649 0000lowastlowastlowast 12263 0272 2041 0024lowast 3887 0000lowastlowastlowast

Significance codes are as follows lowastlowastlowast0001 lowastlowast001 and lowast005 Treatment refers to shelterwood harvest versus control Site refers to replicate sites examined thatspan a 20-year period since harvest

0 5 10 15 200

5000

10000

15000

20000Soil carbon mass per area

Years elapsed since harvest

Carb

on (g

m2)

minusControl R2=

minusHarvested R2=

00221 P = 0791

00042 P = 0404

(a)

0

500

1000

1500

2000Litter carbon mass per area

Carb

on (g

m2)

0 5 10 15 20Years elapsed since harvest

minusControl R2=

Harvested R2=

00151 P = 0551

00017 P = 0294

(b)

0

2000

4000

6000

8000

10000


Carb

on (g

m2)


Control R2=

Harvested R2=

00065 P = 0264

00932 P = 0006

(c)

05000

10000150002000025000

Total carbon mass per area

Carb

on (g

m2)


Control R2= minus

Harvested R2=

00238 P = 0991

00997 P = 0004

(d)

Figure 4 Regressions depicting mass per area change of each carbon pool over time for control and shelterwood harvested sites (a) soilcarbon change (b) litter carbon change (c) coarse woody debris carbon change and (d) total carbon change for three combined pools

Our study also showed higher on average soil carboncontent as compared to a study at the same forest byKulmatiski et al [71] whose objective was to quantify totalsoil carbon at depthTheKulmatiski et al [71] studywas forestwide irrespective of soil type On average it also sampled soilthree times deeper than our study Our soil cores went to a10 cm depth because our focus was the effects of harvestingon carbon in surface horizons They also found topographicposition accounted for 18 of soil carbon variation withwetter lower lying areas containing more carbon than drierupland soils Furthermore carbon content bulk density anddepth are correlated [72] and further research needs to bedone to measure each of these variables to verify and refinemore accurate calculations for soil carbon content unique toeach site

The difference in total carbon stocks between control andharvested sites however was not found to be significantwhenpooling together all sites of varying time since harvest Thepotential reason for the lack of significance is the dynamicchange among the different carbon pools some increasingand some decreasing over time While coarse woody debrisrepresented a smaller carbon pool than mineral soil carbonthe large increase counterbalanced the soil carbon decreaseLitter was the smallest carbon pool and also experienced nosignificant change However when examining trends over the20-year time period total carbon among all pools (soil litterand debris) showed a significant decline This was primarilydue to coarse woody debris which showed the only significantdownward trend over time The trend line for mineral soilcarbon and litter suggested a downward trend over time


relative to the control but was not significant Obtaining datafrom older harvests may have strengthened this relationshipbut no sites exist

Overall these results indicate that these three pools are asignificant portion of total carbon stocks but that the changesthat may occur due to harvest represent a relatively smallportion of total site carbon stocks Coarse woody debrisshowed the strongest significant change with shelterwoodharvest but it was a smaller pool than soil carbon Surfacesoil carbon significantly declined in the top 10 cm butfurther studies that extend over a longer time period will beimportant to do

5 Conclusions

Taking all sites together total pools of carbon (litter woodydebris and surface soil carbon) showed no significant declinewith harvest primarily because declines in surface soil carbonand litter were more than made up for by increases in coarsewoody debris However when comparing sites with timesince shelterwood harvest there are significant declines intotal carbon Our results show this is strongly driven bydecline in coarse woody debris There are also significantdeclines in surface soil carbon with harvest but there isconsiderable variability between sites and this trend is weakwith time since harvest These results suggest that for thepurposes of forest carbon accounting auditors do not needto monitor changes in below-ground carbon stocks withshelterwood harvests for this region We believe there is notenough of a significant decline or change in carbon to meritthe time and cost in measurement

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

The authors would like to thank the Carpenter-Sperry Fundand the Jubitz Family Fund for supporting this study Theywould also like to thank Mark Bradford Elaine HooperJonathan Reuning-Scherer Meredith Martin and Dan Con-stable for their advice on soil sampling methodology statisti-cal methodology and GIS mapping layout

References

[1] R K Dixon S Brown R A Houghton A M Solomon MC Trexler and J Wisniewski ldquoCarbon pools and flux of globalforest ecosystemsrdquo Science vol 263 no 5144 pp 185ndash190 1994

[2] Y Pan R A Birdsey J Fang et al ldquoA large and persistent carbonsink in the worldrsquos forestsrdquo Science vol 333 no 6045 pp 988ndash993 2011

[3] D Eamus and P G Jarvis ldquoThe direct effects of increasein the global atmospheric CO

2concentration on natural and

commercial temperate trees and forestsrdquoAdvances in EcologicalResearch C vol 19 pp 1ndash55 1989

[4] IPCC (Intergovernmental Panel on Climate Change) ClimateChange 2007 The Physical Science Basis Contribution ofWorking Group I to the Fourth Assessment Report of theIntergovernmental Panel on Climate Change IPCC GenevaSwitzerland 2010 httpwwwipccch

[5] R Bradley B Childs T Herzog J Pershing and K A BaumertSlicing the Pie Sector-Based Approaches to International ClimateAgreements World Resources Institute Washington DC USA2007

[6] G R van der Werf D C Morton R S Defries et al ldquoCO2

emissions from forest lossrdquo Nature Geoscience vol 2 no 11 pp737ndash738 2009

[7] T J Fahey P B Woodbury J J Battles et al ldquoForest carbonstorage ecology management and policyrdquo Frontiers in Ecologyand the Environment vol 8 no 5 pp 245ndash252 2010

[8] F Garcıa-Oliva and O R Masera ldquoAssessment and measure-ment issues related to soil carbon sequestration in land-useland-use change and forestry (LULUCF) projects under theKyoto protocolrdquo Climatic Change vol 65 no 3 pp 347ndash3642004

[9] R A Birdsey ldquoCarbon accounting rules and guidelines for theUnited States forest sectorrdquo Journal of Environmental Qualityvol 35 no 4 pp 1518ndash1524 2006

[10] A Gershenson and J Barsimantov ldquoAccounting for carbon insoilsrdquo Climate Action Reserve White Paper 2010

[11] J S Nunery and W S Keeton ldquoForest carbon storage in thenortheastern United States net effects of harvesting frequencypost-harvest retention and wood productsrdquo Forest Ecology andManagement vol 259 no 8 pp 1363ndash1375 2010

[12] M E Harmon W K Ferrell and J F Franklin ldquoEffects oncarbon storage of conversion of old-growth forests to youngforestsrdquo Science vol 247 no 4943 pp 699ndash702 1990

[13] J H M Thornley and M G R Cannell ldquoManaging forestsfor wood yield and carbon storage a theoretical studyrdquo TreePhysiology vol 20 no 7 pp 477ndash484 2000

[14] B E Law P E Thornton J Irvine P M Anthoni and S VanTuyl ldquoCarbon storage and fluxes in ponderosa pine forests atdifferent developmental stagesrdquo Global Change Biology vol 7no 7 pp 755ndash777 2001

[15] E A H SmithwickM E Harmon SM Remillard S A Ackerand J F Franklin ldquoPotential upper bounds of carbon stores inforests of the Pacific Northwestrdquo Ecological Applications vol 12no 5 pp 1303ndash1317 2002

[16] S Luyssaert E-D Schulze A Borner et al ldquoOld-growth forestsas global carbon sinksrdquo Nature vol 455 no 7210 pp 213ndash2152008

[17] D Markewitz ldquoFossil fuel carbon emissions from silvicultureimpacts on net carbon sequestration in forestsrdquo Forest Ecologyand Management vol 236 no 2-3 pp 153ndash161 2006

[18] K I Paul P J Polglase J G Nyakuengama and P K KhannaldquoChange in soil carbon following afforestationrdquo Forest Ecologyand Management vol 168 no 1ndash3 pp 241ndash257 2002

[19] S Schmid B Zierl and H Bugmann ldquoAnalyzing the carbondynamics of central European forests comparison of Biome-BGC simulations with measurementsrdquo Regional EnvironmentalChange vol 6 no 4 pp 167ndash180 2006

[20] Z Xu and G Chen ldquoFingerprinting global climate changeand forest management within rhizosphere carbon and nutri-ent cycling processesrdquo Environmental Science and PollutionResearch vol 13 no 5 pp 293ndash298 2006


[21] R JandlM Lindner L Vesterdal et al ldquoHow strongly can forestmanagement influence soil carbon sequestrationrdquo Geodermavol 137 no 3-4 pp 253ndash268 2007

[22] S P Price M A Bradford and M S Ashton ldquoCharacterizingorganic carbon stocks and flows in forest soilsrdquo in ManagingForest Carbon in a Changing Climate M S Ashton M LTyrrell D Spalding and B Gentry Eds pp 7ndash30 SpringerNew York NY USA 2012

[23] W M Post R C Izaurralde L K Mann and N BlissldquoMonitoring and verifying changes of organic carbon in soilrdquoClimatic Change vol 51 no 1 pp 73ndash99 2001

[24] P S Homann B T Bormann and J R Boyle ldquoDetectingtreatment differences in soil carbon and nitrogen resulting fromforest manipulationsrdquo Soil Science Society of America Journalvol 65 no 2 pp 463ndash469 2001

[25] K M Carney B A Hungate B G Drake and J P MegonigalldquoAltered soil microbial community at elevated CO

2leads to loss

of soil carbonrdquo Proceedings of the National Academy of Sciencesof the United States of America vol 104 no 12 pp 4990ndash49952007

[26] N P A Saby P H Bellamy X Morvan et al ldquoWill Europeansoil-monitoring networks be able to detect changes in topsoilorganic carbon contentrdquo Global Change Biology vol 14 no 10pp 2432ndash2442 2008

[27] D W Johnson and P S Curtis ldquoEffects of forest managementon soil C and N storage meta analysisrdquo Forest Ecology andManagement vol 140 no 2-3 pp 227ndash238 2001

[28] C L Goodale M J Apps R A Birdsey et al ldquoForest carbonsinks in the Northern Hemisphererdquo Ecological Applications vol12 no 3 pp 891ndash899 2002

[29] C C Barford S C Wofsy J W Munger et al ldquoFactorscontrolling long- and short-term sequestration of atmosphericCO2in a mid-latitude forestrdquo Science vol 294 no 5547 pp

1688ndash1691 2001[30] W W Covington ldquoChanges in forest floor organic-matter and

nutrient content following clear cutting in northern hard-woodsrdquo Ecology vol 62 no 1 pp 41ndash48 1981

[31] K G Harrison W M Post and D D Richter ldquoSoil carbonturnover in a recovering temperate forestrdquo Global Biogeochemi-cal Cycles vol 9 no 4 pp 449ndash454 1995

[32] R D Yanai M A Arthur T G Siccama and C A FedererldquoChallenges of measuring forest floor organic matter dynamicsrepeated measures from a chronosequencerdquo Forest Ecology andManagement vol 138 no 1ndash3 pp 273ndash283 2000

[33] T A Black and J W Harden ldquoEffect of timber harvest onsoil carbon storage at Blodgett experimental forest CaliforniardquoCanadian Journal of Forest Research vol 25 no 8 pp 1385ndash1396 1995

[34] J Shan L A Morris and R L Hendrick ldquoThe effects ofmanagement on soil and plant carbon sequestration in slashpine plantationsrdquo Journal of Applied Ecology vol 38 no 5 pp932ndash941 2001

[35] R Lal ldquoForest soils and carbon sequestrationrdquo Forest Ecologyand Management vol 220 no 1ndash3 pp 242ndash258 2005

[36] K G Mattson and W T Swank ldquoSoil and detrital carbondynamics following forest cutting in the Southern Appalachi-ansrdquo Biology and Fertility of Soils vol 7 no 3 pp 247ndash253 1989

[37] R D Yanai S V Stehman M A Arthur et al ldquoDetectingchange in forest floor carbonrdquo Soil Science Society of AmericaJournal vol 67 no 5 pp 1583ndash1593 2003

[38] K Johnson F N Scatena and Y Pan ldquoShort- and long-term responses of total soil organic carbon to harvesting in anorthern hardwood forestrdquo Forest Ecology and Managementvol 259 no 7 pp 1262ndash1267 2010

[39] D D Richter D Markewitz S E Trumbore and C G WellsldquoRapid accumulation and turnover of soil carbon in a re-establishing forestrdquo Nature vol 400 no 6739 pp 56ndash58 1999

[40] R A Houghton ldquoRevised estimates of the annual net fluxof carbon to the atmosphere from changes in land use andland management 1850-2000rdquo Tellus B Chemical and PhysicalMeteorology vol 55 no 2 pp 378ndash390 2003

[41] L E Nave E D Vance C W Swanston and P S CurtisldquoHarvest impacts on soil carbon storage in temperate forestsrdquoForest Ecology and Management vol 259 no 5 pp 857ndash8662010

[42] P J Hanson N T Edwards C T Garten and J A AndrewsldquoSeparating root and soil microbial contributions to soil respi-ration a review of methods and observationsrdquo Biogeochemistryvol 48 no 1 pp 115ndash146 2000

[43] T W Berger E Inselsbacher and S Zechmeister-BoltensternldquoCarbondioxide emissions of soils under pure andmixed standsof beech and spruce affected by decomposing foliage littermixturesrdquo Soil Biology and Biochemistry vol 42 no 6 pp 986ndash997 2010

[44] E A Davidson K Savage P Bolstad et al ldquoBelowgroundcarbon allocation in forests estimated from litterfall and IRGA-based soil respiration measurementsrdquo Agricultural and ForestMeteorology vol 113 no 1ndash4 pp 39ndash51 2002

[45] G D Mroz M J Jurgensen and D J Frederick ldquoSoil nutrientchanges following whole tree harvesting on three northernhardwood sitesrdquo Soil Science Society of America Journal vol 49no 6 pp 1552ndash1557 1985

[46] W H Schlesinger and J Lichter ldquoLimited carbon storage insoil and litter of experimental forest plots under increasedatmospheric CO

2rdquoNature vol 411 no 6836 pp 466ndash469 2001

[47] C Hoover and S Stout ldquoThe carbon consequences of thinningtechniques stand structure makes a differencerdquo Journal ofForestry vol 105 no 5 pp 266ndash270 2007

[48] N A Scott D Y Hollinger E A Davidson C A Rodriguesand D B Dail ldquoImpact of a shelterwood harvest on thenet carbon balance of a sprucehemlock dominated forest inMainerdquo in Proceedings of the New England Society of AmericanForesters 85thWinter Meeting L S T Kenefic and J Mark EdsUS Department of Agriculture Forest Service NortheasternResearch Station Newtown Square Pa USA 2005

[49] J E Smith L S Heath K E Skog and R A Birdsey ldquoMethodsfor calculating forest ecosystem and harvested carbon withstandard estimates for forest types in theUnited Statesrdquo GeneralTechnical Report NE-343 US Department of Agriculture For-est Service Northeastern Research Station Newtown SquarePa USA 2006

[50] J A Gore and W A Patterson ldquoMass of downed wood innorthern hardwood forests in NewHampshire potential effectsof forestmanagementrdquoCanadian Journal of Forest Research vol16 no 2 pp 335ndash339 1986

[51] W H Liu D M Bryant L R Hutyra et al ldquoWoody debriscontribution to the carbon budget of selectively logged andmaturing mid-latitude forestsrdquo Oecologia vol 148 no 1 pp108ndash117 2006

[52] W S Keeton ldquoManaging for late-successionalold-growth char-acteristics in northern hardwood-conifer forestsrdquoForest Ecologyand Management vol 235 no 1ndash3 pp 129ndash142 2006


[53] D M Smith B C Larson M J Kelty and P M S AshtonThe Practice of Silviculture Applied Forest Ecology John Wileyamp Sons New York NY USA 1997

[54] B R Frey M S Ashton J J McKenna D Ellum and AFinkral ldquoTopographic and temporal patterns in tree seedlingestablishment growth and survival among masting speciesof southern New England mixed-deciduous forestsrdquo ForestEcology and Management vol 245 no 1ndash3 pp 54ndash63 2007

[55] W H Meyer and B Plusnin ldquoThe Yale forest in tolland andWindham countiesrdquo Yale School of Forestry and EnvironmentalStudies Bulletin 55 Yale School of Forestry and EnvironmentalStudies New Haven Conn USA 1945

[56] M S Ashton and B C Larson ldquoGermination and seedlinggrowth of Quercus (section Erythrobalanus) across openingsin a mixed-deciduous forest of southern New England USArdquoForest Ecology andManagement vol 80 no 1ndash3 pp 81ndash94 1996

[57] A Roberts Soil Survey of Windham County United StatesDepartment of Agriculture Soil Conservation Service Con-necticut Conn USA 1981

[58] A Kulmatiski D J Vogt T G Siccama and K H BeardldquoDetecting nutrient pool changes in rocky forest soilsrdquo SoilScience Society of America Journal vol 67 no 4 pp 1282ndash12862003

[59] J S Bhatti M J Apps and C Tarnocai ldquoEstimates of soilorganic carbon stocks in central Canada using three differentapproachesrdquo Canadian Journal of Forest Research vol 32 no 5pp 805ndash812 2002

[60] R Lenth ldquoPower and sample-size pagerdquo University of IowaDepartment of Statistics and Actuarial Science 2010 httpwwwstatuiowaedusimrlenthPower

[61] M S Strickland J L Devore J C Maerz and M A BradfordldquoGrass invasion of a hardwood forest is associated with declinesin belowground carbon poolsrdquo Global Change Biology vol 16no 4 pp 1338ndash1350 2010

[62] T E Avery and H E Burkhart Forest Measurements McGrawHill New York NY USA 4th edition 1994

[63] P H Bellamy P J Loveland R I Bradley R M Lark and G JD Kirk ldquoCarbon losses from all soils across England andWales1978-2003rdquo Nature vol 437 no 7056 pp 245ndash248 2005

[64] M A Bradford N Fierer and J F Reynolds ldquoSoil carbon stocksin experimental mesocosms are dependent on the rate of labilecarbon nitrogen and phosphorus inputs to soilsrdquo FunctionalEcology vol 22 no 6 pp 964ndash974 2008

[65] S Brown D Shoch T Pearson and M Delaney Methodsfor Measuring and Monitoring Forestry Carbon Projects inCalifornia Winrock International Arlington Va USA 2004

[66] K G MacDicken A Guide to Monitoring Carbon Storagein Forestry and Agroforestry Projects Winrock InternationalArlington Va USA 1997

[67] R A Birdsey Carbon Storage and Accumulation in UnitedStates Forest Ecosystems US Department of Agriculture ForestService Washington DC USA 1992

[68] J E Smith L S Heath and P B Woodbury ldquoHow to estimateforest carbon for large areas from inventory datardquo Journal ofForestry vol 102 no 5 pp 25ndash31 2004

[69] K L Waddell ldquoSampling coarse woody debris for multipleattributes in extensive resource inventoriesrdquo Ecological Indica-tors vol 1 no 3 pp 139ndash153 2001

[70] The R Development Core Team R A Language and Environ-ment for Statistical Computing The R Foundation for StatisticalComputing Vienna Austria 2011 httpwwwR-projectorg

[71] A Kulmatiski D J Vogt T G Siccama et al ldquoLandscapedeterminants of soil carbon and nitrogen storage in southernNew Englandrdquo Soil Science Society of America Journal vol 68no 6 pp 2014ndash2022 2004

[72] C Perie and R Ouimet ldquoOrganic carbon organic matter andbulk density relationships in boreal forest soilsrdquo CanadianJournal of Soil Science vol 88 no 3 pp 315ndash325 2008

Submit your manuscripts athttpwwwhindawicom

Forestry ResearchInternational Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Environmental and Public Health

Journal of



EcosystemsJournal of


MeteorologyAdvances in

EcologyInternational Journal of


Marine BiologyJournal of


Hindawi Publishing Corporationhttpwwwhindawicom

Applied ampEnvironmentalSoil Science

Volume 2014

Advances in


Environmental Chemistry

Atmospheric SciencesInternational Journal of



Waste ManagementJournal of

Hindawi Publishing Corporation httpwwwhindawicom Volume 2014

International Journal of

Geophysics


Geological ResearchJournal of

EarthquakesJournal of


BiodiversityInternational Journal of


ScientificaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

OceanographyInternational Journal of


The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Journal of Computational Environmental SciencesHindawi Publishing Corporationhttpwwwhindawicom Volume 2014


ClimatologyJournal of


et al [38] found a consistent drop in soil carbon after a timberharvest for a northern hardwood forest in New EnglandCarbon accretion in soils of regenerating forests may befurther limited by soil texture and other environmentalconditions [39] Information on second-growth hardwoodforests is particularly important because conclusions fromstudies on carbon sink quantification conflict [22]

Higher rates of carbon loss occur in the forest floor morethan in the soil for hardwood forests [40 41] The reasonsfor changes in soil carbon relate strongly to litter inputsdecomposition and respiration [42 43] Studies suggest thatnet soil carbon storage may be influenced by small changesin climate and soil condition [41 44] Over time and underthe colder wetter climate conditions the shift of live biomassinto woody debris litter and soil humus may build upstored carbon in soils [7] Studies conflict over whether forestfloor and soil carbon increase or decrease after harvest andit is unclear whether observed changes are more stronglycorrelated with the amount of debris left on-site or the changein efflux and respiration with disturbance of the soil [36 45]Other studies have shown weak ability to correlate litterfalland belowground carbon allocation in mature temperatehardwood forests [44] and weak conversion of litter carboninto long-term forest soil carbon storage [46] Not only arethe dynamics of decomposition from coarse woody debrisand litter to soil carbon quantity poorly understood butalso much remains to be researched about the long-termdynamics of soil carbon after forest harvest

There are few studies which have examined the effectsof silvicultural treatments (regeneration or thinning) on soilcarbon stocks in the northeastern region of the United States[30 47] Some of the limitations of these studies are thatthey looked only at differences between thinnings [47] weremodeled [38] or were focused on coniferous forests ratherthan the predominant hardwood type [48] Oak-hardwoodforests of the northeastern US have shown reduced soilcarbon stocks reduced carbon in litter and increased woodydebris with harvest [49] but this study was after clearcutharvests rather than shelterwood In another study coarsewoody debris was reduced by 17 in the long term by aselective timber harvest [50] but such harvests have showndramatic increases in coarse woody debris in the short term[51] Prescriptions that purposefully leave deadwood as snagsand slash have been shown to negate any loss of coarse woodydebris [52]

Our study had two objectivesWe wanted to first quantifythe change in carbon for the uppermost layers of soiland surface horizons (coarse woody debris and litter) aftershelterwood regeneration harvests in a second-growth oak-hardwood forest Second we wanted to document the changein soil carbon coarse woody debris and litter over timeafter shelterwood harvests We sought to understand howsoil carbon is affected by forest harvest and how surfacecarbon is partitioned between coarse woody debris and litterdecomposition over a 20-year time horizon after harvest Ourstudy focuses on the most common type of forest regenera-tion harvest for oak forests in the northeastern United Statesshelterwoods a method of harvest that retains a limitedcanopy of masting nut trees (oak hickory and beech) to

provide seed for dispersal and germination and shelter fornewly established seedlings [53 54]We expected that surfacecarbon dynamics will show significant declines similar tofindings by Covington [30] and to US Forest Service datafrom clearcuts We also expected rate and amount of declineto be lower than in a clearcut because of the moderatingshade from tree retention in shelterwood systems Lastly wepredicted that the surface carbon pools will decline with themost significant change in the surface litter

2 Methods

21 Site Description of the Study Area This study was con-ducted at the 3173-ha Yale-Myers Forest near Eastford CT inthe northeastern United States (41∘571015840N 72∘071015840W)The Yale-Myers Forest is a research and demonstration forest managedby the School of Forestry and Environmental Studies andowned by Yale University The forest overstory is predomi-nantly oak-maple-pine The climate is characterized as cooltemperate and humid Mean annual summer temperature is20∘C and mean annual winter temperature is ndash4∘C Precipi-tation is evenly distributed throughout the year with a meanof 110 cm [55 56] This is a moist temperate forest underlainby a metamorphic bedrock of schistgneiss and overlain withsoils that originated from till and fluvial sediments of the lastWisconsin glacial period 20000 bp [55]The soils aremostlycoarse-loamy mesic Typic Dystrudepts [57] The topographyis characterized by undulating hills with broad ridges andnarrow valleysThe elevation ranges froma lowpoint of 170mabove sea level to a high point of 300m above sea level [56]

Starting in the early 18th century the forest as with manyareas of New England had a long period of agriculturaluse with land used for various kinds of improved andunimproved pasture tilled for crops or relict forest patchesthat were used as woodlots Since the 1820s this land haslargely been reforested naturally [55]

22 Treatment Description All stands for our study wereon glacial till soils of ablation origin and comprising theCharlton-Chatfield series [57] This soil series is a dominantone for the oak-hardwood of southern New England makingour study results generalizable to a much larger regionthan the forest All study stands were originally cleared formarginal pasture but were never plowed After abandonmentin the 1850s these sites transformed to eastern white pine(Pinus strobus L) and were then cutover in the early 1900s forthe pine boxwood industry After cutting the pine understoryhardwood regeneration comprising oaks hickories birchesand maples was released Today this second-growth hard-wood forest varies in age between 90 and 110 years of ageThis age class and forest history are very representative of thesouthern New England area

Using available GIS maps of soil type and harvests withinYale-Myers a map of shelterwood harvests and reservesof similar land use history (as described above) overlainwith the Charlton-Chatfield soil type was created to identifysites that had suitable treated and control plots adjacent toeach other (Figure 1) Controls were therefore selected as



0 05 1 2 (km)










C(Mgha)


times 100 times

10000m2

1 hatimes

1Mg1000000 g

(1)



Litter sample

3m

15 cm times 15 cm





C (g)area (m2)



(2)


3 Results






Control Harvested


Treatment

0

2000

4000

6000

8000

10000

Carb

on (g

m2)

Pa1

b1


(a)



0

2000

4000

6000

8000

10000

Carb

on (g

m2)

P = 0225

a2 a2

(b)



0

2000

4000

6000

8000

10000

Carb

on (g

m2)

P

a3

b3


(c)



4 Discussion









0 5 10 15 200

5000

10000

15000



Carb

on (g

m2)

minusControl R2=

minusHarvested R2=

00221 P = 0791

00042 P = 0404

(a)

0

500

1000

1500


Carb

on (g

m2)


minusControl R2=

Harvested R2=

00151 P = 0551

00017 P = 0294

(b)

0

2000

4000

6000

8000

10000


Carb

on (g

m2)


Control R2=

Harvested R2=

00065 P = 0264

00932 P = 0006

(c)

05000

10000150002000025000


Carb

on (g

m2)


Control R2= minus

Harvested R2=

00238 P = 0991

00997 P = 0004

(d)







5 Conclusions




Acknowledgments


References






























2leads to loss
























































Journal of












Volume 2014

Advances in









Geophysics
















0 05 1 2 (km)










C(Mgha)


times 100 times

10000m2

1 hatimes

1Mg1000000 g

(1)



Litter sample

3m

15 cm times 15 cm





C (g)area (m2)



(2)


3 Results






Control Harvested


Treatment

0

2000

4000

6000

8000

10000

Carb

on (g

m2)

Pa1

b1


(a)



0

2000

4000

6000

8000

10000

Carb

on (g

m2)

P = 0225

a2 a2

(b)



0

2000

4000

6000

8000

10000

Carb

on (g

m2)

P

a3

b3


(c)



4 Discussion









0 5 10 15 200

5000

10000

15000



Carb

on (g

m2)

minusControl R2=

minusHarvested R2=

00221 P = 0791

00042 P = 0404

(a)

0

500

1000

1500


Carb

on (g

m2)


minusControl R2=

Harvested R2=

00151 P = 0551

00017 P = 0294

(b)

0

2000

4000

6000

8000

10000


Carb

on (g

m2)


Control R2=

Harvested R2=

00065 P = 0264

00932 P = 0006

(c)

05000

10000150002000025000


Carb

on (g

m2)


Control R2= minus

Harvested R2=

00238 P = 0991

00997 P = 0004

(d)







5 Conclusions




Acknowledgments


References






























2leads to loss
























































Journal of












Volume 2014

Advances in









Geophysics















Litter sample

3m

15 cm times 15 cm





C (g)area (m2)



(2)


3 Results






Control Harvested


Treatment

0

2000

4000

6000

8000

10000

Carb

on (g

m2)

Pa1

b1


(a)



0

2000

4000

6000

8000

10000

Carb

on (g

m2)

P = 0225

a2 a2

(b)



0

2000

4000

6000

8000

10000

Carb

on (g

m2)

P

a3

b3


(c)



4 Discussion









0 5 10 15 200

5000

10000

15000



Carb

on (g

m2)

minusControl R2=

minusHarvested R2=

00221 P = 0791

00042 P = 0404

(a)

0

500

1000

1500


Carb

on (g

m2)


minusControl R2=

Harvested R2=

00151 P = 0551

00017 P = 0294

(b)

0

2000

4000

6000

8000

10000


Carb

on (g

m2)


Control R2=

Harvested R2=

00065 P = 0264

00932 P = 0006

(c)

05000

10000150002000025000


Carb

on (g

m2)


Control R2= minus

Harvested R2=

00238 P = 0991

00997 P = 0004

(d)







5 Conclusions




Acknowledgments


References






























2leads to loss
























































Journal of












Volume 2014

Advances in









Geophysics















Control Harvested


Treatment

0

2000

4000

6000

8000

10000

Carb

on (g

m2)

Pa1

b1


(a)



0

2000

4000

6000

8000

10000

Carb

on (g

m2)

P = 0225

a2 a2

(b)



0

2000

4000

6000

8000

10000

Carb

on (g

m2)

P

a3

b3


(c)



4 Discussion









0 5 10 15 200

5000

10000

15000



Carb

on (g

m2)

minusControl R2=

minusHarvested R2=

00221 P = 0791

00042 P = 0404

(a)

0

500

1000

1500


Carb

on (g

m2)


minusControl R2=

Harvested R2=

00151 P = 0551

00017 P = 0294

(b)

0

2000

4000

6000

8000

10000


Carb

on (g

m2)


Control R2=

Harvested R2=

00065 P = 0264

00932 P = 0006

(c)

05000

10000150002000025000


Carb

on (g

m2)


Control R2= minus

Harvested R2=

00238 P = 0991

00997 P = 0004

(d)







5 Conclusions




Acknowledgments


References






























2leads to loss
























































Journal of












Volume 2014

Advances in









Geophysics




















0 5 10 15 200

5000

10000

15000



Carb

on (g

m2)

minusControl R2=

minusHarvested R2=

00221 P = 0791

00042 P = 0404

(a)

0

500

1000

1500


Carb

on (g

m2)


minusControl R2=

Harvested R2=

00151 P = 0551

00017 P = 0294

(b)

0

2000

4000

6000

8000

10000


Carb

on (g

m2)


Control R2=

Harvested R2=

00065 P = 0264

00932 P = 0006

(c)

05000

10000150002000025000


Carb

on (g

m2)


Control R2= minus

Harvested R2=

00238 P = 0991

00997 P = 0004

(d)







5 Conclusions




Acknowledgments


References






























2leads to loss
























































Journal of












Volume 2014

Advances in









Geophysics

















5 Conclusions




Acknowledgments


References






























2leads to loss
























































Journal of












Volume 2014

Advances in









Geophysics




















2leads to loss
























































Journal of












Volume 2014

Advances in









Geophysics







































Journal of












Volume 2014

Advances in









Geophysics














research article change in soil and forest floor carbon...

Documents