Investigating the soil acid–base status in managed boreal forests using the SAFE model

Please download to get full document.

View again

of 21
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Investigating the soil acid–base status in managed boreal forests using the SAFE model
  ecological modelling 206 (2007) 301–321 available at journal homepage: Investigating the soil acid–base status in managed boreal forests using the SAFE model Evelyne Thiffault a , Nicolas B´ elanger b , ∗ , David Par´ e c ,William H. Hendershot d , Alison Munson a a Centre d’´ etude de la forˆ et, Universit´ e Laval, Qu´ ebec, Que., Canada G1K 7P4 b Department of Soil Science, University of Saskatchewan, 51 Campus Drive, Saskatoon, Sask., Canada S7N 5A8 c Natural Resources Canada, Canadian Forest Service, Laurentian Forestry Centre, 1055 du P.E.P.S.,P.O. Box 10380, Succ. Sainte-Foy, Qu´ ebec, Que., Canada G1V 4C7 d Department of Natural Resource Sciences, Macdonald-Stewart Building,21111 Lakeshore Road, Ste-Anne de Bellevue, Que., Canada H9X 3V9 a r t i c l e i n f o  Article history: Received 10 November 2006Received in revised form9 March 2007Accepted 28 March 2007Published on line 30 May 2007 Keywords: Soil acidificationBoreal forestStem-only harvesting Whole-tree harvesting Natural disturbancesAcid depositionBase cationsDynamic modelling SAFE a b s t r a c t Adynamicsoilchemistrymodel,SAFE,wasappliedtoborealstandslocatedinthreeregionsof Quebec (Canada) to investigate the relative contributions of forest harvesting methods(whole-tree and stem-only harvesting), natural disturbances (wildfire and spruce budwormoutbreak) and atmospheric deposition on base cation fluxes and the soil acid–base status.Model parameterization was done using measured and published soil and vegetation data,published trends of nutrient dynamics, governmental databases on atmospheric deposi-tion and climate, the reconstruction method MAKEDEP and throughfall equations that wedevelopedforconiferousstands.Themodelsuggestedthatforestdisturbances,bothanthro-pogenicandnatural,increasedbasesaturation,andthusbasecationavailability,todifferentdegreesforperiodsofonetofivedecades.However,suchdisturbances,nomattertheirinter-valofreturnandtheirintensity,didnotappeartobethemaindrivingforceofsoilchemistryinthelongterm.Ontheotherhand,simulationresultsshowedthattheBhorizonatallsiteshas undergone acidification due to acid atmospheric deposition during the 20th century.Results suggest that atmospheric deposition is the main driver of long-term trends of soilchemistryforthestudysitesandthattheissueofsustainabilityofsoilproductivitydependsmore on air pollution abatement policies than on forest management strategies.© 2007 Elsevier B.V. All rights reserved. 1. Introduction Biogeochemical cycling in forest ecosystems is partly con-trolled by internal factors such as stand development andspecies composition (Alban, 1982; Miller, 1995; Binkley and Giardina, 1998; B´elanger et al., 2004a). Boreal forests of north-eastern Canada are also driven by periodic stand-replacing disturbancessuchasfires,windstorms,insectsandpathogens ∗ Corresponding author . Tel.: +1 306 966 6841; fax: +1 306 966 6881.E-mail address: (N. B´elanger). that alter nutrient fluxes over various periods of time (e.g.Sprugel, 1984; Yorks et al., 1999; Lamontagne et al., 2000;Hunter, 2001). To those natural influences that have shapedborealecosystemsfunctioningforthousandsofyears,humanactionintheformofatmosphericpollutionandintensivehar-vesting has been added in the past century.Many studies have documented the depletion of soil basecation reserves from forest soils of Europe and North America 0304-3800/$ – see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.ecolmodel.2007.03.044  302  ecological modelling 206 (2007) 301–321 overthepastseveraldecades,attributedtoacidicatmosphericdeposition (e.g. Matzner and Murach, 1995; Likens et al., 1996;Driscoll et al., 2001). This depletion reflects the buffering of atmospheric inputs of acidity by dissolution of aluminium(Al) and soil cation exchange reactions; base cations are dis-placed from exchange sites by hydrogen (H) and Al ions, andreadily leached out of the soil profile (Galloway et al., 1983). Declinesinforestsoilexchangeablebasecationpoolsmayalsohave been worsened by decreased atmospheric inputs of basecationsinthelastthirdofthe20thcentury(Hedinetal.,1994). Base-poorforestecosystemsoftheCanadianShieldhavebeendeemed extremely sensitive to acidic deposition and at riskof further base cation depletion if harvested (Watmough etal., 2005; Duchesne and Houle, 2006). Moreover, theoretical geochemical budgets calculated by Par´e et al. (2002) for theboreal forest of Quebec showed that clearcutting, especiallywhen done using an intensive method like whole-tree har-vesting,canleadtoanegativebalanceofbasecations(outputs> inputs) for stands of nutrient-demanding species like bal-sam fir (  Abies balsamea  (L.) Mill.) and trembling aspen ( Populustremuloides  Michx.) located on shallow or coarse sandy soils.On the other hand, Yanai et al. (1999) underlined the dif-ficulty in distinguishing the relative effects of natural forestsuccession and acidic deposition on base cation cycling, evenwith extensive long-term field studies. Also, Thiffault et al.(2006) have concluded that stem-only harvesting, when com-pared with whole-tree harvesting, results in a base cationloading that lasts at least 15–20 years due to the presenceof debris left during the logging operations in boreal conifer-ous stands of Quebec. This loading is however dwarfed whencompared with that associated with wildfire (Thiffault et al.,in press). Additionally, B´elanger et al. (2003) estimated thateven though whole-tree harvesting does represent a loss of soil acid neutralizing capacity (ANC) in black spruce stands of Quebec, this loss is small with respect to the loss of alkalinityassociated with atmospheric deposition.Clearly, there is a need to quantify base cation fluxes,sources and sinks in the boreal forest so that we can assessthe contributions of the various drivers of soil base cationavailability and soil acidification. By doing so, the best mit-igation/prevention strategies could be adopted, either at theforestmanagementleveloratthepollutantemissionscontrollevel.Long-term field data of fluxes and changes in pools may bethe best way of assessing the relative importance of ecosys-tem processes to long-term ecosystem changes ( Johnson andTodd, 1990). However, ecosystem simulation represents aninteresting proxy in the absence of such studies (Wei et al.,2003).Dynamicmodellingprovidesanopportunitytoincreaseourknowledgeoftheinteractionsbetweenbiological,edaphicandchemicalcomponentsoftheforestsystemwhileaddress-ing the issue of time scale in ecosystem response (Kros andWarfvinge, 1995). The soil chemistry model SAFE (Soil Acidifi-cation in Forested Ecosystems) has been developed to providea basic understanding of the main fluxes of ions in borealecosystems, the processes involved and the links betweenthem (Warfvinge et al., 1993). In this paper, we present the results of a modelling exercise aimed at calibrating SAFE inorder to theoretically explore the relative contribution of har-vesting methods (stem-only harvesting: SOH and whole-treeharvesting: WTH), natural disturbances and atmospheric pol-lution to the long-term dynamic pattern of the soil acid–basestatus for various boreal coniferous stands of Quebec. Differ-ent scenarios were used in this exercise, involving three siteswithcontrastingsoilandvegetationcharacteristics.Foroneof these sites, a range of scenarios was also simulated to explorethe influence of harvested species, growth rates, mineraliza-tion rates and canopy structure on soil chemistry. 2. Materials and methods 2.1. Model description SAFE is a dynamic, multi-layer, process-oriented soil modelthatcalculatessoilwaterchemistry,cationexchangereactionsand weathering rates from physical and chemical soil datacoupled with a schematic description of atmospheric deposi-tion as well as nutrient uptake and cycling.SAFE includes mathematical descriptions of weathering of soil minerals, cation exchange reactions, leaching andadsorption of dissolved chemical components, and solutionequilibriumreactions.SAFEalsoincludesthefollowingfluxes,specifiedastime-seriesfiles:(1)atmosphericdepositionofCa,Mg, K, Na, NO 3 , NH 4 , SO 4  and Cl; (2) litterfall of Ca, Mg, K andN;(3)canopyexchangeofCa,Mg,KandN;(4)plantnetuptakeofCa,Mg,KandN;(5)netmineralizationofCa,Mg,K,SandN.Each soil horizon is assumed to be physically and chemicallyhomogeneous. The processes interact only via the soil solu-tion. Some processes not included in the model are sulphateadsorptionandtheorganiccomplexationofmetalssuchasAl.Also, dissolved organic matter is not modelled by the rate of mineralization of organic matter but rather specified as inputdata.Changes in soil acidity are expressed as acid neutralizing capacity (ANC), which is modelled by calculating the massbalance of the various processes contributing or consuming acidity. The buffering of the liquid phase is controlled by thecarbonateequilibriumreactions,thedissolution-precipitationreactions of a solid gibbsite phase which produce variouscharged Al-OH complexes, and the acid–base reactions of amonovalent organic acid (RH):[ANC]  =  [OH − ]  +  [R − ]  +  [HCO 3 − ]  +  2[CO 32 − ] − [H + ] − 3[Al 3 + ] − 2[Al(OH) 2 + ] − [Al(OH) 2 + ]Calcium,MgandKarelumpedtogetherintoonedivalentbasecation component (BC). The Gapon principle is used for con-trolling cation exchange reactions. The rate of exchange islimited by mass transfer of BC, and is proportional to the dif-ference in BC concentration between the exchange complexand the soil solution (Warfvinge et al., 1993). The relationship between Al and H is constrained by gibbsite equilibrium andthuscationexchangeinSAFEisanexchangereactionbetweenBC and acidity. Many studies (e.g. Mulder et al., 1989) have shownthatAlsolubilityiscontrolledbysolid-phaseorganicAlcomplexes rather than some forms of Al hydroxides. It shouldthus be noted that the primary reason for using the apparentgibbsite solubility is that reasonable titration curves can bederived with this model and easily implemented in simula-  ecological modelling 206 (2007) 301–321  303 tion models. According to Alveteg (1998), present knowledgeregarding Al biogeochemistry is too limited to allow the inclu-sion of a widely applicable mechanistic Al submodel based onAl-organic complexation in dynamic simulations.Cationic exchange capacity (CEC) is treated as a constant.This should also be considered a simplification since studieshave shown that disturbances such as clearcut and wildfirecanhaveasignificantinfluenceonCEC(e.g.BockandVanRees,2002; Glaser et al., 2002).A full description of the SAFE model can be found inWarfvinge et al. (1993). A sensitivity test was performed byWarfvinge and Sand´en (1992) on fixed/permanent soil vari-ables; soil bulk density and specific surface area were foundtobethemostsensitiveparametersonthesimulatedacidneu-tralizingcapacity(ANC),followedbysoilmoisturecontentandtemperature, and then by CO 2  partial pressure and CEC. 2.2. Site description Three sites located in regions typical of the balsamfir—white birch bioclimatic domain in the boreal zone of Quebec were chosen for the simulations: Haute-Mauricie(47 ◦ 45 ′ –47 ◦ 55 ′ N, 74 ◦ 07 ′ –74 ◦ 15 ′ W), Forˆ et Montmorency(47 ◦ 17 ′ –47 ◦ 21 ′ N,71 ◦ 09–71 ◦ 08 ′ W)andGasp´esie(48 ◦ 24 ′ –48 ◦ 30 ′ N,66 ◦ 16 ′ –66 ◦ 21 ′ W).Haute-Mauricieislocatedinthewesternpartofthebioclimaticdomainandisdominatedbyjackpine( Pinusbanksiana  Lamb.) and black spruce ( Picea mariana  (Mill) B.S.P.)stands; the main natural disturbance in this area is wildfire.For ˆ et Montmorency and Gasp´esie are located in the easternportion and are dominated by mixed stands of balsam fir andwhite birch; spruce budworm outbreaks are the dominantnatural disturbance events. Sites in Haute-Mauricie and For ˆ etMontmorency lie on granite or granitic gneisses typical of theCanadian Precambrian Shield, whereas the Gasp´esie site islocated on soils developed from Appalachian schists. Soils inGasp´esie are typic Haplohumods developed from sandy loamtill deposits (Soil Survey Staff, 2003). Both Haute-Mauricie andFor ˆ et Montmorency soils are typic Haplorthods developedrespectively from loamy sand glaciofluvial deposits andloamy sand tills (Soil Survey Staff, 2003).The modelling exercise presented here uses soil data (CEC,soilpHandbasesaturation)recentlypublishedinastudycom-paring whole-tree and stem-only harvested plots at the threesites discussed above (Thiffault et al., 2006). For this project, field experiments were established in 2001 in areas wheremature forests were clear-felled in the 1980s. Whole-tree har-vesting (WTH) is defined as the removal of all above-stumpbiomass during clearcutting, whereas stem-only harvesting (SOH) is the removal of the stem, leaving all foliage, twigs andbranches on the forest site. 2.3. Model parameterization and modelling scenarios Ten different modelling scenarios were used in this study(Table 1). Scenarios A to C described sequences of naturaldisturbances and clearcutting events in conditions typicalof the three study sites (Haute-Mauricie, For ˆ et Montmorencyand Gasp´esie, respectively). Using scenario A (Haute-Mauriciesite) as a benchmark, time-series files were then modifiedto explore the sensitivity of modelled soil chemistry to anarrayofvariablesrelatedtotheharvestingmethods:scenariosD to I assessed the impact of the canopy structure of har-vestedstand(i.e.proportionoftwigs,smallbranchesandlargebranches), the mineralization rates of harvesting debris andthe growth rate of the regenerating stand, whereas scenario J was used to investigate the influence of harvested species.Details on specific inputs used in scenarios are given in thefollowing sections.For the purpose of the simulations, the soil profile wasdivided into two layers, i.e. the forest floor and the first 20cmof the podzolic B horizon. The forest floor thickness for eachstudy site was assigned according to values published inThiffault et al. (2006), whereas the B horizon was assigned adepth of 20cm so that the simulation results could be cali-brated from the field results in Thiffault et al. (2006). A depth of20cmcorrespondsroughlytotheobservedlimitoftheroot-ing zone for the study sites. The input data for each layersimulated in each region is described in Table 2. 2.3.1. Geochemical data For each of the three sites used in the modelling scenarios, 10samples of the mineral soil were taken randomly for particle-size distribution and total chemical composition. Particle-sizedistribution was determined by sieving and by sedimentationanalysis using a laser particle sizer (Analysette 22 compact,Fritsch GmbH & Co. KG, Welden, Germany) on NaOCl-treatedsamples. Elemental composition was determined on 32-mmdiam.fusedbeadspreparedfroma1:5soil/lithiumtetraboratemixture using an automated X-ray fluorescence spectrom-eter system (Philips PW2440 4kW, Panalytical, Almelo, TheNetherlands) with a Rhodium 60-kV end window X-ray tube.The results of the elemental analyses were used to assignelements to their respective minerals using the UPPSALAmodel(SverdrupandWarfvinge,1992).Thismodelisanorma-tive back-calculation method for reconstructing the empiricalmineralogy from total digestion analysis in a rapid and easyway,andwasdevelopedfromanimportantdatabasecollectedfromthePrecambrianShieldinSweden.Itisbasedonassump-tionsofthestoichiometriccompositionofthemineralsinsoilsofgraniticsrcin.Themineralshavebeengroupedintoassem-blages of minerals with similar composition and dissolutionrates. For example, chlorite is composed of trioctahedral chlo-rite, primary illite, trioctahedral vermiculite of primary typeand biotite, whereas epidote includes all epidotes and pyrox-enes. To validate the output from the UPPSALA model, themineralogical composition of the clay fraction (<2  m) wasdetermined on a small number of samples (two from eachsite)byX-raydiffractiononMg-andK-saturatedsamplessep-arated from the sedimentation analysis and centrifugation.Oriented samples were analyzed at 25 ◦ C and after glycolationand heating treatments (550 ◦ C) using Cu K   radiation. Thepowders produced for X-ray fluorescence were also mountedas non-oriented slides and analyzed by X-ray diffraction.Specific surface area was calculated from the mean soildensity and texture with the algorithm described by Sverdrup(1990):Aw  =  100 − %CF100  × ((0 . 003 × %sand) + (0 . 022 × %silt) + (0 . 08 × %clay)) × SD × 1000   3   0  4  e  c ol  o gi   cal m ode l l i  n g2 0 6    (    2 0 0  7  )      3 01–  321 Table 1 – Modelling scenarios Scenario Site Species composition Mineralization rateof N, Ca and Mg Canopy structure Disturbance events a Stand growth rate afterharvesting  A Haute-Mauricie Jack pine Lycan Canopy I Wildfire: 1789 and 1905; harvesting: 1985and 2045Same growth for SOH andWTHB Forˆ et Montmorency Balsam fir – white birch Spruce budworm outbreak – light: 1802and 1911; severe: 1832; harvesting: 1930(SOH), 1985 and 2045Same growth for SOH andWTHC Gasp´esie Balsam fir – white birch Spruce budworm outbreak – light: 1802and 1911; severe: 1832 and 1930;harvesting: 1985 and 2045Same growth for SOH andWTHD Haute-Mauricie Jack pine Lycan Canopy I Wildfire: 1789 and 1905; harvesting: 1985and 2045Faster growth for SOHE Haute-Mauricie Jack pine Lycan Canopy II Wildfire: 1789 and 1905; harvesting: 1985and 2045Faster growth for SOHF Haute-Mauricie Jack pine Slogberget Canopy I Wildfire: 1789 and 1905; harvesting: 1985and 2045Faster growth for SOHG Haute-Mauricie Jack pine Slogberget Canopy II Wildfire: 1789 and 1905; harvesting: 1985and 2045Faster growth for SOHH Haute-Mauricie Jack pine T ¨ onnersj ¨ oheden Canopy I Wildfire: 1789 and 1905; harvesting: 1985 and 2045Faster growth for SOHI Haute-Mauricie Jack pine T ¨ onnersj ¨ oheden Canopy II Wildfire: 1789 and 1905; harvesting: 1985and 2045Faster growth for SOH J Haute-Mauricie Black spruce Lycan Canopy I Wildfire: 1789 and 1905; harvesting: 1985and 2045Same growth for SOH andWTH a For each scenario, one simulation was made using SOH and one was made using WTH for the harvesting events of 1985 and 2045.  ecological modelling 206 (2007) 301–321  305 Table 2 – Layer-specific input data for the three simulated sites Parameter Unit Haute-Mauricie For ˆ et Montmorency Gasp´esie1 2 1 2 1 2 Morph. characterization LFH B LFH B LFH BSoil layer thickness m 0.05 0.2 0.08 0.2 0.02 0.2Soil water content m 3 m − 3 0.14 0.3080 0.24 0.3008 0.1054 0.4468Soil bulk density kgm − 3 140 930 140 977 140 1100Specific surface area  × 10 − 6 m 2 m − 3 0 808,031 0 1,185,919 0 1,332,375Cation exchange capacity  × 10 6 kmol c  kg  − 1 256.8 22.3 192.1 64.3 334 194.95CO 2  pressure Times ambient 5 20 5 20 5 20Dissolved organic carbon mgL − 1 20 5 20 5 30 15Log gibbsite eq. constant kmol 2 m − 3 5.5 10 7.5 8.5 5 8.5Outflow % of throughfall 66.7 42.3 89 63 89 48BC and N uptake % of total uptake 75 25 75 25 75 25Mineral % of total matrixK-feldspar 0 12.5 0 20.6 0 15.2Plagioclase 0 12.0 0 11.2 0 0.1Albite 0 25.1 0 22.7 0 16.1Hornblende 0 12.0 0 12.4 0 0Pyroxene 0 0 0 0 0 0Epidote 0 2.0 0 1.9 0 0.1Garnet 0 0 0 0 0 0Biotite 0 0 0 2.5 0 0Muscovite 0 4.4 0 7.3 0 5.4Fe-chlorite 0 0 0 0 0 6.0Mg-vermiculite 0 0 0 0 0 0Apatite 0 0.3 0 1.51 0 0.3Kaolinite 0 0 0 0 0 1Calcite 0 0 0 0 0 0 where Aw is the exposed mineral surface area, SD is soil den-sity, CF represents coarse fragments, and the sum of clay, silt,sand and coarse fragments equals 1.Each simulated layer was assigned a CEC corresponding tothe average value for each site as measured in Thiffault et al.(2006).CO 2  pressure was not measured at the study sites. Inabsence of data on this parameter for our study sites, andto make sure it was confined to reasonable values, defaultvalues similar across sites were adapted from the suggestedparameter sets in Warfvinge et al. (1993). 2.3.2. Climatic and atmospheric data The NAtChem database (Canadian National AtmosphericChemistry Precipitation Database, Meteological Service of Canada, Environment Canada) for stations near our siteswas used to assess precipitation amounts (Parent stationlocated 35km northwest of the Haute-Mauricie site, For ˆ etMontmorency station located within our study site, St-Octavestation located 50km north of the Gasp´esie site). Soil tem-perature was set at the long-term average air temperatureas computed for the climate normals in stations establishedby Environment Canada close to our sites (3.4 ◦ C at the LaTuque station located 50km southeast from Haute-Mauriciesite, 0.3 ◦ C at the For ˆ et Montmorency station located withinthe study site, and 3.5 ◦ C at the Nouvelle station located 20kmsouthwest from Gasp´esie site).Throughfall was set at 92% of bulk precipitation (Piirainenet al., 2002). Average throughfall precipitation therefore was1.09m for Haute-Mauricie, 1.44m for For ˆ et Montmorency and1.34m for Gasp´esie. Infiltration was set equal to throughfallprecipitation (i.e. no surface overland flow). For the purposeof the simulations with SAFE, soil moisture and water per-colation through soil layers were estimated with a separatemodel, FORHYM (Arp and Yin, 1992). FORHYM (Forest Hydrol-ogy Model) is designed to estimate hydrological processes inupland forest soils from monthly climate records (tempera-ture and precipitations) and descriptive site information (e.g.layer depth, texture, stand species composition).To construct the wet deposition time-series, yearly wetdeposition ion fluxes taken from the NAtChem database(Table 3) were scaled through time according to the stan-dardcurvesofemissionsdevelopedforeasternNorthAmericaby B´elanger et al. (2002a) (Fig. 1). This set of assumptions Table 3 – Average of wet deposition ion fluxes (mmol c m − 2 year − 1  ) for the three simulated sites (data from NAtChem) Measurement period SO 42 − NO 3 − NH 4+ Ca 2+ Mg  2+ K + Na + Cl − Haute-Mauricie 1984–1998 33.0 21.7 11.8 4.6 1.1 0.7 1.4 2.4For ˆ et Montmorency 1988–1995 47.5 37.0 19.1 6.3 1.6 0.5 2.2 3.1Gasp´esie 1995–1998 29.1 13.8 6.6 5.1 3.6 1.0 14.4 18.2
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks