Loretta Gratani?Luciano Di Martino?Anna Rita Frattaroli?Andrea Bonito?Valter Di Cecco?Walter De Simone?Giorgia Ferella?Rosangela Catoni
Abstract Terrestrial ecosystems represent a major sink for atmospheric carbon(C)and temperate forests play an important role in global C cycling,contributing to lower atmospheric carbon dioxide(CO2)concentration through photosynthesis.The Intergovernmental Panel of Climate Change highlights that the forestry sector has great potential to decrease atmospheric CO2concentration compared to other sectoral mitigation activities.The aim of this study was to evaluate CO2sequestration(CO2S)capability of Fagus sylvatica(beech)growing in the Orfento Valley within Majella National Park(Abruzzo,Italy).We compared F.sylvatica areas subjected to thinning(one high-forest and one coppice)and no-management areas(two high-forests and two coppices).The results show a mean CO2S of 44.3±2.6 Mg CO2ha-1a-1,corresponding to 12.1±0.7 Mg C ha-1a-1the no-managed areas having a 28%higher value than the managed areas.The results highlight that thinning that allows seed regeneration can support traditional management practices such as civic use in some areas while no management should be carried out in the reserve in order to give priority to the objective of conservation and naturalistic improvement of the forest heritage.
Keywords Beech·Carbon sequestration·Coppice·Highforest·Leaf area index
The United Nations Framework Convention on Climate Change(UNFCCC)de fines carbon(C)sequestration as the process of removing C from the atmosphere and depositing it in a reservoir(Takimoto et al.2008).Moreover,the results of this convention led to an agreement to reduce rising levels of carbon dioxide(CO2)and other greenhouse gases(GHGs)in the atmosphere and the Kyoto protocol proposed C reduction through decreasing fossil fuel emission or accumulating C in vegetation and soil(Oelbermann et al.2004;Jha 2015;Wu et al.2017).In fact,terrestrial ecosystems have a large capacity to sequester atmospheric carbon dioxide(CO2)(Cole et al.2017;Pant and Tewari 2014)through photosynthesis and among them forests are important because they contain a relatively large and stable storage pool of atmospheric C(Pan et al.2011).A recent global C analysis estimated a terrestrial C sink in the range of 2.0–3.4 Pg(petagrams)of C per year-1and the European temperate forest sink was estimated at 0.24 Pg C year-1(Pan et al.2011).Thus,it is important to quantify C storage capacity of different tree species(Jha 2015).Among the European forests,Fagus sylvatica L.(beech)forests are widespread(Piovesan et al.2005).Beech represents the potential natural vegetation of many areas of lowlands in northwest and northcentral Europe up to southern Sweden and the mountains of central,south,and east Europe(Jahn 1991).The distribution of beech is limited by excess soil moisture in riparian floodplain forests and by drought in southern Europe(Ellenberg and Leuschner 2014)due to the high sensitivity of this species to both high and low water availability(Simon et al.2017).Beech forests are among the sites chosen for the European Research Programs EUROFLUX and CARBOEUROFLUX(Damesin et al.Damesin 2003)to monitor longterm atmospheric C fluxes and clarify C storage within components of the forests.Moreover,several old-growth beech forests in 12 European countries(i.e.Albania,Austria,Belgium,Bulgaria,Croatia,Germany,Italy,Romania,Slovakia,Slovenia,Spain and Ukraine)were recently included in the UNESCO World Heritage List(UNESCO,7 July 2017).Beech forests characterize the landscape of mountain areas in Italy from Alps to the southern regions of Campania,Basilicata,Calabria and Sicily(Nocentini 2009).The total area covered by beech in Italy is 1,042,129 ha(National Forest Inventory,INFC 2005)corresponding to 9.4%of the country’s total forested area(Nocentini 2009).On the Apennines,beech usually grows above 900–1000 m a.s.l.(Nocentini 2009)and Abruzzo is the region with the highest surface covered area(122,402 ha)corresponding to 12%of the total beech surface area in Italy(INFC 2005).
Human practices have significantly modified the distribution,composition and structure of beech forests(Nocentini2009).Although beech hascommonly been managed as high-forest(Hahn and Fanta 2001),relatively large areas have been managed as coppice for the production of firewood and charcoal,especially in the lower mountain ranges in southern Europe(Hahn and Fanta 2001;Brunet et al.2010).Forest management has an important impact on C storage(McGrath et al.2015;Granata et al.2016).In particular,thinning can modify stand traits(i.e.tree density,stand basal area,canopy cover)and the related environmental variables,in particular soil moisture and temperature,leading to forest structure variations(Aussenac 2000)that impact on C storage(Granata et al.2016).
The main objective of our research was to analyze CO2sequestration capability(referring to the carbon uptake rate over the year through photosynthesis,Beaumont et al.2014)of F.sylvatica forests in the Orfento Valley within Majella National Park(Abruzzo,Italy).In particular,we analyzed high-forests and coppices(the most common silvicultural practices for beech)subjected to different management regimes(i.e.thinning and no-thinning).We hypothesized that areas subjected to thinning had different CO2sequestration capability due to differences in stand traits,and that among the areas not subjected to thinning,the coppices had greater CO2sequestration capability related to higher tree density.We expected our study results to provide guidance for planning forest management and monitoring of this important type of forest.
The study was carried out during May–September 2015 in the Orfento Valley within Majella National Park(Abruzzo,Italy).The Orfento Valley Natural State Reserve was established in 1971,and subsequently included in Majella National Park.We analyzed six sample areas subjected to one of two silvicultural practices(i.e.high-forest and coppice)that are the most common management practices for beech(Fig.1).Moreover,to understand the effects of thinning on these two silvicultural practices,we analyzed one high-forest and one coppice that were subjected to thinning.The remaining areas are high-forests and coppices where no thinning has been carried in the last 20 years or longer.Information about the management regime was obtained from the reserve’s forest management plan and the Italian State Forestry Corps.All the considered areas were mapped by open source GIS software(QGIS)which was also used to estimate the extent of each area.Our area 1(111 ha,at 1135 m a.s.l.),area 2(93 ha,at 1130 m a.s.l.)and area 3(97 ha,at 1529 m a.s.l.)were F.sylvatica high-forests.Area 1 was subject to thinning at the time of our sampling,while area 2 had not been thinned for 30 years and area 3 for 20 years.Area 4(55 ha,at 1110 m a.s.l.),area 5(40 ha,at 1085 m a.s.l)and area 6(75 ha,1509 m a.s.l.)were coppices.Area 4 was thinned but Area 5 had no thinning for 20 years and Area 6 had not been thinned for 50 years.Soil characteristics of each area are listed in Table 1.The climate was characterized by a mean minimum air temperature(Tmin)of-3.3± 1.7°C(February),a mean maximum air temperature(Tmax)of 22.0± 1.7°C(July)and a mean annual air temperature(Tmean)of 8.2± 6.5°C.Total annual rainfall is 1663 mm.During the study period,Tmaxof the hottest month(July)was 25.6± 1.3°C and total rainfall was 93 mm(Data from the Meteorological Station of Passo Lanciano,Chieti,42°18′62′;14°09′87′for the years 2001–2015,provided by Ufficio Idrograficoe Mareografico Regione Abruzzo).
Ten plots(400 m2each)were randomly selected for each sampling area.Leaf area index(LAI)was measured using an ‘LAI 2000 Plant Canopy Analyzer’(LICOR Inc.,Lincoln,NE,USA)in July in each sample plot.We measured the following parameters for ten trees in each plot:tree diameter at breast height(DBH,cm),measured by callipers to 65 cm,and tree height(H,m)by electronic clinometer(Hagl?f,Sweden).Tree density(Dt,tree number ha-1)was recorded and the total basal area(BA,m2ha-1)was calculated for each plot.
Fig.1 Location map with the Abruzzo Region in Italy and the research area within the Orfento Valley inside the Majella National Park.Area 1=high-forest currently subject to thinning;Area 2=high-forest unmanaged for the last 30 years;Area 3=highforest unmanaged for the last 20 years;Area 4=coppice currently subject to thinning;Area 5=coppice unmanaged for the last 20 years;Area 6=coppice unmanaged for the last 50 years
Table 1 Soil characterization of the six sampling areas
Total photosynthetic leaf surface(TPS,m2)of each area was calculated by multiplying the mean LAI value by the extension of each area measured by QGIS.
Microclimate measurements were recorded monthly in each sample plot from 9:00 to 16:00 on 4 days each month with comparable weather conditions and without clouds.Relative humidity(RH,%)and air temperature(Ta,°C)were recorded at 5 min intervals by HOBO data loggers(H08-003-02,Onset HOBO Data Loggers,Cape Cod,MA,USA)at 2 m above the forest floor.Photon flux density[PFD,μmol(photons)m-2s-1]was measured by a quantum radiometer photometer(LI-189 LI-COR,USA)with a quantum sensor LI-190SA.The photon flux density at soil level below the vegetation(PFD%)was calculated as:(PFD%=PFDb/PFDo)where PFDbwas the photon flux density below the vegetation and PFDothe photon flux density in the open(Gratani et al.2015)for each area.
Gas exchange was measured monthly(3 days per months with comparable weather conditions)in the period of May–September 2015(i.e.presence of fully expanded leaves).Net photosynthetic rate[NP,μmol(CO2)m-2s-1],stomatal conductance[gs,mol(H2O)m-2s-1],leaf transpiration[E,mmol(H2O)m-2s-1]and leaf temperature(Tl,°C)were measured by an infrared gas analyzer(LCPro+,ADC,UK)equipped with a leaf chamber(PLC,Parkinson Leaf Chamber,UK).Measurements were carried out in situ on cloud-free days(PPFD > 1000 μmol m-2s-1),in the morning(from 9:00 am to 12:00 pm)to ensure that nearmaximum daily NPwere measured.On each sampling occasion,fully expanded sun leaves were sampled using a 5-m-long pruning shear(Kimm and Ryu 2015).Moreover,to make the measurements comparable,leaves with a southeast exposure were sampled.Measurements were carried out on five representative trees per sample area(three leaves per tree).During measurements,CO2within the leaf cuvette was 370–380 ppm and relative air humidity 50–65%,according to Varone et al.(2015).
The total yearly CO2sequestration capability[CO2S,Mg CO2ha-1a-1]was calculated multiplying TPS by the mean yearly net photosynthetic rate and the total yearly photosynthetic activity time(in hours),according to Gratani and Varone(2006).The coefficient of CO2/C=3.67 was used to convert the amount of sequestered CO2in equivalent C amount,according to Evrendilek et al.(2006).
Differences between means for the measured parameters were tested by one-way ANOVA and were considered significant at P<0.05 based on the results of the Tuckey test.All data were shown as mean±standard error(±S.E.).All statistic tests were performed by a statistical software package(Statistica 8.0,Stasoft,USA).Linear regression analysis was used to relate NPwith gs,and E with gs,considering separately areas at 1000 m a.s.l.(i.e.areas 1,2,4 and 5)and those above 1500 m a.s.l.(i.e.areas 3 and 6).We also used linear regression analysis to relate PDF%with LAI,and Tawith PFD%,considering all measurements carried out within each area.To understand how LAI,Dt,H,DBH and NPyearly(predictors)affected CO2S(response variable),the predictors were combined via principal component analysis(PCA)across the six sampling areas.Then,linear regression analysis was used to relate the axis explaining the largest proportion of the variance(i.e.PC1)with CO2S.
The highest Taand the lowest RH were measured in July(26.1±1.9°C and 46.8±4.0%,mean value,respectively),area 3 having the lowest Taand the highest RH(25.13±0.12°C and 54.81±0.51%,respectively).The lowest Tawas measured in September(11.7± 1.8°C,mean value)the area 6 having the lowest value(9.50± 0.85°C).The lowest PFD%was recorded in area 5(0.49±0.01%)and highest was recorded in area 4(2.35±0.20%).
Stand measurements are listed in Table 2.Dt was the highest in area 5(1800±23 tree ha-1)and lowest in area 4(450±14 trees ha-1).Greatest DBH and H were measured in area 1 (26.3±0.4 cm and 24.0±0.6 m,respectively)while lowest DBH was recorded in area 5(13.9±0.3 cm).The area 2 had the highest BA(55.2±0.8 m2ha-1) and area 4 the lowest BA(17.2±0.6 m2ha-1).LAI was highest in area 5(4.38±0.12)and lowest in area 4(3.00±0.04).
On average,NPwas highest in May[11.5 ± 0.2 μmol(CO2)m-2s-1,mean value of the six areas]and lowest in July [6.4 ± 0.1 μmol (CO2)m-2s-1, mean value](Figs.2,3a,b).Both gsand NPhad their highest rates in May[0.22±0.01 mol(H2O)m-2s-1,mean value]and their lowest in July[0.07±0.01 mol(H2O)m-2s-1,mean value].The opposite trend was observed for E,with highest rates in July[2.52±0.12 mmol(H2O)m-2s-1,mean value of the six areas]and lowest rates in May[1.91±0.03 mmol(H2O)m-2s-1,mean value].
Mean yearly NPwas highest in area 3 and in area 6[10.5 ± 0.1 μmol(CO2)m-2s-1,mean value of the study period]decreasing by 10%in areas 1,2,4 and 5(mean value).
Highest CO2S was recorded in area 6 [49.8 ±0.67 Mg CO2ha-1a-1],corresponding to 13.6±0.18 Mg C ha-1a-1while lowest CO2S was recorded in area 4[33.1±0.44 Mg CO2ha-1a-1],corresponding to 8.90±0.12 Mg C ha-1a-1(Fig.4).Mean CO2S for all areas was 44.3±2.6 Mg CO2ha-1a-1,corresponding to 12.1±0.7 Mg C ha-1a-1.
Table 2 Stand structural traits of F.sylvatica trees in the six sampling areas
Fig.2 Monthly trend of net photosynthetic rates(NP)for F.sylvatica in the six samplings areas.Mean values±S.E.(n=45).The explanations of Areas 1–6 see Fig.1
Highest coefficient of determination was recorded between NPand gsfor areas at 1000 and 1500 m a.s.l.(R2=0.87 and 0.72,respectively).Moreover,the linear correlations between Taand PFD%showed that 41%of the variation in Tadepended on variation in PFD%,which in turn was influenced by variation in LAI(by 75%)(Table 3).PCA returned two axes of variations across the six sampling areas.PC1accounted for 58%of the total variance and it was positively related to LAI and Dt,and negatively related to DBH.PC2was negatively related to NPyearlyaccounting for 25%of the total variance.There was a significant positive relationship between PC1and CO2S(R2=0.65,p<0.05)(Fig.5).
Fig.3 Monthly trend of a leaf transpiration rates(E)and b stomatal conductance(gs)for F.sylvatica in the six considered areas.Mean values± S.E.(n=45).The explanations of Areas 1–6 see Fig.1
Carbon sequestration capacity depends on several factors including forest structure and species composition(Gratani and Varone 2006,2007;Gratani et al.2011,2013).LAI is an important variable for characterizing vegetation structure and function(Garrigues et al.2008),scaling up photosynthesis from leaves to the canopy(Leuning et al.1995;Baldocchi 1997;de Pury and Farquhar 1997;Ryu et al.2011).
Fig.4 Carbon dioxide sequestration capability[CO2S,Mg CO2-ha-1a-1]in the considered areas.Mean values±S.E.(n=10).The explanations of Areas 1–6 see Fig.1
Table 3 Regression analysis between net photosynthetic rates(NP)and stomatal conductance(gs),between leaf transpiration(E)and gs,between air temperature inside the area(Ta)and the photon flux density at soil level below the vegetation(PFD%)and between PFD%and leaf area index(LAI)
Fig.5 Linear relationship between the first principal component combining LAI,Dt and DBH(PC1)and the CO2sequestration capacity per hectare[CO2S,Mg CO2ha-1year-1]across the six considered areas.The equations as well as their R2are shown.The relationships were significant at p<0.05.The explanations of Areas 1–6 see Fig.1
On the whole,our results show that structural stand traits,particularly LAI and Dt are good predictors of CO2S,as attested by the significant relationship between the first component,explaining the largest proportion of the variance(i.e.PC1)and CO2S.The managed areas(i.e.areas 1 and 4)were distinguished from the others by 28%lower CO2S(mean value)due to 18%lower LAI and 58%lower Dt.Thinning is a common management strategy in European beech forests(Simon et al.2017)and has been reported to show positive effects on tree radial growth(Van der Maaten 2013).In fact,DBH in these areas were 9 and 60% higher in high-forest and coppice,respectively,compared to the un-managed counterpart.Thinning improves natural regeneration in beech forests(Fotelli et al.2004)through formation of gaps which determine light penetration at soil level(Gower and Norman 1991),consequently improving the abiotic variables(i.e.a greater light intensity,mineral content,water availability and temperature)for seedlings in beech forests(Fotelli et al.2004).Accordingly,the higher PFD%in areas 1 and 4 due to thinning allowed regeneration of F.sylvatica seedlings(Frattaroli,data not pubblished).With regard to the unmanaged areas,the results show a clear separation between the high-forests(areas 2 and 3)and coppices(areas 5 and 6)with the latter having 6%higher CO2S.This result is justified by 73%more trees with a 42%lower DBH and a 27%lower H than in the high-forests,according to the results of Scolastri et al.(2017)for F.sylvatica in Italy.Thus,the higher Dt and LAI determine the higher CO2S in coppices.
Mean yearly NPwas not a major predictor of CO2S.The sampled areas had similar trends in NP,with highest NPrecorded in May[11.5 ± 0.2 μmol(CO2)m-2s-1,mean value of the six areas]decreasing by 44%in July(mean value of the six areas)due to high air temperature and low water availability.This trend highlights the beech sensitivity to these environmental factors.In particular,area 3 and area 6(1500 m a.s.l.)had 8%higher NPdue to favorable air temperature and humidity(11%lower Taand 8%higher RH)than at areas 1,2,4 and 5(1000 m a.s.l.).In these areas rainfall shortage in summer months can be partially compensated by the higher elevation,which moderates air temperature and humidity so that evapotranspiration is lower and fog formation partially compensates for rainfall shortage(Thiébaut 1984;Gutiérrez 1988;Aranda et al.2000).This justified also the highest gsand E rates in these areas in July,as con firmed by regression analysis.
Our results highlight a mean CO2sequestration capacity of 44.3±2.6 Mg CO2ha-1a-1at the F.sylvatica stands,corresponding to 12.1±0.7 Mg C ha-1a-1.Extending the mean CO2S value over the entire area covered by beech forests in Abruzzo(122,402 ha;INFC 2005),we calculate a total CO2sequestration capacity of 5,422,408 Mg CO2year-1, which corresponds to 1,477,496 Mg C sequestered per year.
Therefore it must be taken into consideration that forests not only have economic value from the commercial production of timber,but are also of value to the public through their important environmental functions including their roles as carbon sinks,contributing to biodiversity,and protecting water resources(Pukkala 2002;Nakajima et al.2011).
In conclusion,we hypothesized less CO2would be sequestered in managed high-forests and coppices,and more CO2would be sequestered in coppices than in high forests.This was supported by our results.Thus,considering that natural regeneration can play a crucial role in forest restoration(Puettmann et al.2015;Muscolo et al.2017),selective thinning can be recommended to support traditional forest management practices such as civic use in some areas of Majella National Park.At the same time,the un-managed areas should be maintained in the reserve to give priority to the objective of conservation and naturalistic improvement of forest heritage(Feliziani 2006).In fact,the obtained information would be helpful to devise conservation strategies for diverse forest types(Jain and Ansari 2013).Moreover,our results can be used as a database for management projects at Majella National Park.This complies with the UNFCCC(1997)specification that the effects of forest management can be incorporated into the national green-house gas emissions budget as an additional human-induced activity(de Simon et al.2012).Moreover,these results can be incorporated in a geographic information system to monitor spatial variations in CO2concentrations over time.
AcknowledgementsThe authors thank the State Forestry Corps for valuable information and for the great support during the surveys.
Journal of Forestry Research2018年6期