Biogeosciences 14 1593ndash1602 2017wwwbiogeosciencesnet1415932017doi105194bg-14-1593-2017copy Author(s) 2017 CC Attribution 30 License

Estimating global nitrous oxide emissions by lichens and bryophyteswith a process-based productivity modelPhilipp Porada12 Ulrich Poumlschl3 Axel Kleidon4 Christian Beer12 and Bettina Weber3

1Department of Environmental Science and Analytical Chemistry (ACES) Stockholm University10691 Stockholm Sweden2Bolin Centre for Climate Research Stockholm University 10691 Stockholm Sweden3Max Planck Institute for Chemistry PO Box 3060 55020 Mainz Germany4Max Planck Institute for Biogeochemistry PO Box 10 01 64 07701 Jena Germany

Correspondence to Philipp Porada (philippporadaacessuse)

Received 3 October 2016 ndash Discussion started 14 October 2016Revised 30 January 2017 ndash Accepted 3 March 2017 ndash Published 28 March 2017

Abstract Nitrous oxide is a strong greenhouse gas and at-mospheric ozone-depleting agent which is largely emittedby soils Recently lichens and bryophytes have also beenshown to release significant amounts of nitrous oxide Thisfinding relies on ecosystem-scale estimates of net primaryproductivity of lichens and bryophytes which are convertedto nitrous oxide emissions by empirical relationships be-tween productivity and respiration as well as between res-piration and nitrous oxide release Here we obtain an alter-native estimate of nitrous oxide emissions which is basedon a global process-based non-vascular vegetation modelof lichens and bryophytes The model quantifies photosyn-thesis and respiration of lichens and bryophytes directly asa function of environmental conditions such as light andtemperature Nitrous oxide emissions are then derived fromsimulated respiration assuming a fixed relationship betweenthe two fluxes This approach yields a global estimate of027 (019ndash035) (Tg N2O) yearminus1 released by lichens andbryophytes This is lower than previous estimates but cor-responds to about 50 of the atmospheric deposition of ni-trous oxide into the oceans or 25 of the atmospheric depo-sition on land Uncertainty in our simulated estimate resultsfrom large variation in emission rates due to both physiolog-ical differences between species and spatial heterogeneity ofclimatic conditions To constrain our predictions combinedonline gas exchange measurements of respiration and nitrousoxide emissions may be helpful

1 Introduction

Lichens and bryophytes have increasingly been recognizedto play a relevant role in global biogeochemical cycles (El-bert et al 2012 Sancho et al 2016 Barger et al 2016)They are globally abundant growing on soils rocks and epi-phytically on trees At high latitudes they may form exten-sive covers on the forest floor and in wetlands mosses fre-quently represent the dominant vegetation type In drylandslichens and bryophytes form so-called biological soil cruststogether with photosynthesizing cyanobacteria algae fungiand bacteria These crusts cover vast areas in arid and semi-arid ecosystems

In a first approach based on empirical upscaling of fieldmeasurements according to ecosystem categories Elbertet al (2012) calculated that lichens and bryophytes to-gether with free-living cyanobacteria and algae fix around143 (Gt CO2) yearminus1 (39 Gt carbon) at the global scale Thiscorresponds to about 7 of the net primary productivity(NPP) by terrestrial vegetation In an alternative approach tothe empirical upscaling of observations Porada et al (2013)utilized a process-based non-vascular vegetation model forlichens and bryophytes called LiBry to calculate the NPP ofthese organism groups at the global scale obtaining similarresults

In addition to photosynthetic carbon uptake lichensand bryophytes are able to fix nitrogen through symbio-sis with cyanobacteria (DeLuca et al 2002 Barger et al2016) Together with free-living cyanobacteria their nitro-

Published by Copernicus Publications on behalf of the European Geosciences Union

1594 P Porada et al Nitrous oxide emissions by lichens and bryophytes

gen fixation was estimated to sum up to a global valueof sim 49 (Tg N) yearminus1 (Elbert et al 2012) which accountsfor nearly half of the biological nitrogen fixation on landThe LiBry model yielded a similar estimate of up to34 (Tg N) yearminus1 based on the nitrogen requirements oflichens and bryophytes determined by Porada et al (2014)Moreover it was found in the same study that the organismsmay contribute significantly to biotic enhancement of globalchemical weathering by release of weathering agents such asorganic acids Their potential for chemical weathering wasderived from their phosphorus demand assuming that theydissolve surface rocks to acquire phosphorus

Recently lichen- and bryophyte-related nitrogen fluxesother than fixation of nitrogen have been shown to be sig-nificant at the global scale Weber et al (2015) found thatbiological soil crusts which may contain large fractionsof lichens or bryophytes emit considerable quantities ofthe reactive trace gases NO and HONO accounting forsim 17 (Tg N) yearminus1 This corresponds to sim 20 of globalnitrogen oxide emissions from soils under natural vegetation(Ciais et al 2013)

Furthermore Lenhart et al (2015) showed that a largevariety of lichen and bryophyte species release nitrous ox-ide (N2O) They estimated that the organisms emit a to-tal value of 045 (032ndash059) (Tg N2O) yearminus1 at the globalscale which corresponds to 4ndash9 of natural terrestrial N2Oemissions (Zhuang et al 2012) Since N2O is an impor-tant greenhouse gas and also the main depleting substance ofstratospheric ozone which is still emitted today quantifyingall contributing sources is of high importance (Butterbach-Bahl et al 2013 Ravishankara et al 2009 Gaumlrdenaumls et al2011 Ciais et al 2013)

Absolute values of N2O release estimated by Lenhart et al(2015) were highest for lichens and bryophytes living onthe ground in the boreal zone and for epiphytic lichens andbryophytes in the humid tropics The relative contributionsof lichens and bryophytes to total ecosystem N2O emissionshowever were highest in desert and tundra biomes due tothe low emissions by other vegetation and the soil there Thehigh relevance of lichens and bryophytes for N2O emissionsin drylands and at high latitudes is in accordance with theirstrong impacts on other components of the nitrogen cyclein these regions Bryophytes for instance have been sug-gested to be the main source of nitrogen input into borealforests through fixation from the atmosphere by cyanobacte-rial partners (DeLuca et al 2002) Also in drylands lichensand bryophytes are crucial for input of nitrogen into theecosystem (Barger et al 2016) and they may even be essen-tial providers of nitrogen for vascular plants (Stewart 1967Hawkes 2003)

The estimate by Lenhart et al (2015) is derived from mea-suring emissions of N2O by the organisms in the laboratoryunder a range of environmental conditions All lichen andbryophyte species analysed by Lenhart et al (2015) showedrelease of N2O Lichens and bryophytes were shown to uti-

lize 15N labelled NOminus3 but not NH+4 indicating that N2Ois likely formed during denitrification The exact process ofN2O formation however remains largely unknown One op-tion is that the organisms themselves release N2O during themetabolization of nitrate in a similar way to that suggestedby Smart and Bloom (2001) for vascular plants Another op-tion is that bacteria growing on lichen and moss cushionsare responsible for the emissions of N2O This second op-tion is supported by a recently published study wherein sev-eral strains of the bacterial genus Burkholderia which wereshown to emit N2O were isolated from the boreal peat mossSphagnum fuscum (Nie et al 2015) While Lenhart et al(2015) describe that the substrate which the organisms grewon was thoroughly removed further cleaning steps to re-move potential bacterial colonies have not been conducted

Another finding by Lenhart et al (2015) is that N2O emis-sions are related to respiration by a relatively constant factorBy applying this factor and furthermore assuming a fixedratio between respiration and NPP the authors utilized theglobal NPP data of Elbert et al (2012) to obtain globally re-solved N2O emissions by lichens and bryophytes The relia-bility of global estimates derived from upscaling of small-scale measurements depends on the variation of the mea-sured fluxes The field measurements of NPP which wereextrapolated to the spatial scale of a biome by Elbert et al(2012) vary by around 2 orders of magnitude Measurementsof N2O emissions by lichens and bryophytes too show con-siderable variation Regarding biological soil crusts severalstudies analysed denitrification rates to be negligible (John-son et al 2007 Strauss et al 2012) and N2O productionwas calculated to constitute only 3ndash4 of the N fixationrate (Barger et al 2013) Other studies however describedhigh denitrification rates that either increased (Brankatschket al 2013) or decreased with advancing crust development(Abed et al 2013) One possibility to increase the reliabil-ity of large-scale estimates of N2O emissions by lichens andbryophytes is the application of alternative methodically dif-ferent approaches

For this reason we apply here the process-based non-vascular vegetation model LiBry (Porada et al 2013) toassess the contribution of these organisms to the globalN2O budget LiBry simulates photosynthesis respiration andgrowth of lichens and bryophytes as a function of environ-mental conditions To distinguish global patterns of produc-tivity on the ground and in the canopy the model repre-sents these locations and their differing environmental con-ditions separately We calculate respiration by lichens andbryophytes directly as a function of environmental condi-tions and we derive N2O emissions based on the simulatedrespiration By doing this we obtain physiologically drivenand spatially resolved data on the N2O emissions by lichensand bryophytes at the global scale Since we estimate res-piration with LiBry we do not need to make assumptionsregarding the ratio of NPP to respiration contrary to Lenhartet al (2015) Furthermore we quantify different sources of

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1595

variation in N2O emissions and determine their relative im-portance

2 Methods

The non-vascular vegetation model LiBry estimates globalpatterns of photosynthesis respiration and net primary pro-ductivity of lichens and bryophytes (Porada et al 2013) Themodel calculates these physiological processes as a functionof climate and additional environmental conditions whichare provided in the form of time series of global griddedmaps Photosynthesis in LiBry is determined by ambient lev-els of light CO2 and temperature according to the Farquharapproach (Farquhar and von Caemmerer 1982) Respirationis simulated as a function of temperature via a Q10 rela-tionship Both processes also depend on the water status ofthe simulated lichens and bryophytes which includes limita-tion of CO2 diffusion at high water content NPP is derivedfrom the difference between gross photosynthesis and respi-ration A unique feature of LiBry is that functional diversityof lichens and bryophytes is represented by a large numberof artificial species instead of being aggregated into one or afew average functional types The advantage of this approachis that adaptation of the organisms to differing environmentalconditions is simulated in a more realistic way Physiologi-cal processes such as photosynthesis and respiration are cal-culated separately for each artificial species LiBry has beensuccessfully applied to estimate global NPP by lichens andbryophytes (Porada et al 2013) and other impacts of theseorganisms on global biogeochemical cycles (Porada et al2014 2016b)

The model version presented here contains several exten-sions compared to the original version first an NPP-basedweighting scheme was introduced which assigns relativeabundances to all artificial species that survive in a grid cellof the model in the steady state (Porada et al 2016b) Thisallows an average grid cell value of NPP based on the relativeabundances of the simulated species in that cell to be derivedIn the original version grid cell NPP could only be predictedin the form of a range of values due to unknown abundancesof the species The average grid cell NPP is close to the upperend of the range of productivity values since the most pro-ductive simulated species are assumed to be the most abun-dant ones Secondly a dynamic disturbance scheme was im-plemented which replaces the equilibrium computation ofsurface coverage by a monthly update of coverage (Poradaet al 2016a) This makes the new model applicable to tran-sient scenarios of climatic and environmental change whilethe original model required the assumption of a steady stateto compute coverage

For this study we run LiBry with an initial value of 3000artificial species in each grid cell for a period of 600 years toreach steady state with climatic fields and other forcing datafrom Porada et al (2013) Our global estimates are based on

average values over the last 50 years of the simulation Weevaluate the new version of LiBry in the same way as theoriginal one (Porada et al 2013) by comparing simulatedNPP to field measurements on a biome basis

LiBry does not include an explicit representation ofprocesses that directly result in emission of nitrous ox-ide However it has been determined experimentally byLenhart et al (2015) that N2O emissions by lichens andbryophytes are related to their respiration by a conversionfactor of 16 ng N2O (mg CO2)minus1 The conversion factor hasa 90 confidence interval of 11 to 21 ng N2O (mg CO2)minus1Since LiBry explicitly calculates respiration by lichens andbryophytes we derive N2O emissions from simulated respi-ration using the conversion factor of Lenhart et al (2015)

The study by Lenhart et al (2015) uses NPP of lichensand bryophytes together with free-living cyanobacteria andalgae to estimate N2O emissions since global upscaled dataon respiration of these organisms are not available from El-bert et al (2012) Thereby Lenhart et al (2015) assume afixed ratio of respiration to NPP To determine this ratio theyevaluate literature data obtaining a rate of respiration rela-tive to net photosynthesis of sim 49 (Lenhart et al 2015Table S6) Since measurements have been made in the sun-light but respiration continues in the dark respiration is mul-tiplied by a factor of 2 assuming a 12 h day This leads to anestimated ratio of respiration to NPP which is roughly 1 1To evaluate LiBry further we compute the ratio of respirationto NPP in LiBry to assess whether the model is in agreementwith these observations

In the study of Lenhart et al (2015) the substrate of thesamples was removed to avoid biases resulting from N2Orelease by microbes in the substrate Foliose and fruticoselichens as well as mosses were collected which grew on soilrocks and epiphytically on trees and the authors found novariation in N2O emissions depending on the underlying sub-strate Endolithic and crustose lichens were not included inthat study as for these growth forms the dry weight which isneeded for calculations could not be determined in a reliablemanner

Variation in field measurements of N2O emissions may re-sult not only from physiological differences between speciesbut also from variation in climatic conditions which canbe significant at the small scale To upscale emissions frompoint measurements to the large scale it is important to quan-tify the relative contributions of these different sources ofvariation When for instance the variation between speciesregarding their N2O emissions is small it suffices to samplea low number of species to obtain an average emission for acertain climatic condition LiBry can provide an indication ofthe relative importance of these sources of variation since themodel not only represents climate variability but also simu-lates diverse physiological strategies Each grid cell of themodel contains a range of surviving species at the end ofthe simulation and that consequently shows a range of N2Oemissions LiBry does not simulate spatial variation in cli-

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1596 P Porada et al Nitrous oxide emissions by lichens and bryophytes

matic conditions within a grid cell However by comparingaverage emission rates of grid cells from different climates itis possible to assess the relative importance of climatic con-ditions for variation in N2O emissions We select five modelgrid cells from different ecosystem classes to analyse the rel-ative importance of differences between species and climaticheterogeneity on variation in N2O emissions It should bepointed out that LiBry does not compute directly N2O emis-sions by lichens and bryophytes but it derives them fromsimulated respiration through an empirical linear relation-ship Hence differences in the sensitivities of respiration andN2O emissions to climatic conditions may lead to uncertain-ties in our predicted effects of climate on N2O emissions

3 Results

The global distribution of net primary productivity simulatedby the updated version of LiBry is shown in Fig 1 Produc-tivity by lichens and bryophytes is highest in forested re-gions and lowest in deserts and agricultural regions Hencethe spatial pattern is mainly controlled by water availabil-ity except for cropland In LiBry it is assumed that lichensand bryophytes only grow on the area fraction of a grid cellwhich is not occupied by crops Therefore on a grid cell ba-sis regions with a high fractional cover of cropland show lowproductivity by lichens and bryophytes in spite of favourableclimatic conditions The high productivity in the humid trop-ics mainly results from epiphytic lichens and bryophytes inthe canopy while in the boreal zone the larger fraction ofproductivity stems from the ground

As a result of the dynamic surface coverage the spatialpatterns of NPP differ slightly between the new and the orig-inal version of LiBry but the large-scale gradients remain thesame Comparing the global pattern of lichen and bryophyteNPP simulated by the new version of LiBry to an empiri-cal estimate by Elbert et al (2012) shows good agreementsimilar to the original version Furthermore the total globalNPP predicted by the new LiBry differs from the original es-timate due to the updated calculation of coverage The maindifference is found for the tropical forest canopy where sim-ulated NPP increases significantly The total global NPP of43 (Gt C) yearminus1 estimated by the new LiBry compares wellto the value of 39 (Gt C) yearminus1 calculated by Elbert et al(2012)

Comparison of simulated NPP to field measurements ona biome basis suggests that LiBry predicts realistic valuesof NPP for a range of ecosystems (Fig 2) In particularsimulated NPP in the tropical and the boreal forest matcheswell with observations while the original version of LiBryseemed to underestimate NPP in these biomes In the biomesdesert and to a lesser extent tundra LiBry seems to overes-timate productivity which may have also been the case withthe original version A potential explanation for this is thatproductivity in dry and cold areas may be limited not only by

NPP

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

a)(

10 100 1000

[g C mminus2 yrminus1]

NPP in canopy

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

b)(

10 100 1000

[g C mminus2 yrminus1]

NPP on ground

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

c)(

10 100 1000

[g C mminus2 yrminus1]

Figure 1 Global patterns of NPP Lichen and bryophyte NPP esti-mated by LiBry for (a) all locations of growth (b) the canopy and(c) the ground The estimates are in grams of carbon per square me-tre and they are average values over the last 50 years of a 600-yearsimulation with 3000 initial species Grey colour denotes regionswhere no simulated species is able to survive such as ice shieldsand the driest regions of deserts

climatic factors but also by nutrient availability (Porada et al2016b) Since photosynthesis and growth are only controlledby climatic factors in LiBry the effect of spatial variation innutrient availability on productivity cannot yet be simulatedIt should be pointed out however that except for the borealbiome the number of field measurements is quite low andconsequently the observation-based characteristic values foreach biome are subject to considerable uncertainty

Figure 3 shows simulated global patterns of nitrous ox-ide emissions by lichens and bryophytes Nitrous oxideemission is highest in the humid tropics and subtropicswith values up to 10 (mg N2O) mminus2 yearminus1 (Fig 3a) Asecond region of high emissions is the boreal zone withvalues up to 8 (mg N2O) mminus2 yearminus1 Dry regions showlowest values of nitrous oxide emissions in general lessthan 1 (mg N2O) mminus2 yearminus1 Considering only lichens andbryophytes which grow as epiphytes in the canopy (Fig 3b)

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1597

001

01

1

10

100

1000

Tundra Boreal f orestf loor

Desert Tropical f orestcanopy

Net carb

on u

pta

ke [(g

C)

mndash 2

yrndash 1

]

16

69

9

2

Figure 2 Comparison of LiBry estimates to field measurementsNPP estimated by LiBry compared to field measurements from fourbiomes defined after Olson et al (2001) The blue dots show theaverage simulated NPP for each biome and the blue vertical barsshow the range of NPP values between the different grid cells ina biome The magenta diamonds correspond to the median of NPPvalues measured in the field on the small scale the magenta verticalbars denote the range of the field measurements Left of the ma-genta diamonds the number of field measurements is shown that isconsidered for the respective biome Details can be found in Poradaet al (2013)

emissions in the humid tropics are around 3 times higher thanin the boreal and temperate zones Lichens and bryophytes onthe ground show highest values of nitrous oxide emissions inthe boreal zone with values around 3 (mg N2O) mminus2 yearminus1

(Fig 3c) Regarding the ground tropical and subtropical re-gions only partly show N2O emissions comparable to thoseof the boreal zone The reason for this is low simulatedproductivity and coverage of lichens and bryophytes on theground in tropical and subtropical climates which also leadsto low respiration on a grid cell level and hence to low N2Oemissions

Figure 4 shows the simulated global spatial distribution ofthe ratio of respiration to NPP The assumption of a globallyconstant ratio of respiration to NPP is used by Lenhart et al(2015) to derive ecosystem-scale N2O emissions by lichensand bryophytes from their NPP Alternatively this ratio canbe derived from the independent LiBry estimates of NPP andrespiration The simulated ratio shows a latitudinal patternwith increasing values towards the tropics (Fig 4a) This re-sults from the influence of surface temperature on respira-tion in combination with high nighttime temperatures in thehumid tropics which cause high respiration rates during thenight Note that high respiration relative to NPP of tropicallichens and bryophytes does not necessarily mean high res-piration at the grid cell level since net productivity and cov-erage may be low Respiration by lichens and bryophytes inthe canopy shows a slightly weaker latitudinal gradient thanon the ground which can be explained by efficient evapo-rative cooling in the canopy (Fig 4b) In contrast lichens

N2O release

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

(a)

00 15 30 45 60 75 90 105 120 135 150

[mg mminus2 yrminus1]

N2O release in canopy

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

(b)

00 10 20 30 40 50 60 70 80 90 100

[mg mminus2 yrminus1]

N2O release on ground

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

(c)

00 05 10 15 20 25 30 35 40 45 50

[mg mminus2 yrminus1]

Figure 3 Global patterns of N2O release Nitrous oxide emissionsby lichens and bryophytes estimated by LiBry for (a) all locations ofgrowth (b) the canopy and (c) the ground Note the differing rangesof the colour bars Grey colour denotes regions where no simulatedspecies is able to survive such as ice shields and the driest regionsof deserts

and bryophytes on the ground usually grow within the sur-face boundary layer which reduces cooling by turbulent heattransfer leading to a strong influence of incoming radiationon surface temperature Since radiation input increases to-ward the Equator the ratio of respiration to NPP on theground in the tropics is markedly higher than at high lati-tudes (Fig 4c) The ratio of respiration to NPP varies fromless than 1 to around 2 while most values are around 1 Thismeans that gross primary productivity (GPP) is partitionedroughly equally into NPP and respiration which agrees wellwith the observational data from Lenhart et al (2015)

An overview of global total values of N2O emissions res-piration NPP and the ratio of respiration to NPP estimatedby LiBry is shown in Table 1 Table 2 shows N2O emis-sions by lichens and bryophytes for individual grid cells fromfive different ecosystem classes (see also Table 1) Variationin emissions between species within a grid cell is large itcan exceed 3 orders of magnitude The variation due to cli-

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1598 P Porada et al Nitrous oxide emissions by lichens and bryophytes

Table 1 Annual global total values of N2O emissions NPP respiration and the ratio of respiration to NPP estimated by LiBry and separatedinto lichens and bryophytes living in the canopy and on the ground The values in brackets in the first column show the uncertainty in N2Oemissions due to the conversion of released CO2 to N2O (90 confidence interval from Lenhart et al (2015)) Ecosystem classes shown arebased on the categories made by Olson et al (2001) which were aggregated by us in the same way as in Elbert et al (2012) ldquoGt Crdquo standsfor gigatons of carbon

N2O emissions NPP Respiration Respiration NPP(Tg N2O) yearminus1 (Gt C) yearminus1 (Gt C) yearminus1 [ ]

Canopy+ groundGlobal 027 (019ndash035) 43 45 110Tropical forest 011 (008ndash014) 15 18 133Extratropical forest 011 (008ndash014) 20 18 093Steppe amp savannah 003 (002ndash004) 04 04 121Desert 002 (001ndash003) 04 04 105Tundra 001 (0007ndash0013) 02 02 087

Canopy global 013 (009ndash017) 21 22 101Ground global 014 (010ndash018) 22 23 116

Table 2 Simulated nitrous oxide emissions by lichens and bryophytes in (mg N2O) mminus2 yearminus1 for individual grid cells of the LiBry modelThe values are averages over the last 50 years of a 600-year simulation with 3000 initial species Grid cells are selected from five differentecosystem classes In the two forest classes emissions are separated into canopy and ground In the other classes the model does notrepresent lichens and bryophytes in the canopy The range of N2O emissions based on all surviving artificial species in a grid cell is shownThe average value for all species in a grid cell is derived by an NPP-based weighting scheme (see Sect 2)

Ecosystem class Location Minimum Average Maximum

Tropical forest Central Amazon ground 031 059 088canopy 0081 33 82

Extratropical forest West Siberia ground 0023 17 48canopy 00040 21 62

Steppe amp savannah Central Sahel 00095 0088 032Desert Central Australia 0019 16 55Tundra North Alaska 0012 0095 017

matic conditions is smaller but it still amounts to almost 2orders of magnitude based on the grid cells with the high-est and lowest average emission rates Comparing Table 2 tothe global range of N2O emissions by lichens and bryophytes(Fig 3) shows that the five selected grid cells represent wellthe global variation in emissions due to climatic conditionsThus both functional diversity of the artificial species anddifferent climatic conditions are important for variation ofN2O emissions according to the LiBry simulation

4 Discussion

In this study we estimate nitrous oxide emissions by lichensand bryophytes with the global process-based non-vascularvegetation model LiBry Thereby we derive N2O emissionsfrom respiration fluxes which are together with photosyn-thesis and net primary productivity simulated by LiBry

We use an updated version of LiBry which contains signif-icant modifications with regard to the original version pub-lished in Porada et al (2013) Regarding NPP the new ver-

sion estimates 43 (Gt C) yearminus1 while the original versionof LiBry predicted a range of 034 to 33 (Gt C) yearminus1 Theincrease in predicted NPP is mainly attributed to a highersimulated productivity in the tropical forest canopy sincea new disturbance scheme allows for a higher surface cov-erage of lichens and bryophytes there An empirical globalestimate of NPP by lichens bryophytes free-living terres-trial cyanobacteria and algae (Elbert et al 2012) amountsto 39 (Gt C) yearminus1 Our new estimate is higher than thatby Elbert et al (2012) although LiBry does not considerfree-living cyanobacteria and algae This may be explainedby the small contribution of cyanobacteria and algae to theoverall global carbon uptake which can be compensatedfor by minor relative changes in productivity of lichensand bryophytes (Darrouzet-Nardi et al 2015 Sancho et al2016) It is not straightforward to determine which numberis closest to reality since both the process-based estimate byLiBry and the empirical one by Elbert et al (2012) are sub-ject to uncertainty In the study of Elbert et al (2012) forinstance it is assumed that productivity and active time areuniform within a biome Furthermore Elbert et al (2012) use

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1599

Ratio respiration to NPP

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

a)(

00 03 06 09 12 15 18 21 24

[ ]

Ratio respiration to NPP in canopy

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

b)(

00 03 06 09 12 15 18 21 24

[ ]

Ratio respiration to NPP on ground

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

c)(

00 03 06 09 12 15 18 21 24

[ ]

Figure 4 Global patterns of the ratio of respiration to NPP Ratioof respiration to NPP of lichens and bryophytes estimated by LiBryfor (a) all locations of growth (b) the canopy and (c) the ground

a globally uniform value of surface cover fraction to scaleup local field measurements of productivity to the globalscale However values of surface coverage by lichens andbryophytes compiled by Elbert et al (2012) vary greatly atthe small scale which makes upscaling to larger scales chal-lenging

While productivity estimated by LiBry is evaluated in thisstudy large-scale surface coverage of lichens and bryophytessimulated by LiBry has been evaluated for regions north of50 N in Porada et al (2016a) It was shown that LiBry pre-dicts realistic values of cover fraction Moreover values ofsurface cover predicted by LiBry for other regions of theworld (Porada et al 2016b) are in agreement with the es-timate of Elbert et al (2012) In spite of uncertainties regard-ing productivity and abundance of lichens and bryophytescomparing the empirical and process-based approaches givesconfidence in the order of magnitude of the LiBry simulationresults

As a 50-year steady-state average value we estimate to-tal N2O emissions by lichens and bryophytes of 027 (019ndash035) (Tg N2O) yearminus1 which is at the lower end of the rangeof 032 to 059 (Tg N2O) yearminus1 calculated by Lenhart et al(2015) The evaluation of LiBry regarding simulated NPPshows that our global patterns and total values of NPP arevery similar to the empirical estimate by Elbert et al (2012)Since Lenhart et al (2015) use this NPP estimate by Elbertet al (2012) to derive N2O emissions differences in NPPare most likely not the reason for our lower estimate of N2Oemissions compared to Lenhart et al (2015) Instead thismay be explained by differing methods to compute respira-tion while Lenhart et al (2015) assume a globally uniformratio of respiration to NPP of a value of 2 to estimate respira-tion LiBry simulates respiration independently as a species-specific function of temperature and water status This re-sults in a lower global average value of around 1 for the ra-tio of respiration to NPP predicted by LiBry Our estimatedratio of respiration to NPP agrees well with laboratory mea-surements but it is in general difficult to compare a globalecosystem-scale value to small-scale and short-term obser-vations

Our simulated global pattern of N2O emissions is slightlydifferent than that shown in Lenhart et al (2015) who esti-mate highest values in the boreal zone and only intermedi-ate values in the humid tropics This can be explained bytheir assumed constant ratio of respiration to NPP whichmakes their global pattern of N2O emissions identical to thatof NPP which is shown in Elbert et al (2012) In LiBryhowever the simulated ratio of respiration to NPP increasestowards higher surface temperatures in the tropics (Fig 4)Furthermore the ratio shows large spatial variation Evaluat-ing this simulated pattern is difficult since estimates whichare extrapolated to the large scale such as the NPP estimateby Elbert et al (2012) are not available for respiration bylichens and bryophytes However observed ratios of respira-tion to NPP of lichens and bryophytes vary considerably atthe species level as shown by for example Lenhart et al(2015) Using a constant ratio of respiration to NPP maytherefore introduce a bias in the estimated spatial distribu-tion of N2O emissions

Small-scale measurements of N2O emissions by lichensand bryophytes may show considerable variation Thesources of this variation may be physiological differencesbetween species variation of associated microbial communi-ties andor heterogeneity in climatic conditions We examinethe relative importance for respiration of differences betweenspecies compared to climatic differences with LiBry sincethe model simulates various physiological strategies and rep-resents variation in climatic conditions at the global scaleThereby we assume that the relationship between respira-tion and N2O emissions is relatively insensitive to climaticconditions and physiological differences between species assuggested by the experiments by Lenhart et al (2015) Ta-ble 2 shows that both differences between artificial species

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1600 P Porada et al Nitrous oxide emissions by lichens and bryophytes

and different climatic conditions are important for variationof N2O emissions Upscaling of N2O emission rates mea-sured in the field may therefore be subject to considerableuncertainty Modelling approaches in this direction shouldprobably account for both interspecific variation in processesassociated with N2O release by lichens and bryophytes aswell as variation in climatic conditions

Although our approach considers the most importantsources of variation in N2O emissions by lichens andbryophytes it is associated with uncertainties that should bediscussed further These uncertainties mainly result from ourmethod to estimate respiration and from assumptions con-cerning the empirical relationship between respiration andN2O emissions

Respiration and the ratio of respiration to NPP simulatedby LiBry are difficult to validate since the number of labo-ratory or field studies which measure not only NPP but alsoGPP and respiration is not very high Moreover long-termmeasurements of respiration would be required to determinethe ratio of respiration to NPP Otherwise assumptions aboutthe contribution of respiration in the dark to total respirationare necessary

To obtain N2O emissions from respiration our resultsrely on the laboratory incubation measurements and the cal-culated ratio of N2O emissions to respiration presented inLenhart et al (2015) Furthermore our approach considerseffects of variation in climatic conditions on N2O emissionsby lichens and bryophytes Hence it is necessary to dis-cuss the sensitivity of the relationship between respirationand N2O emissions to a range of climatic conditions Asshown in Lenhart et al (2015 Fig 3) the relationship be-tween respiration and N2O emissions seems to be insensi-tive to temperature changes for the tested species Likewisevariations in water content have no clear effect on the rela-tionship between N2O release and respiration (Lenhart et al2015 Fig S3) Although the sensitivities of N2O release totemperature and water content are similar to those of respira-tion across species the relationship between N2O release andrespiration shows interspecific variation However in spite ofa large number of around 40 sampled species the relation-ship shows a relatively narrow 90 confidence interval of113 to 207 ng N2O (mg CO2)minus1 (Lenhart et al 2015) Thissuggests that the mechanism of N2O release by lichens andbryophytes is similar between different species

To analyse the relation between the production of N2Oand respiratory CO2 in greater detail measurements of bothfluxes by means of online gas exchange measurements wouldbe needed which then could be linked to the observed wa-ter status of the organisms Since LiBry explicitly representsthe dynamic water saturation of lichens and bryophytes thiswould allow a more process-based prediction of the durationand magnitude of N2O emissions In this way the uncertaintyassociated with our approach would be reduced facilitatingan improved estimate of global N2O emissions by lichensand bryophytes

In order to assess model-based estimates of N2O emis-sions by lichens and bryophytes a relatively large number offield measurements are necessary Currently most N2O mea-surements independent of the substrate or organisms mea-sured generally suffer from major uncertainties additionallyto variation from functional diversity and differing climaticconditions first the majority of these studies have been con-ducted using the acetylene inhibition technique The idea ofthis method is to inhibit the last denitrification step so thatthe measured N2O amounts should reveal the sum of N2Oand N2 release during denitrification under natural condi-tions It has however been shown quite a while ago thatthis method leads to an underestimation of denitrificationunder oxic conditions (Bollmann and Conrad 1997) Sec-ondly the most widely used measuring technique has beenthe closed-chamber method which is inexpensive and easyto use This however has major shortcomings as environ-mental conditions are hard to control and only limited surfaceareas can be measured (Butterbach-Bahl et al 2013 Groff-man 2012) Furthermore the limited temporal resolution ofchamber measurements may affect estimated N2O emissions(Barton et al 2015) Thirdly depending on the environmen-tal conditions under which the experiment is performed ndashparticularly water temperature and nutrient conditions ndash theobtained N2O emission rates could differ widely Thus it isindispensable to report and consider the exact environmen-tal conditions under which the measurements were made andto restrict natural emission data to those assessed under typi-cally occurring natural conditions

Respiration by lichens and bryophytes is not the only pro-cess which can be used to estimate their N2O emissionsBarger et al (2013) report a relationship between nitrogenfixation and N2O release in biological soil crusts which in-clude lichens and bryophytes as well as soil bacteria andalgae For this approach however reliable nitrogen fixationdata are sparse It is also possible to estimate the demand fornitrogen by lichens and bryophytes with LiBry with an un-certainty range of around 1 order of magnitude (Porada et al2014) However it is not straightforward to derive realizednitrogen uptake or nitrogen fixation from this since LiBrydoes not yet include processes related to nitrogen uptake ormetabolization of nitrogen species Therefore for this studywe chose the relation between respiration and N2O release toquantify N2O emissions by lichens and bryophytes

Our simulated global N2O emissions by lichens andbryophytes of 027 (019ndash035) (Tg N2O) yearminus1 amount toaround 3 of global N2O emissions from natural sources onland (Ciais et al 2013) This value may sound low at firstglance but it equals about 50 of the atmospheric deposi-tion of N2O into the oceans or 25 of the deposition on land(Ciais et al 2013) Considering that N2O has a strong nega-tive effect on stratospheric ozone and a significant warmingpotential as a greenhouse gas even relatively small emissionsshould not be neglected in global budgets

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1601

The study by Zhuang et al (2012) estimates global pat-terns of N2O emission from soils and finds that the hu-mid tropics contribute most to global N2O emission due tohigh temperature and precipitation Our simulated pattern ofglobal N2O emissions by lichens and bryophytes also showsa hotspot in the humid tropics but the relative contributionof the boreal zone to the global flux seems to be higher thanin Zhuang et al (2012) This probably results from the highsimulated NPP in the boreal zone particularly on the groundwhich compensates for the lower respiration and thereforeN2O emission per productivity due to low temperatures Rel-ative contributions of lichens and bryophytes to N2O emis-sions are highest for ecosystems in desert regions and athigh latitudes which agrees with the results by Lenhart et al(2015)

5 Conclusions

We estimate large-scale spatial patterns and global values ofN2O emissions by lichens and bryophytes from a process-based model of their productivity and respiration Our resultssuggest a significant contribution of lichens and bryophytesto global N2O emissions albeit at the lower end of the rangeof a previous empirical estimate Since both approachesuse respiration to derive N2O emissions our lower estimatelikely results from a different method to predict respirationcompared to the empirical approach Hence while estimatesof productivity are relatively well constrained evaluatingmodels with regard to estimated respiration may improvepredictions of N2O emissions by lichens and bryophytesOne important finding derived from our simulation is that theratio of respiration to NPP by lichens and bryophytes showsspatial variation and a latitudinal gradient at the global scaleThis means that productivity and N2O emissions by the or-ganisms are not necessarily correlated and that tropical re-gions may show higher emissions than polar regions giventhe same NPP Furthermore we show that both physiologicalvariation among species and variation in climatic conditionsare relevant for variation in respiration and consequentlyN2O emissions Ecosystem-scale estimates of N2O emis-sions by lichens and bryophytes should therefore include suf-ficient ranges of species and climatic conditions to avoid bi-ased results Our results build on the empirical finding thatN2O emissions by lichens and bryophytes are linearly re-lated to their respiration This relationship is relatively in-sensitive to climatic conditions and shows no large varia-tion between species However the relationship is based onclosed-chamber measurements Therefore it would be use-ful to perform online gas exchange measurements of N2Oemissions and respiration to test the effect of climatic con-ditions on the relationship between N2O release and respi-ration Furthermore using alternative approaches to estimateN2O emissions by lichens and bryophytes may be helpful toconstrain our approach

Code and data availability The non-vascular vegetation model Li-Bry used here is combined with an interface for parallel computingwhich was developed at the Max Planck Institute for Biogeochem-istry Jena Germany LiBry without the interface is freely availableas long as the names of the copyright holders and a disclaimer aredistributed along with the code in source or binary form The codeis available from the corresponding author upon requestModel output data which are presented as maps in this study areavailable as netCDF files from the authors on request

Author contributions Philipp Porada Bettina Weber andAxel Kleidon designed the model simulations and Philipp Poradacarried them out Philipp Porada prepared the manuscript withcontributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This work has been supported by the PAGE21project grant agreement number 282700 funded by the EC SeventhFramework Programme theme FP7-ENV-2011 and the CARBOP-ERM project grant agreement number 03G0836B funded by theBMBF (German Ministry for Science and Education) The authorsthank the Max Planck Society (Nobel Laureate Fellowship forBettina Weber) for financial support The Max Planck Institute forBiogeochemistry provided computational resources

Edited by V BrovkinReviewed by two anonymous referees

References

Abed R Lam P de Beer D and Stief P High rates of den-itrification and nitrous oxide emission in arid biological soilcrusts from the Sultanate of Oman ISME J 7 1862ndash1875doi101038ismej201355 2013

Barger N Castle S and Dean G Denitrification from nitrogen-fixing biologically crusted soils in a cool desert environ-ment southeast Utah USA Ecological Processes 2 1ndash9doi1011862192-1709-2-16 2013

Barger N Weber B Garcia-Pichel F Zaady E and Belnap JPatterns and Controls on Nitrogen Cycling of Biological SoilCrusts in Biological Soil Crusts An Organizing Principle inDrylands edited by Weber B Buumldel B and Belnap J 257ndash285 Springer International Publishing Cham doi101007978-3-319-30214-0_14 2016

Barton L Wolf B Rowlings D Scheer C Kiese R Grace PStefanova K and Butterbach-Bahl K Sampling frequency af-fects estimates of annual nitrous oxide fluxes Scientific Reports5 15912 doi101038srep15912 2015

Bollmann A and Conrad R Acetylene blockage technique leadsto underestimation of denitrification rates in oxic soils due toscavenging of intermediate nitric oxide Soil Biol Biochem 291067ndash1077 doi101016S0038-0717(97)00007-2 1997

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1594 P Porada et al Nitrous oxide emissions by lichens and bryophytes

gen fixation was estimated to sum up to a global valueof sim 49 (Tg N) yearminus1 (Elbert et al 2012) which accountsfor nearly half of the biological nitrogen fixation on landThe LiBry model yielded a similar estimate of up to34 (Tg N) yearminus1 based on the nitrogen requirements oflichens and bryophytes determined by Porada et al (2014)Moreover it was found in the same study that the organismsmay contribute significantly to biotic enhancement of globalchemical weathering by release of weathering agents such asorganic acids Their potential for chemical weathering wasderived from their phosphorus demand assuming that theydissolve surface rocks to acquire phosphorus

Recently lichen- and bryophyte-related nitrogen fluxesother than fixation of nitrogen have been shown to be sig-nificant at the global scale Weber et al (2015) found thatbiological soil crusts which may contain large fractionsof lichens or bryophytes emit considerable quantities ofthe reactive trace gases NO and HONO accounting forsim 17 (Tg N) yearminus1 This corresponds to sim 20 of globalnitrogen oxide emissions from soils under natural vegetation(Ciais et al 2013)

Furthermore Lenhart et al (2015) showed that a largevariety of lichen and bryophyte species release nitrous ox-ide (N2O) They estimated that the organisms emit a to-tal value of 045 (032ndash059) (Tg N2O) yearminus1 at the globalscale which corresponds to 4ndash9 of natural terrestrial N2Oemissions (Zhuang et al 2012) Since N2O is an impor-tant greenhouse gas and also the main depleting substance ofstratospheric ozone which is still emitted today quantifyingall contributing sources is of high importance (Butterbach-Bahl et al 2013 Ravishankara et al 2009 Gaumlrdenaumls et al2011 Ciais et al 2013)

Absolute values of N2O release estimated by Lenhart et al(2015) were highest for lichens and bryophytes living onthe ground in the boreal zone and for epiphytic lichens andbryophytes in the humid tropics The relative contributionsof lichens and bryophytes to total ecosystem N2O emissionshowever were highest in desert and tundra biomes due tothe low emissions by other vegetation and the soil there Thehigh relevance of lichens and bryophytes for N2O emissionsin drylands and at high latitudes is in accordance with theirstrong impacts on other components of the nitrogen cyclein these regions Bryophytes for instance have been sug-gested to be the main source of nitrogen input into borealforests through fixation from the atmosphere by cyanobacte-rial partners (DeLuca et al 2002) Also in drylands lichensand bryophytes are crucial for input of nitrogen into theecosystem (Barger et al 2016) and they may even be essen-tial providers of nitrogen for vascular plants (Stewart 1967Hawkes 2003)

The estimate by Lenhart et al (2015) is derived from mea-suring emissions of N2O by the organisms in the laboratoryunder a range of environmental conditions All lichen andbryophyte species analysed by Lenhart et al (2015) showedrelease of N2O Lichens and bryophytes were shown to uti-

lize 15N labelled NOminus3 but not NH+4 indicating that N2Ois likely formed during denitrification The exact process ofN2O formation however remains largely unknown One op-tion is that the organisms themselves release N2O during themetabolization of nitrate in a similar way to that suggestedby Smart and Bloom (2001) for vascular plants Another op-tion is that bacteria growing on lichen and moss cushionsare responsible for the emissions of N2O This second op-tion is supported by a recently published study wherein sev-eral strains of the bacterial genus Burkholderia which wereshown to emit N2O were isolated from the boreal peat mossSphagnum fuscum (Nie et al 2015) While Lenhart et al(2015) describe that the substrate which the organisms grewon was thoroughly removed further cleaning steps to re-move potential bacterial colonies have not been conducted

Another finding by Lenhart et al (2015) is that N2O emis-sions are related to respiration by a relatively constant factorBy applying this factor and furthermore assuming a fixedratio between respiration and NPP the authors utilized theglobal NPP data of Elbert et al (2012) to obtain globally re-solved N2O emissions by lichens and bryophytes The relia-bility of global estimates derived from upscaling of small-scale measurements depends on the variation of the mea-sured fluxes The field measurements of NPP which wereextrapolated to the spatial scale of a biome by Elbert et al(2012) vary by around 2 orders of magnitude Measurementsof N2O emissions by lichens and bryophytes too show con-siderable variation Regarding biological soil crusts severalstudies analysed denitrification rates to be negligible (John-son et al 2007 Strauss et al 2012) and N2O productionwas calculated to constitute only 3ndash4 of the N fixationrate (Barger et al 2013) Other studies however describedhigh denitrification rates that either increased (Brankatschket al 2013) or decreased with advancing crust development(Abed et al 2013) One possibility to increase the reliabil-ity of large-scale estimates of N2O emissions by lichens andbryophytes is the application of alternative methodically dif-ferent approaches

For this reason we apply here the process-based non-vascular vegetation model LiBry (Porada et al 2013) toassess the contribution of these organisms to the globalN2O budget LiBry simulates photosynthesis respiration andgrowth of lichens and bryophytes as a function of environ-mental conditions To distinguish global patterns of produc-tivity on the ground and in the canopy the model repre-sents these locations and their differing environmental con-ditions separately We calculate respiration by lichens andbryophytes directly as a function of environmental condi-tions and we derive N2O emissions based on the simulatedrespiration By doing this we obtain physiologically drivenand spatially resolved data on the N2O emissions by lichensand bryophytes at the global scale Since we estimate res-piration with LiBry we do not need to make assumptionsregarding the ratio of NPP to respiration contrary to Lenhartet al (2015) Furthermore we quantify different sources of

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1595

variation in N2O emissions and determine their relative im-portance

2 Methods

The non-vascular vegetation model LiBry estimates globalpatterns of photosynthesis respiration and net primary pro-ductivity of lichens and bryophytes (Porada et al 2013) Themodel calculates these physiological processes as a functionof climate and additional environmental conditions whichare provided in the form of time series of global griddedmaps Photosynthesis in LiBry is determined by ambient lev-els of light CO2 and temperature according to the Farquharapproach (Farquhar and von Caemmerer 1982) Respirationis simulated as a function of temperature via a Q10 rela-tionship Both processes also depend on the water status ofthe simulated lichens and bryophytes which includes limita-tion of CO2 diffusion at high water content NPP is derivedfrom the difference between gross photosynthesis and respi-ration A unique feature of LiBry is that functional diversityof lichens and bryophytes is represented by a large numberof artificial species instead of being aggregated into one or afew average functional types The advantage of this approachis that adaptation of the organisms to differing environmentalconditions is simulated in a more realistic way Physiologi-cal processes such as photosynthesis and respiration are cal-culated separately for each artificial species LiBry has beensuccessfully applied to estimate global NPP by lichens andbryophytes (Porada et al 2013) and other impacts of theseorganisms on global biogeochemical cycles (Porada et al2014 2016b)

The model version presented here contains several exten-sions compared to the original version first an NPP-basedweighting scheme was introduced which assigns relativeabundances to all artificial species that survive in a grid cellof the model in the steady state (Porada et al 2016b) Thisallows an average grid cell value of NPP based on the relativeabundances of the simulated species in that cell to be derivedIn the original version grid cell NPP could only be predictedin the form of a range of values due to unknown abundancesof the species The average grid cell NPP is close to the upperend of the range of productivity values since the most pro-ductive simulated species are assumed to be the most abun-dant ones Secondly a dynamic disturbance scheme was im-plemented which replaces the equilibrium computation ofsurface coverage by a monthly update of coverage (Poradaet al 2016a) This makes the new model applicable to tran-sient scenarios of climatic and environmental change whilethe original model required the assumption of a steady stateto compute coverage

For this study we run LiBry with an initial value of 3000artificial species in each grid cell for a period of 600 years toreach steady state with climatic fields and other forcing datafrom Porada et al (2013) Our global estimates are based on

average values over the last 50 years of the simulation Weevaluate the new version of LiBry in the same way as theoriginal one (Porada et al 2013) by comparing simulatedNPP to field measurements on a biome basis

LiBry does not include an explicit representation ofprocesses that directly result in emission of nitrous ox-ide However it has been determined experimentally byLenhart et al (2015) that N2O emissions by lichens andbryophytes are related to their respiration by a conversionfactor of 16 ng N2O (mg CO2)minus1 The conversion factor hasa 90 confidence interval of 11 to 21 ng N2O (mg CO2)minus1Since LiBry explicitly calculates respiration by lichens andbryophytes we derive N2O emissions from simulated respi-ration using the conversion factor of Lenhart et al (2015)

The study by Lenhart et al (2015) uses NPP of lichensand bryophytes together with free-living cyanobacteria andalgae to estimate N2O emissions since global upscaled dataon respiration of these organisms are not available from El-bert et al (2012) Thereby Lenhart et al (2015) assume afixed ratio of respiration to NPP To determine this ratio theyevaluate literature data obtaining a rate of respiration rela-tive to net photosynthesis of sim 49 (Lenhart et al 2015Table S6) Since measurements have been made in the sun-light but respiration continues in the dark respiration is mul-tiplied by a factor of 2 assuming a 12 h day This leads to anestimated ratio of respiration to NPP which is roughly 1 1To evaluate LiBry further we compute the ratio of respirationto NPP in LiBry to assess whether the model is in agreementwith these observations

In the study of Lenhart et al (2015) the substrate of thesamples was removed to avoid biases resulting from N2Orelease by microbes in the substrate Foliose and fruticoselichens as well as mosses were collected which grew on soilrocks and epiphytically on trees and the authors found novariation in N2O emissions depending on the underlying sub-strate Endolithic and crustose lichens were not included inthat study as for these growth forms the dry weight which isneeded for calculations could not be determined in a reliablemanner

Variation in field measurements of N2O emissions may re-sult not only from physiological differences between speciesbut also from variation in climatic conditions which canbe significant at the small scale To upscale emissions frompoint measurements to the large scale it is important to quan-tify the relative contributions of these different sources ofvariation When for instance the variation between speciesregarding their N2O emissions is small it suffices to samplea low number of species to obtain an average emission for acertain climatic condition LiBry can provide an indication ofthe relative importance of these sources of variation since themodel not only represents climate variability but also simu-lates diverse physiological strategies Each grid cell of themodel contains a range of surviving species at the end ofthe simulation and that consequently shows a range of N2Oemissions LiBry does not simulate spatial variation in cli-

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1596 P Porada et al Nitrous oxide emissions by lichens and bryophytes

matic conditions within a grid cell However by comparingaverage emission rates of grid cells from different climates itis possible to assess the relative importance of climatic con-ditions for variation in N2O emissions We select five modelgrid cells from different ecosystem classes to analyse the rel-ative importance of differences between species and climaticheterogeneity on variation in N2O emissions It should bepointed out that LiBry does not compute directly N2O emis-sions by lichens and bryophytes but it derives them fromsimulated respiration through an empirical linear relation-ship Hence differences in the sensitivities of respiration andN2O emissions to climatic conditions may lead to uncertain-ties in our predicted effects of climate on N2O emissions

3 Results

The global distribution of net primary productivity simulatedby the updated version of LiBry is shown in Fig 1 Produc-tivity by lichens and bryophytes is highest in forested re-gions and lowest in deserts and agricultural regions Hencethe spatial pattern is mainly controlled by water availabil-ity except for cropland In LiBry it is assumed that lichensand bryophytes only grow on the area fraction of a grid cellwhich is not occupied by crops Therefore on a grid cell ba-sis regions with a high fractional cover of cropland show lowproductivity by lichens and bryophytes in spite of favourableclimatic conditions The high productivity in the humid trop-ics mainly results from epiphytic lichens and bryophytes inthe canopy while in the boreal zone the larger fraction ofproductivity stems from the ground

As a result of the dynamic surface coverage the spatialpatterns of NPP differ slightly between the new and the orig-inal version of LiBry but the large-scale gradients remain thesame Comparing the global pattern of lichen and bryophyteNPP simulated by the new version of LiBry to an empiri-cal estimate by Elbert et al (2012) shows good agreementsimilar to the original version Furthermore the total globalNPP predicted by the new LiBry differs from the original es-timate due to the updated calculation of coverage The maindifference is found for the tropical forest canopy where sim-ulated NPP increases significantly The total global NPP of43 (Gt C) yearminus1 estimated by the new LiBry compares wellto the value of 39 (Gt C) yearminus1 calculated by Elbert et al(2012)

Comparison of simulated NPP to field measurements ona biome basis suggests that LiBry predicts realistic valuesof NPP for a range of ecosystems (Fig 2) In particularsimulated NPP in the tropical and the boreal forest matcheswell with observations while the original version of LiBryseemed to underestimate NPP in these biomes In the biomesdesert and to a lesser extent tundra LiBry seems to overes-timate productivity which may have also been the case withthe original version A potential explanation for this is thatproductivity in dry and cold areas may be limited not only by

NPP

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

a)(

10 100 1000

[g C mminus2 yrminus1]

NPP in canopy

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

b)(

10 100 1000

[g C mminus2 yrminus1]

NPP on ground

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

c)(

10 100 1000

[g C mminus2 yrminus1]

Figure 1 Global patterns of NPP Lichen and bryophyte NPP esti-mated by LiBry for (a) all locations of growth (b) the canopy and(c) the ground The estimates are in grams of carbon per square me-tre and they are average values over the last 50 years of a 600-yearsimulation with 3000 initial species Grey colour denotes regionswhere no simulated species is able to survive such as ice shieldsand the driest regions of deserts

climatic factors but also by nutrient availability (Porada et al2016b) Since photosynthesis and growth are only controlledby climatic factors in LiBry the effect of spatial variation innutrient availability on productivity cannot yet be simulatedIt should be pointed out however that except for the borealbiome the number of field measurements is quite low andconsequently the observation-based characteristic values foreach biome are subject to considerable uncertainty

Figure 3 shows simulated global patterns of nitrous ox-ide emissions by lichens and bryophytes Nitrous oxideemission is highest in the humid tropics and subtropicswith values up to 10 (mg N2O) mminus2 yearminus1 (Fig 3a) Asecond region of high emissions is the boreal zone withvalues up to 8 (mg N2O) mminus2 yearminus1 Dry regions showlowest values of nitrous oxide emissions in general lessthan 1 (mg N2O) mminus2 yearminus1 Considering only lichens andbryophytes which grow as epiphytes in the canopy (Fig 3b)

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1597

001

01

1

10

100

1000

Tundra Boreal f orestf loor

Desert Tropical f orestcanopy

Net carb

on u

pta

ke [(g

C)

mndash 2

yrndash 1

]

16

69

9

2

Figure 2 Comparison of LiBry estimates to field measurementsNPP estimated by LiBry compared to field measurements from fourbiomes defined after Olson et al (2001) The blue dots show theaverage simulated NPP for each biome and the blue vertical barsshow the range of NPP values between the different grid cells ina biome The magenta diamonds correspond to the median of NPPvalues measured in the field on the small scale the magenta verticalbars denote the range of the field measurements Left of the ma-genta diamonds the number of field measurements is shown that isconsidered for the respective biome Details can be found in Poradaet al (2013)

emissions in the humid tropics are around 3 times higher thanin the boreal and temperate zones Lichens and bryophytes onthe ground show highest values of nitrous oxide emissions inthe boreal zone with values around 3 (mg N2O) mminus2 yearminus1

(Fig 3c) Regarding the ground tropical and subtropical re-gions only partly show N2O emissions comparable to thoseof the boreal zone The reason for this is low simulatedproductivity and coverage of lichens and bryophytes on theground in tropical and subtropical climates which also leadsto low respiration on a grid cell level and hence to low N2Oemissions

Figure 4 shows the simulated global spatial distribution ofthe ratio of respiration to NPP The assumption of a globallyconstant ratio of respiration to NPP is used by Lenhart et al(2015) to derive ecosystem-scale N2O emissions by lichensand bryophytes from their NPP Alternatively this ratio canbe derived from the independent LiBry estimates of NPP andrespiration The simulated ratio shows a latitudinal patternwith increasing values towards the tropics (Fig 4a) This re-sults from the influence of surface temperature on respira-tion in combination with high nighttime temperatures in thehumid tropics which cause high respiration rates during thenight Note that high respiration relative to NPP of tropicallichens and bryophytes does not necessarily mean high res-piration at the grid cell level since net productivity and cov-erage may be low Respiration by lichens and bryophytes inthe canopy shows a slightly weaker latitudinal gradient thanon the ground which can be explained by efficient evapo-rative cooling in the canopy (Fig 4b) In contrast lichens

N2O release

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

(a)

00 15 30 45 60 75 90 105 120 135 150

[mg mminus2 yrminus1]

N2O release in canopy

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

(b)

00 10 20 30 40 50 60 70 80 90 100

[mg mminus2 yrminus1]

N2O release on ground

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

(c)

00 05 10 15 20 25 30 35 40 45 50

[mg mminus2 yrminus1]

Figure 3 Global patterns of N2O release Nitrous oxide emissionsby lichens and bryophytes estimated by LiBry for (a) all locations ofgrowth (b) the canopy and (c) the ground Note the differing rangesof the colour bars Grey colour denotes regions where no simulatedspecies is able to survive such as ice shields and the driest regionsof deserts

and bryophytes on the ground usually grow within the sur-face boundary layer which reduces cooling by turbulent heattransfer leading to a strong influence of incoming radiationon surface temperature Since radiation input increases to-ward the Equator the ratio of respiration to NPP on theground in the tropics is markedly higher than at high lati-tudes (Fig 4c) The ratio of respiration to NPP varies fromless than 1 to around 2 while most values are around 1 Thismeans that gross primary productivity (GPP) is partitionedroughly equally into NPP and respiration which agrees wellwith the observational data from Lenhart et al (2015)

An overview of global total values of N2O emissions res-piration NPP and the ratio of respiration to NPP estimatedby LiBry is shown in Table 1 Table 2 shows N2O emis-sions by lichens and bryophytes for individual grid cells fromfive different ecosystem classes (see also Table 1) Variationin emissions between species within a grid cell is large itcan exceed 3 orders of magnitude The variation due to cli-

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1598 P Porada et al Nitrous oxide emissions by lichens and bryophytes

Table 1 Annual global total values of N2O emissions NPP respiration and the ratio of respiration to NPP estimated by LiBry and separatedinto lichens and bryophytes living in the canopy and on the ground The values in brackets in the first column show the uncertainty in N2Oemissions due to the conversion of released CO2 to N2O (90 confidence interval from Lenhart et al (2015)) Ecosystem classes shown arebased on the categories made by Olson et al (2001) which were aggregated by us in the same way as in Elbert et al (2012) ldquoGt Crdquo standsfor gigatons of carbon

N2O emissions NPP Respiration Respiration NPP(Tg N2O) yearminus1 (Gt C) yearminus1 (Gt C) yearminus1 [ ]

Canopy+ groundGlobal 027 (019ndash035) 43 45 110Tropical forest 011 (008ndash014) 15 18 133Extratropical forest 011 (008ndash014) 20 18 093Steppe amp savannah 003 (002ndash004) 04 04 121Desert 002 (001ndash003) 04 04 105Tundra 001 (0007ndash0013) 02 02 087

Canopy global 013 (009ndash017) 21 22 101Ground global 014 (010ndash018) 22 23 116

Table 2 Simulated nitrous oxide emissions by lichens and bryophytes in (mg N2O) mminus2 yearminus1 for individual grid cells of the LiBry modelThe values are averages over the last 50 years of a 600-year simulation with 3000 initial species Grid cells are selected from five differentecosystem classes In the two forest classes emissions are separated into canopy and ground In the other classes the model does notrepresent lichens and bryophytes in the canopy The range of N2O emissions based on all surviving artificial species in a grid cell is shownThe average value for all species in a grid cell is derived by an NPP-based weighting scheme (see Sect 2)

Ecosystem class Location Minimum Average Maximum

Tropical forest Central Amazon ground 031 059 088canopy 0081 33 82

Extratropical forest West Siberia ground 0023 17 48canopy 00040 21 62

Steppe amp savannah Central Sahel 00095 0088 032Desert Central Australia 0019 16 55Tundra North Alaska 0012 0095 017

matic conditions is smaller but it still amounts to almost 2orders of magnitude based on the grid cells with the high-est and lowest average emission rates Comparing Table 2 tothe global range of N2O emissions by lichens and bryophytes(Fig 3) shows that the five selected grid cells represent wellthe global variation in emissions due to climatic conditionsThus both functional diversity of the artificial species anddifferent climatic conditions are important for variation ofN2O emissions according to the LiBry simulation

4 Discussion

In this study we estimate nitrous oxide emissions by lichensand bryophytes with the global process-based non-vascularvegetation model LiBry Thereby we derive N2O emissionsfrom respiration fluxes which are together with photosyn-thesis and net primary productivity simulated by LiBry

We use an updated version of LiBry which contains signif-icant modifications with regard to the original version pub-lished in Porada et al (2013) Regarding NPP the new ver-

sion estimates 43 (Gt C) yearminus1 while the original versionof LiBry predicted a range of 034 to 33 (Gt C) yearminus1 Theincrease in predicted NPP is mainly attributed to a highersimulated productivity in the tropical forest canopy sincea new disturbance scheme allows for a higher surface cov-erage of lichens and bryophytes there An empirical globalestimate of NPP by lichens bryophytes free-living terres-trial cyanobacteria and algae (Elbert et al 2012) amountsto 39 (Gt C) yearminus1 Our new estimate is higher than thatby Elbert et al (2012) although LiBry does not considerfree-living cyanobacteria and algae This may be explainedby the small contribution of cyanobacteria and algae to theoverall global carbon uptake which can be compensatedfor by minor relative changes in productivity of lichensand bryophytes (Darrouzet-Nardi et al 2015 Sancho et al2016) It is not straightforward to determine which numberis closest to reality since both the process-based estimate byLiBry and the empirical one by Elbert et al (2012) are sub-ject to uncertainty In the study of Elbert et al (2012) forinstance it is assumed that productivity and active time areuniform within a biome Furthermore Elbert et al (2012) use

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1599

Ratio respiration to NPP

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

a)(

00 03 06 09 12 15 18 21 24

[ ]

Ratio respiration to NPP in canopy

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

b)(

00 03 06 09 12 15 18 21 24

[ ]

Ratio respiration to NPP on ground

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

c)(

00 03 06 09 12 15 18 21 24

[ ]

Figure 4 Global patterns of the ratio of respiration to NPP Ratioof respiration to NPP of lichens and bryophytes estimated by LiBryfor (a) all locations of growth (b) the canopy and (c) the ground

a globally uniform value of surface cover fraction to scaleup local field measurements of productivity to the globalscale However values of surface coverage by lichens andbryophytes compiled by Elbert et al (2012) vary greatly atthe small scale which makes upscaling to larger scales chal-lenging

While productivity estimated by LiBry is evaluated in thisstudy large-scale surface coverage of lichens and bryophytessimulated by LiBry has been evaluated for regions north of50 N in Porada et al (2016a) It was shown that LiBry pre-dicts realistic values of cover fraction Moreover values ofsurface cover predicted by LiBry for other regions of theworld (Porada et al 2016b) are in agreement with the es-timate of Elbert et al (2012) In spite of uncertainties regard-ing productivity and abundance of lichens and bryophytescomparing the empirical and process-based approaches givesconfidence in the order of magnitude of the LiBry simulationresults

As a 50-year steady-state average value we estimate to-tal N2O emissions by lichens and bryophytes of 027 (019ndash035) (Tg N2O) yearminus1 which is at the lower end of the rangeof 032 to 059 (Tg N2O) yearminus1 calculated by Lenhart et al(2015) The evaluation of LiBry regarding simulated NPPshows that our global patterns and total values of NPP arevery similar to the empirical estimate by Elbert et al (2012)Since Lenhart et al (2015) use this NPP estimate by Elbertet al (2012) to derive N2O emissions differences in NPPare most likely not the reason for our lower estimate of N2Oemissions compared to Lenhart et al (2015) Instead thismay be explained by differing methods to compute respira-tion while Lenhart et al (2015) assume a globally uniformratio of respiration to NPP of a value of 2 to estimate respira-tion LiBry simulates respiration independently as a species-specific function of temperature and water status This re-sults in a lower global average value of around 1 for the ra-tio of respiration to NPP predicted by LiBry Our estimatedratio of respiration to NPP agrees well with laboratory mea-surements but it is in general difficult to compare a globalecosystem-scale value to small-scale and short-term obser-vations

Our simulated global pattern of N2O emissions is slightlydifferent than that shown in Lenhart et al (2015) who esti-mate highest values in the boreal zone and only intermedi-ate values in the humid tropics This can be explained bytheir assumed constant ratio of respiration to NPP whichmakes their global pattern of N2O emissions identical to thatof NPP which is shown in Elbert et al (2012) In LiBryhowever the simulated ratio of respiration to NPP increasestowards higher surface temperatures in the tropics (Fig 4)Furthermore the ratio shows large spatial variation Evaluat-ing this simulated pattern is difficult since estimates whichare extrapolated to the large scale such as the NPP estimateby Elbert et al (2012) are not available for respiration bylichens and bryophytes However observed ratios of respira-tion to NPP of lichens and bryophytes vary considerably atthe species level as shown by for example Lenhart et al(2015) Using a constant ratio of respiration to NPP maytherefore introduce a bias in the estimated spatial distribu-tion of N2O emissions

Small-scale measurements of N2O emissions by lichensand bryophytes may show considerable variation Thesources of this variation may be physiological differencesbetween species variation of associated microbial communi-ties andor heterogeneity in climatic conditions We examinethe relative importance for respiration of differences betweenspecies compared to climatic differences with LiBry sincethe model simulates various physiological strategies and rep-resents variation in climatic conditions at the global scaleThereby we assume that the relationship between respira-tion and N2O emissions is relatively insensitive to climaticconditions and physiological differences between species assuggested by the experiments by Lenhart et al (2015) Ta-ble 2 shows that both differences between artificial species

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1600 P Porada et al Nitrous oxide emissions by lichens and bryophytes

and different climatic conditions are important for variationof N2O emissions Upscaling of N2O emission rates mea-sured in the field may therefore be subject to considerableuncertainty Modelling approaches in this direction shouldprobably account for both interspecific variation in processesassociated with N2O release by lichens and bryophytes aswell as variation in climatic conditions

Although our approach considers the most importantsources of variation in N2O emissions by lichens andbryophytes it is associated with uncertainties that should bediscussed further These uncertainties mainly result from ourmethod to estimate respiration and from assumptions con-cerning the empirical relationship between respiration andN2O emissions

Respiration and the ratio of respiration to NPP simulatedby LiBry are difficult to validate since the number of labo-ratory or field studies which measure not only NPP but alsoGPP and respiration is not very high Moreover long-termmeasurements of respiration would be required to determinethe ratio of respiration to NPP Otherwise assumptions aboutthe contribution of respiration in the dark to total respirationare necessary

To obtain N2O emissions from respiration our resultsrely on the laboratory incubation measurements and the cal-culated ratio of N2O emissions to respiration presented inLenhart et al (2015) Furthermore our approach considerseffects of variation in climatic conditions on N2O emissionsby lichens and bryophytes Hence it is necessary to dis-cuss the sensitivity of the relationship between respirationand N2O emissions to a range of climatic conditions Asshown in Lenhart et al (2015 Fig 3) the relationship be-tween respiration and N2O emissions seems to be insensi-tive to temperature changes for the tested species Likewisevariations in water content have no clear effect on the rela-tionship between N2O release and respiration (Lenhart et al2015 Fig S3) Although the sensitivities of N2O release totemperature and water content are similar to those of respira-tion across species the relationship between N2O release andrespiration shows interspecific variation However in spite ofa large number of around 40 sampled species the relation-ship shows a relatively narrow 90 confidence interval of113 to 207 ng N2O (mg CO2)minus1 (Lenhart et al 2015) Thissuggests that the mechanism of N2O release by lichens andbryophytes is similar between different species

To analyse the relation between the production of N2Oand respiratory CO2 in greater detail measurements of bothfluxes by means of online gas exchange measurements wouldbe needed which then could be linked to the observed wa-ter status of the organisms Since LiBry explicitly representsthe dynamic water saturation of lichens and bryophytes thiswould allow a more process-based prediction of the durationand magnitude of N2O emissions In this way the uncertaintyassociated with our approach would be reduced facilitatingan improved estimate of global N2O emissions by lichensand bryophytes

In order to assess model-based estimates of N2O emis-sions by lichens and bryophytes a relatively large number offield measurements are necessary Currently most N2O mea-surements independent of the substrate or organisms mea-sured generally suffer from major uncertainties additionallyto variation from functional diversity and differing climaticconditions first the majority of these studies have been con-ducted using the acetylene inhibition technique The idea ofthis method is to inhibit the last denitrification step so thatthe measured N2O amounts should reveal the sum of N2Oand N2 release during denitrification under natural condi-tions It has however been shown quite a while ago thatthis method leads to an underestimation of denitrificationunder oxic conditions (Bollmann and Conrad 1997) Sec-ondly the most widely used measuring technique has beenthe closed-chamber method which is inexpensive and easyto use This however has major shortcomings as environ-mental conditions are hard to control and only limited surfaceareas can be measured (Butterbach-Bahl et al 2013 Groff-man 2012) Furthermore the limited temporal resolution ofchamber measurements may affect estimated N2O emissions(Barton et al 2015) Thirdly depending on the environmen-tal conditions under which the experiment is performed ndashparticularly water temperature and nutrient conditions ndash theobtained N2O emission rates could differ widely Thus it isindispensable to report and consider the exact environmen-tal conditions under which the measurements were made andto restrict natural emission data to those assessed under typi-cally occurring natural conditions

Respiration by lichens and bryophytes is not the only pro-cess which can be used to estimate their N2O emissionsBarger et al (2013) report a relationship between nitrogenfixation and N2O release in biological soil crusts which in-clude lichens and bryophytes as well as soil bacteria andalgae For this approach however reliable nitrogen fixationdata are sparse It is also possible to estimate the demand fornitrogen by lichens and bryophytes with LiBry with an un-certainty range of around 1 order of magnitude (Porada et al2014) However it is not straightforward to derive realizednitrogen uptake or nitrogen fixation from this since LiBrydoes not yet include processes related to nitrogen uptake ormetabolization of nitrogen species Therefore for this studywe chose the relation between respiration and N2O release toquantify N2O emissions by lichens and bryophytes

Our simulated global N2O emissions by lichens andbryophytes of 027 (019ndash035) (Tg N2O) yearminus1 amount toaround 3 of global N2O emissions from natural sources onland (Ciais et al 2013) This value may sound low at firstglance but it equals about 50 of the atmospheric deposi-tion of N2O into the oceans or 25 of the deposition on land(Ciais et al 2013) Considering that N2O has a strong nega-tive effect on stratospheric ozone and a significant warmingpotential as a greenhouse gas even relatively small emissionsshould not be neglected in global budgets

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1601

The study by Zhuang et al (2012) estimates global pat-terns of N2O emission from soils and finds that the hu-mid tropics contribute most to global N2O emission due tohigh temperature and precipitation Our simulated pattern ofglobal N2O emissions by lichens and bryophytes also showsa hotspot in the humid tropics but the relative contributionof the boreal zone to the global flux seems to be higher thanin Zhuang et al (2012) This probably results from the highsimulated NPP in the boreal zone particularly on the groundwhich compensates for the lower respiration and thereforeN2O emission per productivity due to low temperatures Rel-ative contributions of lichens and bryophytes to N2O emis-sions are highest for ecosystems in desert regions and athigh latitudes which agrees with the results by Lenhart et al(2015)

5 Conclusions

We estimate large-scale spatial patterns and global values ofN2O emissions by lichens and bryophytes from a process-based model of their productivity and respiration Our resultssuggest a significant contribution of lichens and bryophytesto global N2O emissions albeit at the lower end of the rangeof a previous empirical estimate Since both approachesuse respiration to derive N2O emissions our lower estimatelikely results from a different method to predict respirationcompared to the empirical approach Hence while estimatesof productivity are relatively well constrained evaluatingmodels with regard to estimated respiration may improvepredictions of N2O emissions by lichens and bryophytesOne important finding derived from our simulation is that theratio of respiration to NPP by lichens and bryophytes showsspatial variation and a latitudinal gradient at the global scaleThis means that productivity and N2O emissions by the or-ganisms are not necessarily correlated and that tropical re-gions may show higher emissions than polar regions giventhe same NPP Furthermore we show that both physiologicalvariation among species and variation in climatic conditionsare relevant for variation in respiration and consequentlyN2O emissions Ecosystem-scale estimates of N2O emis-sions by lichens and bryophytes should therefore include suf-ficient ranges of species and climatic conditions to avoid bi-ased results Our results build on the empirical finding thatN2O emissions by lichens and bryophytes are linearly re-lated to their respiration This relationship is relatively in-sensitive to climatic conditions and shows no large varia-tion between species However the relationship is based onclosed-chamber measurements Therefore it would be use-ful to perform online gas exchange measurements of N2Oemissions and respiration to test the effect of climatic con-ditions on the relationship between N2O release and respi-ration Furthermore using alternative approaches to estimateN2O emissions by lichens and bryophytes may be helpful toconstrain our approach

Code and data availability The non-vascular vegetation model Li-Bry used here is combined with an interface for parallel computingwhich was developed at the Max Planck Institute for Biogeochem-istry Jena Germany LiBry without the interface is freely availableas long as the names of the copyright holders and a disclaimer aredistributed along with the code in source or binary form The codeis available from the corresponding author upon requestModel output data which are presented as maps in this study areavailable as netCDF files from the authors on request

Author contributions Philipp Porada Bettina Weber andAxel Kleidon designed the model simulations and Philipp Poradacarried them out Philipp Porada prepared the manuscript withcontributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This work has been supported by the PAGE21project grant agreement number 282700 funded by the EC SeventhFramework Programme theme FP7-ENV-2011 and the CARBOP-ERM project grant agreement number 03G0836B funded by theBMBF (German Ministry for Science and Education) The authorsthank the Max Planck Society (Nobel Laureate Fellowship forBettina Weber) for financial support The Max Planck Institute forBiogeochemistry provided computational resources

Edited by V BrovkinReviewed by two anonymous referees

References

Abed R Lam P de Beer D and Stief P High rates of den-itrification and nitrous oxide emission in arid biological soilcrusts from the Sultanate of Oman ISME J 7 1862ndash1875doi101038ismej201355 2013

Barger N Castle S and Dean G Denitrification from nitrogen-fixing biologically crusted soils in a cool desert environ-ment southeast Utah USA Ecological Processes 2 1ndash9doi1011862192-1709-2-16 2013

Barger N Weber B Garcia-Pichel F Zaady E and Belnap JPatterns and Controls on Nitrogen Cycling of Biological SoilCrusts in Biological Soil Crusts An Organizing Principle inDrylands edited by Weber B Buumldel B and Belnap J 257ndash285 Springer International Publishing Cham doi101007978-3-319-30214-0_14 2016

Barton L Wolf B Rowlings D Scheer C Kiese R Grace PStefanova K and Butterbach-Bahl K Sampling frequency af-fects estimates of annual nitrous oxide fluxes Scientific Reports5 15912 doi101038srep15912 2015

Bollmann A and Conrad R Acetylene blockage technique leadsto underestimation of denitrification rates in oxic soils due toscavenging of intermediate nitric oxide Soil Biol Biochem 291067ndash1077 doi101016S0038-0717(97)00007-2 1997

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1595

variation in N2O emissions and determine their relative im-portance

2 Methods

The non-vascular vegetation model LiBry estimates globalpatterns of photosynthesis respiration and net primary pro-ductivity of lichens and bryophytes (Porada et al 2013) Themodel calculates these physiological processes as a functionof climate and additional environmental conditions whichare provided in the form of time series of global griddedmaps Photosynthesis in LiBry is determined by ambient lev-els of light CO2 and temperature according to the Farquharapproach (Farquhar and von Caemmerer 1982) Respirationis simulated as a function of temperature via a Q10 rela-tionship Both processes also depend on the water status ofthe simulated lichens and bryophytes which includes limita-tion of CO2 diffusion at high water content NPP is derivedfrom the difference between gross photosynthesis and respi-ration A unique feature of LiBry is that functional diversityof lichens and bryophytes is represented by a large numberof artificial species instead of being aggregated into one or afew average functional types The advantage of this approachis that adaptation of the organisms to differing environmentalconditions is simulated in a more realistic way Physiologi-cal processes such as photosynthesis and respiration are cal-culated separately for each artificial species LiBry has beensuccessfully applied to estimate global NPP by lichens andbryophytes (Porada et al 2013) and other impacts of theseorganisms on global biogeochemical cycles (Porada et al2014 2016b)

The model version presented here contains several exten-sions compared to the original version first an NPP-basedweighting scheme was introduced which assigns relativeabundances to all artificial species that survive in a grid cellof the model in the steady state (Porada et al 2016b) Thisallows an average grid cell value of NPP based on the relativeabundances of the simulated species in that cell to be derivedIn the original version grid cell NPP could only be predictedin the form of a range of values due to unknown abundancesof the species The average grid cell NPP is close to the upperend of the range of productivity values since the most pro-ductive simulated species are assumed to be the most abun-dant ones Secondly a dynamic disturbance scheme was im-plemented which replaces the equilibrium computation ofsurface coverage by a monthly update of coverage (Poradaet al 2016a) This makes the new model applicable to tran-sient scenarios of climatic and environmental change whilethe original model required the assumption of a steady stateto compute coverage

For this study we run LiBry with an initial value of 3000artificial species in each grid cell for a period of 600 years toreach steady state with climatic fields and other forcing datafrom Porada et al (2013) Our global estimates are based on

average values over the last 50 years of the simulation Weevaluate the new version of LiBry in the same way as theoriginal one (Porada et al 2013) by comparing simulatedNPP to field measurements on a biome basis

LiBry does not include an explicit representation ofprocesses that directly result in emission of nitrous ox-ide However it has been determined experimentally byLenhart et al (2015) that N2O emissions by lichens andbryophytes are related to their respiration by a conversionfactor of 16 ng N2O (mg CO2)minus1 The conversion factor hasa 90 confidence interval of 11 to 21 ng N2O (mg CO2)minus1Since LiBry explicitly calculates respiration by lichens andbryophytes we derive N2O emissions from simulated respi-ration using the conversion factor of Lenhart et al (2015)

The study by Lenhart et al (2015) uses NPP of lichensand bryophytes together with free-living cyanobacteria andalgae to estimate N2O emissions since global upscaled dataon respiration of these organisms are not available from El-bert et al (2012) Thereby Lenhart et al (2015) assume afixed ratio of respiration to NPP To determine this ratio theyevaluate literature data obtaining a rate of respiration rela-tive to net photosynthesis of sim 49 (Lenhart et al 2015Table S6) Since measurements have been made in the sun-light but respiration continues in the dark respiration is mul-tiplied by a factor of 2 assuming a 12 h day This leads to anestimated ratio of respiration to NPP which is roughly 1 1To evaluate LiBry further we compute the ratio of respirationto NPP in LiBry to assess whether the model is in agreementwith these observations

In the study of Lenhart et al (2015) the substrate of thesamples was removed to avoid biases resulting from N2Orelease by microbes in the substrate Foliose and fruticoselichens as well as mosses were collected which grew on soilrocks and epiphytically on trees and the authors found novariation in N2O emissions depending on the underlying sub-strate Endolithic and crustose lichens were not included inthat study as for these growth forms the dry weight which isneeded for calculations could not be determined in a reliablemanner

Variation in field measurements of N2O emissions may re-sult not only from physiological differences between speciesbut also from variation in climatic conditions which canbe significant at the small scale To upscale emissions frompoint measurements to the large scale it is important to quan-tify the relative contributions of these different sources ofvariation When for instance the variation between speciesregarding their N2O emissions is small it suffices to samplea low number of species to obtain an average emission for acertain climatic condition LiBry can provide an indication ofthe relative importance of these sources of variation since themodel not only represents climate variability but also simu-lates diverse physiological strategies Each grid cell of themodel contains a range of surviving species at the end ofthe simulation and that consequently shows a range of N2Oemissions LiBry does not simulate spatial variation in cli-

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1596 P Porada et al Nitrous oxide emissions by lichens and bryophytes

matic conditions within a grid cell However by comparingaverage emission rates of grid cells from different climates itis possible to assess the relative importance of climatic con-ditions for variation in N2O emissions We select five modelgrid cells from different ecosystem classes to analyse the rel-ative importance of differences between species and climaticheterogeneity on variation in N2O emissions It should bepointed out that LiBry does not compute directly N2O emis-sions by lichens and bryophytes but it derives them fromsimulated respiration through an empirical linear relation-ship Hence differences in the sensitivities of respiration andN2O emissions to climatic conditions may lead to uncertain-ties in our predicted effects of climate on N2O emissions

3 Results

The global distribution of net primary productivity simulatedby the updated version of LiBry is shown in Fig 1 Produc-tivity by lichens and bryophytes is highest in forested re-gions and lowest in deserts and agricultural regions Hencethe spatial pattern is mainly controlled by water availabil-ity except for cropland In LiBry it is assumed that lichensand bryophytes only grow on the area fraction of a grid cellwhich is not occupied by crops Therefore on a grid cell ba-sis regions with a high fractional cover of cropland show lowproductivity by lichens and bryophytes in spite of favourableclimatic conditions The high productivity in the humid trop-ics mainly results from epiphytic lichens and bryophytes inthe canopy while in the boreal zone the larger fraction ofproductivity stems from the ground

As a result of the dynamic surface coverage the spatialpatterns of NPP differ slightly between the new and the orig-inal version of LiBry but the large-scale gradients remain thesame Comparing the global pattern of lichen and bryophyteNPP simulated by the new version of LiBry to an empiri-cal estimate by Elbert et al (2012) shows good agreementsimilar to the original version Furthermore the total globalNPP predicted by the new LiBry differs from the original es-timate due to the updated calculation of coverage The maindifference is found for the tropical forest canopy where sim-ulated NPP increases significantly The total global NPP of43 (Gt C) yearminus1 estimated by the new LiBry compares wellto the value of 39 (Gt C) yearminus1 calculated by Elbert et al(2012)

Comparison of simulated NPP to field measurements ona biome basis suggests that LiBry predicts realistic valuesof NPP for a range of ecosystems (Fig 2) In particularsimulated NPP in the tropical and the boreal forest matcheswell with observations while the original version of LiBryseemed to underestimate NPP in these biomes In the biomesdesert and to a lesser extent tundra LiBry seems to overes-timate productivity which may have also been the case withthe original version A potential explanation for this is thatproductivity in dry and cold areas may be limited not only by

NPP

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

a)(

10 100 1000

[g C mminus2 yrminus1]

NPP in canopy

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

b)(

10 100 1000

[g C mminus2 yrminus1]

NPP on ground

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

c)(

10 100 1000

[g C mminus2 yrminus1]

Figure 1 Global patterns of NPP Lichen and bryophyte NPP esti-mated by LiBry for (a) all locations of growth (b) the canopy and(c) the ground The estimates are in grams of carbon per square me-tre and they are average values over the last 50 years of a 600-yearsimulation with 3000 initial species Grey colour denotes regionswhere no simulated species is able to survive such as ice shieldsand the driest regions of deserts

climatic factors but also by nutrient availability (Porada et al2016b) Since photosynthesis and growth are only controlledby climatic factors in LiBry the effect of spatial variation innutrient availability on productivity cannot yet be simulatedIt should be pointed out however that except for the borealbiome the number of field measurements is quite low andconsequently the observation-based characteristic values foreach biome are subject to considerable uncertainty

Figure 3 shows simulated global patterns of nitrous ox-ide emissions by lichens and bryophytes Nitrous oxideemission is highest in the humid tropics and subtropicswith values up to 10 (mg N2O) mminus2 yearminus1 (Fig 3a) Asecond region of high emissions is the boreal zone withvalues up to 8 (mg N2O) mminus2 yearminus1 Dry regions showlowest values of nitrous oxide emissions in general lessthan 1 (mg N2O) mminus2 yearminus1 Considering only lichens andbryophytes which grow as epiphytes in the canopy (Fig 3b)

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1597

001

01

1

10

100

1000

Tundra Boreal f orestf loor

Desert Tropical f orestcanopy

Net carb

on u

pta

ke [(g

C)

mndash 2

yrndash 1

]

16

69

9

2

Figure 2 Comparison of LiBry estimates to field measurementsNPP estimated by LiBry compared to field measurements from fourbiomes defined after Olson et al (2001) The blue dots show theaverage simulated NPP for each biome and the blue vertical barsshow the range of NPP values between the different grid cells ina biome The magenta diamonds correspond to the median of NPPvalues measured in the field on the small scale the magenta verticalbars denote the range of the field measurements Left of the ma-genta diamonds the number of field measurements is shown that isconsidered for the respective biome Details can be found in Poradaet al (2013)

emissions in the humid tropics are around 3 times higher thanin the boreal and temperate zones Lichens and bryophytes onthe ground show highest values of nitrous oxide emissions inthe boreal zone with values around 3 (mg N2O) mminus2 yearminus1

(Fig 3c) Regarding the ground tropical and subtropical re-gions only partly show N2O emissions comparable to thoseof the boreal zone The reason for this is low simulatedproductivity and coverage of lichens and bryophytes on theground in tropical and subtropical climates which also leadsto low respiration on a grid cell level and hence to low N2Oemissions

Figure 4 shows the simulated global spatial distribution ofthe ratio of respiration to NPP The assumption of a globallyconstant ratio of respiration to NPP is used by Lenhart et al(2015) to derive ecosystem-scale N2O emissions by lichensand bryophytes from their NPP Alternatively this ratio canbe derived from the independent LiBry estimates of NPP andrespiration The simulated ratio shows a latitudinal patternwith increasing values towards the tropics (Fig 4a) This re-sults from the influence of surface temperature on respira-tion in combination with high nighttime temperatures in thehumid tropics which cause high respiration rates during thenight Note that high respiration relative to NPP of tropicallichens and bryophytes does not necessarily mean high res-piration at the grid cell level since net productivity and cov-erage may be low Respiration by lichens and bryophytes inthe canopy shows a slightly weaker latitudinal gradient thanon the ground which can be explained by efficient evapo-rative cooling in the canopy (Fig 4b) In contrast lichens

N2O release

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

(a)

00 15 30 45 60 75 90 105 120 135 150

[mg mminus2 yrminus1]

N2O release in canopy

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

(b)

00 10 20 30 40 50 60 70 80 90 100

[mg mminus2 yrminus1]

N2O release on ground

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

(c)

00 05 10 15 20 25 30 35 40 45 50

[mg mminus2 yrminus1]

Figure 3 Global patterns of N2O release Nitrous oxide emissionsby lichens and bryophytes estimated by LiBry for (a) all locations ofgrowth (b) the canopy and (c) the ground Note the differing rangesof the colour bars Grey colour denotes regions where no simulatedspecies is able to survive such as ice shields and the driest regionsof deserts

and bryophytes on the ground usually grow within the sur-face boundary layer which reduces cooling by turbulent heattransfer leading to a strong influence of incoming radiationon surface temperature Since radiation input increases to-ward the Equator the ratio of respiration to NPP on theground in the tropics is markedly higher than at high lati-tudes (Fig 4c) The ratio of respiration to NPP varies fromless than 1 to around 2 while most values are around 1 Thismeans that gross primary productivity (GPP) is partitionedroughly equally into NPP and respiration which agrees wellwith the observational data from Lenhart et al (2015)

An overview of global total values of N2O emissions res-piration NPP and the ratio of respiration to NPP estimatedby LiBry is shown in Table 1 Table 2 shows N2O emis-sions by lichens and bryophytes for individual grid cells fromfive different ecosystem classes (see also Table 1) Variationin emissions between species within a grid cell is large itcan exceed 3 orders of magnitude The variation due to cli-

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1598 P Porada et al Nitrous oxide emissions by lichens and bryophytes

Table 1 Annual global total values of N2O emissions NPP respiration and the ratio of respiration to NPP estimated by LiBry and separatedinto lichens and bryophytes living in the canopy and on the ground The values in brackets in the first column show the uncertainty in N2Oemissions due to the conversion of released CO2 to N2O (90 confidence interval from Lenhart et al (2015)) Ecosystem classes shown arebased on the categories made by Olson et al (2001) which were aggregated by us in the same way as in Elbert et al (2012) ldquoGt Crdquo standsfor gigatons of carbon

N2O emissions NPP Respiration Respiration NPP(Tg N2O) yearminus1 (Gt C) yearminus1 (Gt C) yearminus1 [ ]

Canopy+ groundGlobal 027 (019ndash035) 43 45 110Tropical forest 011 (008ndash014) 15 18 133Extratropical forest 011 (008ndash014) 20 18 093Steppe amp savannah 003 (002ndash004) 04 04 121Desert 002 (001ndash003) 04 04 105Tundra 001 (0007ndash0013) 02 02 087

Canopy global 013 (009ndash017) 21 22 101Ground global 014 (010ndash018) 22 23 116

Table 2 Simulated nitrous oxide emissions by lichens and bryophytes in (mg N2O) mminus2 yearminus1 for individual grid cells of the LiBry modelThe values are averages over the last 50 years of a 600-year simulation with 3000 initial species Grid cells are selected from five differentecosystem classes In the two forest classes emissions are separated into canopy and ground In the other classes the model does notrepresent lichens and bryophytes in the canopy The range of N2O emissions based on all surviving artificial species in a grid cell is shownThe average value for all species in a grid cell is derived by an NPP-based weighting scheme (see Sect 2)

Ecosystem class Location Minimum Average Maximum

Tropical forest Central Amazon ground 031 059 088canopy 0081 33 82

Extratropical forest West Siberia ground 0023 17 48canopy 00040 21 62

Steppe amp savannah Central Sahel 00095 0088 032Desert Central Australia 0019 16 55Tundra North Alaska 0012 0095 017

matic conditions is smaller but it still amounts to almost 2orders of magnitude based on the grid cells with the high-est and lowest average emission rates Comparing Table 2 tothe global range of N2O emissions by lichens and bryophytes(Fig 3) shows that the five selected grid cells represent wellthe global variation in emissions due to climatic conditionsThus both functional diversity of the artificial species anddifferent climatic conditions are important for variation ofN2O emissions according to the LiBry simulation

4 Discussion

In this study we estimate nitrous oxide emissions by lichensand bryophytes with the global process-based non-vascularvegetation model LiBry Thereby we derive N2O emissionsfrom respiration fluxes which are together with photosyn-thesis and net primary productivity simulated by LiBry

We use an updated version of LiBry which contains signif-icant modifications with regard to the original version pub-lished in Porada et al (2013) Regarding NPP the new ver-

sion estimates 43 (Gt C) yearminus1 while the original versionof LiBry predicted a range of 034 to 33 (Gt C) yearminus1 Theincrease in predicted NPP is mainly attributed to a highersimulated productivity in the tropical forest canopy sincea new disturbance scheme allows for a higher surface cov-erage of lichens and bryophytes there An empirical globalestimate of NPP by lichens bryophytes free-living terres-trial cyanobacteria and algae (Elbert et al 2012) amountsto 39 (Gt C) yearminus1 Our new estimate is higher than thatby Elbert et al (2012) although LiBry does not considerfree-living cyanobacteria and algae This may be explainedby the small contribution of cyanobacteria and algae to theoverall global carbon uptake which can be compensatedfor by minor relative changes in productivity of lichensand bryophytes (Darrouzet-Nardi et al 2015 Sancho et al2016) It is not straightforward to determine which numberis closest to reality since both the process-based estimate byLiBry and the empirical one by Elbert et al (2012) are sub-ject to uncertainty In the study of Elbert et al (2012) forinstance it is assumed that productivity and active time areuniform within a biome Furthermore Elbert et al (2012) use

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1599

Ratio respiration to NPP

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

a)(

00 03 06 09 12 15 18 21 24

[ ]

Ratio respiration to NPP in canopy

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

b)(

00 03 06 09 12 15 18 21 24

[ ]

Ratio respiration to NPP on ground

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

c)(

00 03 06 09 12 15 18 21 24

[ ]

Figure 4 Global patterns of the ratio of respiration to NPP Ratioof respiration to NPP of lichens and bryophytes estimated by LiBryfor (a) all locations of growth (b) the canopy and (c) the ground

a globally uniform value of surface cover fraction to scaleup local field measurements of productivity to the globalscale However values of surface coverage by lichens andbryophytes compiled by Elbert et al (2012) vary greatly atthe small scale which makes upscaling to larger scales chal-lenging

While productivity estimated by LiBry is evaluated in thisstudy large-scale surface coverage of lichens and bryophytessimulated by LiBry has been evaluated for regions north of50 N in Porada et al (2016a) It was shown that LiBry pre-dicts realistic values of cover fraction Moreover values ofsurface cover predicted by LiBry for other regions of theworld (Porada et al 2016b) are in agreement with the es-timate of Elbert et al (2012) In spite of uncertainties regard-ing productivity and abundance of lichens and bryophytescomparing the empirical and process-based approaches givesconfidence in the order of magnitude of the LiBry simulationresults

As a 50-year steady-state average value we estimate to-tal N2O emissions by lichens and bryophytes of 027 (019ndash035) (Tg N2O) yearminus1 which is at the lower end of the rangeof 032 to 059 (Tg N2O) yearminus1 calculated by Lenhart et al(2015) The evaluation of LiBry regarding simulated NPPshows that our global patterns and total values of NPP arevery similar to the empirical estimate by Elbert et al (2012)Since Lenhart et al (2015) use this NPP estimate by Elbertet al (2012) to derive N2O emissions differences in NPPare most likely not the reason for our lower estimate of N2Oemissions compared to Lenhart et al (2015) Instead thismay be explained by differing methods to compute respira-tion while Lenhart et al (2015) assume a globally uniformratio of respiration to NPP of a value of 2 to estimate respira-tion LiBry simulates respiration independently as a species-specific function of temperature and water status This re-sults in a lower global average value of around 1 for the ra-tio of respiration to NPP predicted by LiBry Our estimatedratio of respiration to NPP agrees well with laboratory mea-surements but it is in general difficult to compare a globalecosystem-scale value to small-scale and short-term obser-vations

Our simulated global pattern of N2O emissions is slightlydifferent than that shown in Lenhart et al (2015) who esti-mate highest values in the boreal zone and only intermedi-ate values in the humid tropics This can be explained bytheir assumed constant ratio of respiration to NPP whichmakes their global pattern of N2O emissions identical to thatof NPP which is shown in Elbert et al (2012) In LiBryhowever the simulated ratio of respiration to NPP increasestowards higher surface temperatures in the tropics (Fig 4)Furthermore the ratio shows large spatial variation Evaluat-ing this simulated pattern is difficult since estimates whichare extrapolated to the large scale such as the NPP estimateby Elbert et al (2012) are not available for respiration bylichens and bryophytes However observed ratios of respira-tion to NPP of lichens and bryophytes vary considerably atthe species level as shown by for example Lenhart et al(2015) Using a constant ratio of respiration to NPP maytherefore introduce a bias in the estimated spatial distribu-tion of N2O emissions

Small-scale measurements of N2O emissions by lichensand bryophytes may show considerable variation Thesources of this variation may be physiological differencesbetween species variation of associated microbial communi-ties andor heterogeneity in climatic conditions We examinethe relative importance for respiration of differences betweenspecies compared to climatic differences with LiBry sincethe model simulates various physiological strategies and rep-resents variation in climatic conditions at the global scaleThereby we assume that the relationship between respira-tion and N2O emissions is relatively insensitive to climaticconditions and physiological differences between species assuggested by the experiments by Lenhart et al (2015) Ta-ble 2 shows that both differences between artificial species

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1600 P Porada et al Nitrous oxide emissions by lichens and bryophytes

and different climatic conditions are important for variationof N2O emissions Upscaling of N2O emission rates mea-sured in the field may therefore be subject to considerableuncertainty Modelling approaches in this direction shouldprobably account for both interspecific variation in processesassociated with N2O release by lichens and bryophytes aswell as variation in climatic conditions

Although our approach considers the most importantsources of variation in N2O emissions by lichens andbryophytes it is associated with uncertainties that should bediscussed further These uncertainties mainly result from ourmethod to estimate respiration and from assumptions con-cerning the empirical relationship between respiration andN2O emissions

Respiration and the ratio of respiration to NPP simulatedby LiBry are difficult to validate since the number of labo-ratory or field studies which measure not only NPP but alsoGPP and respiration is not very high Moreover long-termmeasurements of respiration would be required to determinethe ratio of respiration to NPP Otherwise assumptions aboutthe contribution of respiration in the dark to total respirationare necessary

To obtain N2O emissions from respiration our resultsrely on the laboratory incubation measurements and the cal-culated ratio of N2O emissions to respiration presented inLenhart et al (2015) Furthermore our approach considerseffects of variation in climatic conditions on N2O emissionsby lichens and bryophytes Hence it is necessary to dis-cuss the sensitivity of the relationship between respirationand N2O emissions to a range of climatic conditions Asshown in Lenhart et al (2015 Fig 3) the relationship be-tween respiration and N2O emissions seems to be insensi-tive to temperature changes for the tested species Likewisevariations in water content have no clear effect on the rela-tionship between N2O release and respiration (Lenhart et al2015 Fig S3) Although the sensitivities of N2O release totemperature and water content are similar to those of respira-tion across species the relationship between N2O release andrespiration shows interspecific variation However in spite ofa large number of around 40 sampled species the relation-ship shows a relatively narrow 90 confidence interval of113 to 207 ng N2O (mg CO2)minus1 (Lenhart et al 2015) Thissuggests that the mechanism of N2O release by lichens andbryophytes is similar between different species

To analyse the relation between the production of N2Oand respiratory CO2 in greater detail measurements of bothfluxes by means of online gas exchange measurements wouldbe needed which then could be linked to the observed wa-ter status of the organisms Since LiBry explicitly representsthe dynamic water saturation of lichens and bryophytes thiswould allow a more process-based prediction of the durationand magnitude of N2O emissions In this way the uncertaintyassociated with our approach would be reduced facilitatingan improved estimate of global N2O emissions by lichensand bryophytes

In order to assess model-based estimates of N2O emis-sions by lichens and bryophytes a relatively large number offield measurements are necessary Currently most N2O mea-surements independent of the substrate or organisms mea-sured generally suffer from major uncertainties additionallyto variation from functional diversity and differing climaticconditions first the majority of these studies have been con-ducted using the acetylene inhibition technique The idea ofthis method is to inhibit the last denitrification step so thatthe measured N2O amounts should reveal the sum of N2Oand N2 release during denitrification under natural condi-tions It has however been shown quite a while ago thatthis method leads to an underestimation of denitrificationunder oxic conditions (Bollmann and Conrad 1997) Sec-ondly the most widely used measuring technique has beenthe closed-chamber method which is inexpensive and easyto use This however has major shortcomings as environ-mental conditions are hard to control and only limited surfaceareas can be measured (Butterbach-Bahl et al 2013 Groff-man 2012) Furthermore the limited temporal resolution ofchamber measurements may affect estimated N2O emissions(Barton et al 2015) Thirdly depending on the environmen-tal conditions under which the experiment is performed ndashparticularly water temperature and nutrient conditions ndash theobtained N2O emission rates could differ widely Thus it isindispensable to report and consider the exact environmen-tal conditions under which the measurements were made andto restrict natural emission data to those assessed under typi-cally occurring natural conditions

Respiration by lichens and bryophytes is not the only pro-cess which can be used to estimate their N2O emissionsBarger et al (2013) report a relationship between nitrogenfixation and N2O release in biological soil crusts which in-clude lichens and bryophytes as well as soil bacteria andalgae For this approach however reliable nitrogen fixationdata are sparse It is also possible to estimate the demand fornitrogen by lichens and bryophytes with LiBry with an un-certainty range of around 1 order of magnitude (Porada et al2014) However it is not straightforward to derive realizednitrogen uptake or nitrogen fixation from this since LiBrydoes not yet include processes related to nitrogen uptake ormetabolization of nitrogen species Therefore for this studywe chose the relation between respiration and N2O release toquantify N2O emissions by lichens and bryophytes

Our simulated global N2O emissions by lichens andbryophytes of 027 (019ndash035) (Tg N2O) yearminus1 amount toaround 3 of global N2O emissions from natural sources onland (Ciais et al 2013) This value may sound low at firstglance but it equals about 50 of the atmospheric deposi-tion of N2O into the oceans or 25 of the deposition on land(Ciais et al 2013) Considering that N2O has a strong nega-tive effect on stratospheric ozone and a significant warmingpotential as a greenhouse gas even relatively small emissionsshould not be neglected in global budgets

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1601

The study by Zhuang et al (2012) estimates global pat-terns of N2O emission from soils and finds that the hu-mid tropics contribute most to global N2O emission due tohigh temperature and precipitation Our simulated pattern ofglobal N2O emissions by lichens and bryophytes also showsa hotspot in the humid tropics but the relative contributionof the boreal zone to the global flux seems to be higher thanin Zhuang et al (2012) This probably results from the highsimulated NPP in the boreal zone particularly on the groundwhich compensates for the lower respiration and thereforeN2O emission per productivity due to low temperatures Rel-ative contributions of lichens and bryophytes to N2O emis-sions are highest for ecosystems in desert regions and athigh latitudes which agrees with the results by Lenhart et al(2015)

5 Conclusions

We estimate large-scale spatial patterns and global values ofN2O emissions by lichens and bryophytes from a process-based model of their productivity and respiration Our resultssuggest a significant contribution of lichens and bryophytesto global N2O emissions albeit at the lower end of the rangeof a previous empirical estimate Since both approachesuse respiration to derive N2O emissions our lower estimatelikely results from a different method to predict respirationcompared to the empirical approach Hence while estimatesof productivity are relatively well constrained evaluatingmodels with regard to estimated respiration may improvepredictions of N2O emissions by lichens and bryophytesOne important finding derived from our simulation is that theratio of respiration to NPP by lichens and bryophytes showsspatial variation and a latitudinal gradient at the global scaleThis means that productivity and N2O emissions by the or-ganisms are not necessarily correlated and that tropical re-gions may show higher emissions than polar regions giventhe same NPP Furthermore we show that both physiologicalvariation among species and variation in climatic conditionsare relevant for variation in respiration and consequentlyN2O emissions Ecosystem-scale estimates of N2O emis-sions by lichens and bryophytes should therefore include suf-ficient ranges of species and climatic conditions to avoid bi-ased results Our results build on the empirical finding thatN2O emissions by lichens and bryophytes are linearly re-lated to their respiration This relationship is relatively in-sensitive to climatic conditions and shows no large varia-tion between species However the relationship is based onclosed-chamber measurements Therefore it would be use-ful to perform online gas exchange measurements of N2Oemissions and respiration to test the effect of climatic con-ditions on the relationship between N2O release and respi-ration Furthermore using alternative approaches to estimateN2O emissions by lichens and bryophytes may be helpful toconstrain our approach

Code and data availability The non-vascular vegetation model Li-Bry used here is combined with an interface for parallel computingwhich was developed at the Max Planck Institute for Biogeochem-istry Jena Germany LiBry without the interface is freely availableas long as the names of the copyright holders and a disclaimer aredistributed along with the code in source or binary form The codeis available from the corresponding author upon requestModel output data which are presented as maps in this study areavailable as netCDF files from the authors on request

Author contributions Philipp Porada Bettina Weber andAxel Kleidon designed the model simulations and Philipp Poradacarried them out Philipp Porada prepared the manuscript withcontributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This work has been supported by the PAGE21project grant agreement number 282700 funded by the EC SeventhFramework Programme theme FP7-ENV-2011 and the CARBOP-ERM project grant agreement number 03G0836B funded by theBMBF (German Ministry for Science and Education) The authorsthank the Max Planck Society (Nobel Laureate Fellowship forBettina Weber) for financial support The Max Planck Institute forBiogeochemistry provided computational resources

Edited by V BrovkinReviewed by two anonymous referees

References

Abed R Lam P de Beer D and Stief P High rates of den-itrification and nitrous oxide emission in arid biological soilcrusts from the Sultanate of Oman ISME J 7 1862ndash1875doi101038ismej201355 2013

Barger N Castle S and Dean G Denitrification from nitrogen-fixing biologically crusted soils in a cool desert environ-ment southeast Utah USA Ecological Processes 2 1ndash9doi1011862192-1709-2-16 2013

Barger N Weber B Garcia-Pichel F Zaady E and Belnap JPatterns and Controls on Nitrogen Cycling of Biological SoilCrusts in Biological Soil Crusts An Organizing Principle inDrylands edited by Weber B Buumldel B and Belnap J 257ndash285 Springer International Publishing Cham doi101007978-3-319-30214-0_14 2016

Barton L Wolf B Rowlings D Scheer C Kiese R Grace PStefanova K and Butterbach-Bahl K Sampling frequency af-fects estimates of annual nitrous oxide fluxes Scientific Reports5 15912 doi101038srep15912 2015

Bollmann A and Conrad R Acetylene blockage technique leadsto underestimation of denitrification rates in oxic soils due toscavenging of intermediate nitric oxide Soil Biol Biochem 291067ndash1077 doi101016S0038-0717(97)00007-2 1997

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1596 P Porada et al Nitrous oxide emissions by lichens and bryophytes

matic conditions within a grid cell However by comparingaverage emission rates of grid cells from different climates itis possible to assess the relative importance of climatic con-ditions for variation in N2O emissions We select five modelgrid cells from different ecosystem classes to analyse the rel-ative importance of differences between species and climaticheterogeneity on variation in N2O emissions It should bepointed out that LiBry does not compute directly N2O emis-sions by lichens and bryophytes but it derives them fromsimulated respiration through an empirical linear relation-ship Hence differences in the sensitivities of respiration andN2O emissions to climatic conditions may lead to uncertain-ties in our predicted effects of climate on N2O emissions

3 Results

The global distribution of net primary productivity simulatedby the updated version of LiBry is shown in Fig 1 Produc-tivity by lichens and bryophytes is highest in forested re-gions and lowest in deserts and agricultural regions Hencethe spatial pattern is mainly controlled by water availabil-ity except for cropland In LiBry it is assumed that lichensand bryophytes only grow on the area fraction of a grid cellwhich is not occupied by crops Therefore on a grid cell ba-sis regions with a high fractional cover of cropland show lowproductivity by lichens and bryophytes in spite of favourableclimatic conditions The high productivity in the humid trop-ics mainly results from epiphytic lichens and bryophytes inthe canopy while in the boreal zone the larger fraction ofproductivity stems from the ground

As a result of the dynamic surface coverage the spatialpatterns of NPP differ slightly between the new and the orig-inal version of LiBry but the large-scale gradients remain thesame Comparing the global pattern of lichen and bryophyteNPP simulated by the new version of LiBry to an empiri-cal estimate by Elbert et al (2012) shows good agreementsimilar to the original version Furthermore the total globalNPP predicted by the new LiBry differs from the original es-timate due to the updated calculation of coverage The maindifference is found for the tropical forest canopy where sim-ulated NPP increases significantly The total global NPP of43 (Gt C) yearminus1 estimated by the new LiBry compares wellto the value of 39 (Gt C) yearminus1 calculated by Elbert et al(2012)

Comparison of simulated NPP to field measurements ona biome basis suggests that LiBry predicts realistic valuesof NPP for a range of ecosystems (Fig 2) In particularsimulated NPP in the tropical and the boreal forest matcheswell with observations while the original version of LiBryseemed to underestimate NPP in these biomes In the biomesdesert and to a lesser extent tundra LiBry seems to overes-timate productivity which may have also been the case withthe original version A potential explanation for this is thatproductivity in dry and cold areas may be limited not only by

NPP

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

a)(

10 100 1000

[g C mminus2 yrminus1]

NPP in canopy

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

b)(

10 100 1000

[g C mminus2 yrminus1]

NPP on ground

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

c)(

10 100 1000

[g C mminus2 yrminus1]

Figure 1 Global patterns of NPP Lichen and bryophyte NPP esti-mated by LiBry for (a) all locations of growth (b) the canopy and(c) the ground The estimates are in grams of carbon per square me-tre and they are average values over the last 50 years of a 600-yearsimulation with 3000 initial species Grey colour denotes regionswhere no simulated species is able to survive such as ice shieldsand the driest regions of deserts

climatic factors but also by nutrient availability (Porada et al2016b) Since photosynthesis and growth are only controlledby climatic factors in LiBry the effect of spatial variation innutrient availability on productivity cannot yet be simulatedIt should be pointed out however that except for the borealbiome the number of field measurements is quite low andconsequently the observation-based characteristic values foreach biome are subject to considerable uncertainty

Figure 3 shows simulated global patterns of nitrous ox-ide emissions by lichens and bryophytes Nitrous oxideemission is highest in the humid tropics and subtropicswith values up to 10 (mg N2O) mminus2 yearminus1 (Fig 3a) Asecond region of high emissions is the boreal zone withvalues up to 8 (mg N2O) mminus2 yearminus1 Dry regions showlowest values of nitrous oxide emissions in general lessthan 1 (mg N2O) mminus2 yearminus1 Considering only lichens andbryophytes which grow as epiphytes in the canopy (Fig 3b)

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1597

001

01

1

10

100

1000

Tundra Boreal f orestf loor

Desert Tropical f orestcanopy

Net carb

on u

pta

ke [(g

C)

mndash 2

yrndash 1

]

16

69

9

2

Figure 2 Comparison of LiBry estimates to field measurementsNPP estimated by LiBry compared to field measurements from fourbiomes defined after Olson et al (2001) The blue dots show theaverage simulated NPP for each biome and the blue vertical barsshow the range of NPP values between the different grid cells ina biome The magenta diamonds correspond to the median of NPPvalues measured in the field on the small scale the magenta verticalbars denote the range of the field measurements Left of the ma-genta diamonds the number of field measurements is shown that isconsidered for the respective biome Details can be found in Poradaet al (2013)

emissions in the humid tropics are around 3 times higher thanin the boreal and temperate zones Lichens and bryophytes onthe ground show highest values of nitrous oxide emissions inthe boreal zone with values around 3 (mg N2O) mminus2 yearminus1

(Fig 3c) Regarding the ground tropical and subtropical re-gions only partly show N2O emissions comparable to thoseof the boreal zone The reason for this is low simulatedproductivity and coverage of lichens and bryophytes on theground in tropical and subtropical climates which also leadsto low respiration on a grid cell level and hence to low N2Oemissions

Figure 4 shows the simulated global spatial distribution ofthe ratio of respiration to NPP The assumption of a globallyconstant ratio of respiration to NPP is used by Lenhart et al(2015) to derive ecosystem-scale N2O emissions by lichensand bryophytes from their NPP Alternatively this ratio canbe derived from the independent LiBry estimates of NPP andrespiration The simulated ratio shows a latitudinal patternwith increasing values towards the tropics (Fig 4a) This re-sults from the influence of surface temperature on respira-tion in combination with high nighttime temperatures in thehumid tropics which cause high respiration rates during thenight Note that high respiration relative to NPP of tropicallichens and bryophytes does not necessarily mean high res-piration at the grid cell level since net productivity and cov-erage may be low Respiration by lichens and bryophytes inthe canopy shows a slightly weaker latitudinal gradient thanon the ground which can be explained by efficient evapo-rative cooling in the canopy (Fig 4b) In contrast lichens

N2O release

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

(a)

00 15 30 45 60 75 90 105 120 135 150

[mg mminus2 yrminus1]

N2O release in canopy

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

(b)

00 10 20 30 40 50 60 70 80 90 100

[mg mminus2 yrminus1]

N2O release on ground

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

(c)

00 05 10 15 20 25 30 35 40 45 50

[mg mminus2 yrminus1]

Figure 3 Global patterns of N2O release Nitrous oxide emissionsby lichens and bryophytes estimated by LiBry for (a) all locations ofgrowth (b) the canopy and (c) the ground Note the differing rangesof the colour bars Grey colour denotes regions where no simulatedspecies is able to survive such as ice shields and the driest regionsof deserts

and bryophytes on the ground usually grow within the sur-face boundary layer which reduces cooling by turbulent heattransfer leading to a strong influence of incoming radiationon surface temperature Since radiation input increases to-ward the Equator the ratio of respiration to NPP on theground in the tropics is markedly higher than at high lati-tudes (Fig 4c) The ratio of respiration to NPP varies fromless than 1 to around 2 while most values are around 1 Thismeans that gross primary productivity (GPP) is partitionedroughly equally into NPP and respiration which agrees wellwith the observational data from Lenhart et al (2015)

An overview of global total values of N2O emissions res-piration NPP and the ratio of respiration to NPP estimatedby LiBry is shown in Table 1 Table 2 shows N2O emis-sions by lichens and bryophytes for individual grid cells fromfive different ecosystem classes (see also Table 1) Variationin emissions between species within a grid cell is large itcan exceed 3 orders of magnitude The variation due to cli-

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1598 P Porada et al Nitrous oxide emissions by lichens and bryophytes

Table 1 Annual global total values of N2O emissions NPP respiration and the ratio of respiration to NPP estimated by LiBry and separatedinto lichens and bryophytes living in the canopy and on the ground The values in brackets in the first column show the uncertainty in N2Oemissions due to the conversion of released CO2 to N2O (90 confidence interval from Lenhart et al (2015)) Ecosystem classes shown arebased on the categories made by Olson et al (2001) which were aggregated by us in the same way as in Elbert et al (2012) ldquoGt Crdquo standsfor gigatons of carbon

N2O emissions NPP Respiration Respiration NPP(Tg N2O) yearminus1 (Gt C) yearminus1 (Gt C) yearminus1 [ ]

Canopy+ groundGlobal 027 (019ndash035) 43 45 110Tropical forest 011 (008ndash014) 15 18 133Extratropical forest 011 (008ndash014) 20 18 093Steppe amp savannah 003 (002ndash004) 04 04 121Desert 002 (001ndash003) 04 04 105Tundra 001 (0007ndash0013) 02 02 087

Canopy global 013 (009ndash017) 21 22 101Ground global 014 (010ndash018) 22 23 116

Table 2 Simulated nitrous oxide emissions by lichens and bryophytes in (mg N2O) mminus2 yearminus1 for individual grid cells of the LiBry modelThe values are averages over the last 50 years of a 600-year simulation with 3000 initial species Grid cells are selected from five differentecosystem classes In the two forest classes emissions are separated into canopy and ground In the other classes the model does notrepresent lichens and bryophytes in the canopy The range of N2O emissions based on all surviving artificial species in a grid cell is shownThe average value for all species in a grid cell is derived by an NPP-based weighting scheme (see Sect 2)

Ecosystem class Location Minimum Average Maximum

Tropical forest Central Amazon ground 031 059 088canopy 0081 33 82

Extratropical forest West Siberia ground 0023 17 48canopy 00040 21 62

Steppe amp savannah Central Sahel 00095 0088 032Desert Central Australia 0019 16 55Tundra North Alaska 0012 0095 017

matic conditions is smaller but it still amounts to almost 2orders of magnitude based on the grid cells with the high-est and lowest average emission rates Comparing Table 2 tothe global range of N2O emissions by lichens and bryophytes(Fig 3) shows that the five selected grid cells represent wellthe global variation in emissions due to climatic conditionsThus both functional diversity of the artificial species anddifferent climatic conditions are important for variation ofN2O emissions according to the LiBry simulation

4 Discussion

In this study we estimate nitrous oxide emissions by lichensand bryophytes with the global process-based non-vascularvegetation model LiBry Thereby we derive N2O emissionsfrom respiration fluxes which are together with photosyn-thesis and net primary productivity simulated by LiBry

We use an updated version of LiBry which contains signif-icant modifications with regard to the original version pub-lished in Porada et al (2013) Regarding NPP the new ver-

sion estimates 43 (Gt C) yearminus1 while the original versionof LiBry predicted a range of 034 to 33 (Gt C) yearminus1 Theincrease in predicted NPP is mainly attributed to a highersimulated productivity in the tropical forest canopy sincea new disturbance scheme allows for a higher surface cov-erage of lichens and bryophytes there An empirical globalestimate of NPP by lichens bryophytes free-living terres-trial cyanobacteria and algae (Elbert et al 2012) amountsto 39 (Gt C) yearminus1 Our new estimate is higher than thatby Elbert et al (2012) although LiBry does not considerfree-living cyanobacteria and algae This may be explainedby the small contribution of cyanobacteria and algae to theoverall global carbon uptake which can be compensatedfor by minor relative changes in productivity of lichensand bryophytes (Darrouzet-Nardi et al 2015 Sancho et al2016) It is not straightforward to determine which numberis closest to reality since both the process-based estimate byLiBry and the empirical one by Elbert et al (2012) are sub-ject to uncertainty In the study of Elbert et al (2012) forinstance it is assumed that productivity and active time areuniform within a biome Furthermore Elbert et al (2012) use

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1599

Ratio respiration to NPP

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

a)(

00 03 06 09 12 15 18 21 24

[ ]

Ratio respiration to NPP in canopy

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

b)(

00 03 06 09 12 15 18 21 24

[ ]

Ratio respiration to NPP on ground

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

c)(

00 03 06 09 12 15 18 21 24

[ ]

Figure 4 Global patterns of the ratio of respiration to NPP Ratioof respiration to NPP of lichens and bryophytes estimated by LiBryfor (a) all locations of growth (b) the canopy and (c) the ground

a globally uniform value of surface cover fraction to scaleup local field measurements of productivity to the globalscale However values of surface coverage by lichens andbryophytes compiled by Elbert et al (2012) vary greatly atthe small scale which makes upscaling to larger scales chal-lenging

While productivity estimated by LiBry is evaluated in thisstudy large-scale surface coverage of lichens and bryophytessimulated by LiBry has been evaluated for regions north of50 N in Porada et al (2016a) It was shown that LiBry pre-dicts realistic values of cover fraction Moreover values ofsurface cover predicted by LiBry for other regions of theworld (Porada et al 2016b) are in agreement with the es-timate of Elbert et al (2012) In spite of uncertainties regard-ing productivity and abundance of lichens and bryophytescomparing the empirical and process-based approaches givesconfidence in the order of magnitude of the LiBry simulationresults

As a 50-year steady-state average value we estimate to-tal N2O emissions by lichens and bryophytes of 027 (019ndash035) (Tg N2O) yearminus1 which is at the lower end of the rangeof 032 to 059 (Tg N2O) yearminus1 calculated by Lenhart et al(2015) The evaluation of LiBry regarding simulated NPPshows that our global patterns and total values of NPP arevery similar to the empirical estimate by Elbert et al (2012)Since Lenhart et al (2015) use this NPP estimate by Elbertet al (2012) to derive N2O emissions differences in NPPare most likely not the reason for our lower estimate of N2Oemissions compared to Lenhart et al (2015) Instead thismay be explained by differing methods to compute respira-tion while Lenhart et al (2015) assume a globally uniformratio of respiration to NPP of a value of 2 to estimate respira-tion LiBry simulates respiration independently as a species-specific function of temperature and water status This re-sults in a lower global average value of around 1 for the ra-tio of respiration to NPP predicted by LiBry Our estimatedratio of respiration to NPP agrees well with laboratory mea-surements but it is in general difficult to compare a globalecosystem-scale value to small-scale and short-term obser-vations

Our simulated global pattern of N2O emissions is slightlydifferent than that shown in Lenhart et al (2015) who esti-mate highest values in the boreal zone and only intermedi-ate values in the humid tropics This can be explained bytheir assumed constant ratio of respiration to NPP whichmakes their global pattern of N2O emissions identical to thatof NPP which is shown in Elbert et al (2012) In LiBryhowever the simulated ratio of respiration to NPP increasestowards higher surface temperatures in the tropics (Fig 4)Furthermore the ratio shows large spatial variation Evaluat-ing this simulated pattern is difficult since estimates whichare extrapolated to the large scale such as the NPP estimateby Elbert et al (2012) are not available for respiration bylichens and bryophytes However observed ratios of respira-tion to NPP of lichens and bryophytes vary considerably atthe species level as shown by for example Lenhart et al(2015) Using a constant ratio of respiration to NPP maytherefore introduce a bias in the estimated spatial distribu-tion of N2O emissions

Small-scale measurements of N2O emissions by lichensand bryophytes may show considerable variation Thesources of this variation may be physiological differencesbetween species variation of associated microbial communi-ties andor heterogeneity in climatic conditions We examinethe relative importance for respiration of differences betweenspecies compared to climatic differences with LiBry sincethe model simulates various physiological strategies and rep-resents variation in climatic conditions at the global scaleThereby we assume that the relationship between respira-tion and N2O emissions is relatively insensitive to climaticconditions and physiological differences between species assuggested by the experiments by Lenhart et al (2015) Ta-ble 2 shows that both differences between artificial species

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1600 P Porada et al Nitrous oxide emissions by lichens and bryophytes

and different climatic conditions are important for variationof N2O emissions Upscaling of N2O emission rates mea-sured in the field may therefore be subject to considerableuncertainty Modelling approaches in this direction shouldprobably account for both interspecific variation in processesassociated with N2O release by lichens and bryophytes aswell as variation in climatic conditions

Although our approach considers the most importantsources of variation in N2O emissions by lichens andbryophytes it is associated with uncertainties that should bediscussed further These uncertainties mainly result from ourmethod to estimate respiration and from assumptions con-cerning the empirical relationship between respiration andN2O emissions

Respiration and the ratio of respiration to NPP simulatedby LiBry are difficult to validate since the number of labo-ratory or field studies which measure not only NPP but alsoGPP and respiration is not very high Moreover long-termmeasurements of respiration would be required to determinethe ratio of respiration to NPP Otherwise assumptions aboutthe contribution of respiration in the dark to total respirationare necessary

To obtain N2O emissions from respiration our resultsrely on the laboratory incubation measurements and the cal-culated ratio of N2O emissions to respiration presented inLenhart et al (2015) Furthermore our approach considerseffects of variation in climatic conditions on N2O emissionsby lichens and bryophytes Hence it is necessary to dis-cuss the sensitivity of the relationship between respirationand N2O emissions to a range of climatic conditions Asshown in Lenhart et al (2015 Fig 3) the relationship be-tween respiration and N2O emissions seems to be insensi-tive to temperature changes for the tested species Likewisevariations in water content have no clear effect on the rela-tionship between N2O release and respiration (Lenhart et al2015 Fig S3) Although the sensitivities of N2O release totemperature and water content are similar to those of respira-tion across species the relationship between N2O release andrespiration shows interspecific variation However in spite ofa large number of around 40 sampled species the relation-ship shows a relatively narrow 90 confidence interval of113 to 207 ng N2O (mg CO2)minus1 (Lenhart et al 2015) Thissuggests that the mechanism of N2O release by lichens andbryophytes is similar between different species

To analyse the relation between the production of N2Oand respiratory CO2 in greater detail measurements of bothfluxes by means of online gas exchange measurements wouldbe needed which then could be linked to the observed wa-ter status of the organisms Since LiBry explicitly representsthe dynamic water saturation of lichens and bryophytes thiswould allow a more process-based prediction of the durationand magnitude of N2O emissions In this way the uncertaintyassociated with our approach would be reduced facilitatingan improved estimate of global N2O emissions by lichensand bryophytes

In order to assess model-based estimates of N2O emis-sions by lichens and bryophytes a relatively large number offield measurements are necessary Currently most N2O mea-surements independent of the substrate or organisms mea-sured generally suffer from major uncertainties additionallyto variation from functional diversity and differing climaticconditions first the majority of these studies have been con-ducted using the acetylene inhibition technique The idea ofthis method is to inhibit the last denitrification step so thatthe measured N2O amounts should reveal the sum of N2Oand N2 release during denitrification under natural condi-tions It has however been shown quite a while ago thatthis method leads to an underestimation of denitrificationunder oxic conditions (Bollmann and Conrad 1997) Sec-ondly the most widely used measuring technique has beenthe closed-chamber method which is inexpensive and easyto use This however has major shortcomings as environ-mental conditions are hard to control and only limited surfaceareas can be measured (Butterbach-Bahl et al 2013 Groff-man 2012) Furthermore the limited temporal resolution ofchamber measurements may affect estimated N2O emissions(Barton et al 2015) Thirdly depending on the environmen-tal conditions under which the experiment is performed ndashparticularly water temperature and nutrient conditions ndash theobtained N2O emission rates could differ widely Thus it isindispensable to report and consider the exact environmen-tal conditions under which the measurements were made andto restrict natural emission data to those assessed under typi-cally occurring natural conditions

Respiration by lichens and bryophytes is not the only pro-cess which can be used to estimate their N2O emissionsBarger et al (2013) report a relationship between nitrogenfixation and N2O release in biological soil crusts which in-clude lichens and bryophytes as well as soil bacteria andalgae For this approach however reliable nitrogen fixationdata are sparse It is also possible to estimate the demand fornitrogen by lichens and bryophytes with LiBry with an un-certainty range of around 1 order of magnitude (Porada et al2014) However it is not straightforward to derive realizednitrogen uptake or nitrogen fixation from this since LiBrydoes not yet include processes related to nitrogen uptake ormetabolization of nitrogen species Therefore for this studywe chose the relation between respiration and N2O release toquantify N2O emissions by lichens and bryophytes

Our simulated global N2O emissions by lichens andbryophytes of 027 (019ndash035) (Tg N2O) yearminus1 amount toaround 3 of global N2O emissions from natural sources onland (Ciais et al 2013) This value may sound low at firstglance but it equals about 50 of the atmospheric deposi-tion of N2O into the oceans or 25 of the deposition on land(Ciais et al 2013) Considering that N2O has a strong nega-tive effect on stratospheric ozone and a significant warmingpotential as a greenhouse gas even relatively small emissionsshould not be neglected in global budgets

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1601

The study by Zhuang et al (2012) estimates global pat-terns of N2O emission from soils and finds that the hu-mid tropics contribute most to global N2O emission due tohigh temperature and precipitation Our simulated pattern ofglobal N2O emissions by lichens and bryophytes also showsa hotspot in the humid tropics but the relative contributionof the boreal zone to the global flux seems to be higher thanin Zhuang et al (2012) This probably results from the highsimulated NPP in the boreal zone particularly on the groundwhich compensates for the lower respiration and thereforeN2O emission per productivity due to low temperatures Rel-ative contributions of lichens and bryophytes to N2O emis-sions are highest for ecosystems in desert regions and athigh latitudes which agrees with the results by Lenhart et al(2015)

5 Conclusions

We estimate large-scale spatial patterns and global values ofN2O emissions by lichens and bryophytes from a process-based model of their productivity and respiration Our resultssuggest a significant contribution of lichens and bryophytesto global N2O emissions albeit at the lower end of the rangeof a previous empirical estimate Since both approachesuse respiration to derive N2O emissions our lower estimatelikely results from a different method to predict respirationcompared to the empirical approach Hence while estimatesof productivity are relatively well constrained evaluatingmodels with regard to estimated respiration may improvepredictions of N2O emissions by lichens and bryophytesOne important finding derived from our simulation is that theratio of respiration to NPP by lichens and bryophytes showsspatial variation and a latitudinal gradient at the global scaleThis means that productivity and N2O emissions by the or-ganisms are not necessarily correlated and that tropical re-gions may show higher emissions than polar regions giventhe same NPP Furthermore we show that both physiologicalvariation among species and variation in climatic conditionsare relevant for variation in respiration and consequentlyN2O emissions Ecosystem-scale estimates of N2O emis-sions by lichens and bryophytes should therefore include suf-ficient ranges of species and climatic conditions to avoid bi-ased results Our results build on the empirical finding thatN2O emissions by lichens and bryophytes are linearly re-lated to their respiration This relationship is relatively in-sensitive to climatic conditions and shows no large varia-tion between species However the relationship is based onclosed-chamber measurements Therefore it would be use-ful to perform online gas exchange measurements of N2Oemissions and respiration to test the effect of climatic con-ditions on the relationship between N2O release and respi-ration Furthermore using alternative approaches to estimateN2O emissions by lichens and bryophytes may be helpful toconstrain our approach

Code and data availability The non-vascular vegetation model Li-Bry used here is combined with an interface for parallel computingwhich was developed at the Max Planck Institute for Biogeochem-istry Jena Germany LiBry without the interface is freely availableas long as the names of the copyright holders and a disclaimer aredistributed along with the code in source or binary form The codeis available from the corresponding author upon requestModel output data which are presented as maps in this study areavailable as netCDF files from the authors on request

Author contributions Philipp Porada Bettina Weber andAxel Kleidon designed the model simulations and Philipp Poradacarried them out Philipp Porada prepared the manuscript withcontributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This work has been supported by the PAGE21project grant agreement number 282700 funded by the EC SeventhFramework Programme theme FP7-ENV-2011 and the CARBOP-ERM project grant agreement number 03G0836B funded by theBMBF (German Ministry for Science and Education) The authorsthank the Max Planck Society (Nobel Laureate Fellowship forBettina Weber) for financial support The Max Planck Institute forBiogeochemistry provided computational resources

Edited by V BrovkinReviewed by two anonymous referees

References

Abed R Lam P de Beer D and Stief P High rates of den-itrification and nitrous oxide emission in arid biological soilcrusts from the Sultanate of Oman ISME J 7 1862ndash1875doi101038ismej201355 2013

Barger N Castle S and Dean G Denitrification from nitrogen-fixing biologically crusted soils in a cool desert environ-ment southeast Utah USA Ecological Processes 2 1ndash9doi1011862192-1709-2-16 2013

Barger N Weber B Garcia-Pichel F Zaady E and Belnap JPatterns and Controls on Nitrogen Cycling of Biological SoilCrusts in Biological Soil Crusts An Organizing Principle inDrylands edited by Weber B Buumldel B and Belnap J 257ndash285 Springer International Publishing Cham doi101007978-3-319-30214-0_14 2016

Barton L Wolf B Rowlings D Scheer C Kiese R Grace PStefanova K and Butterbach-Bahl K Sampling frequency af-fects estimates of annual nitrous oxide fluxes Scientific Reports5 15912 doi101038srep15912 2015

Bollmann A and Conrad R Acetylene blockage technique leadsto underestimation of denitrification rates in oxic soils due toscavenging of intermediate nitric oxide Soil Biol Biochem 291067ndash1077 doi101016S0038-0717(97)00007-2 1997

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1597

001

01

1

10

100

1000

Tundra Boreal f orestf loor

Desert Tropical f orestcanopy

Net carb

on u

pta

ke [(g

C)

mndash 2

yrndash 1

]

16

69

9

2

Figure 2 Comparison of LiBry estimates to field measurementsNPP estimated by LiBry compared to field measurements from fourbiomes defined after Olson et al (2001) The blue dots show theaverage simulated NPP for each biome and the blue vertical barsshow the range of NPP values between the different grid cells ina biome The magenta diamonds correspond to the median of NPPvalues measured in the field on the small scale the magenta verticalbars denote the range of the field measurements Left of the ma-genta diamonds the number of field measurements is shown that isconsidered for the respective biome Details can be found in Poradaet al (2013)

emissions in the humid tropics are around 3 times higher thanin the boreal and temperate zones Lichens and bryophytes onthe ground show highest values of nitrous oxide emissions inthe boreal zone with values around 3 (mg N2O) mminus2 yearminus1

(Fig 3c) Regarding the ground tropical and subtropical re-gions only partly show N2O emissions comparable to thoseof the boreal zone The reason for this is low simulatedproductivity and coverage of lichens and bryophytes on theground in tropical and subtropical climates which also leadsto low respiration on a grid cell level and hence to low N2Oemissions

Figure 4 shows the simulated global spatial distribution ofthe ratio of respiration to NPP The assumption of a globallyconstant ratio of respiration to NPP is used by Lenhart et al(2015) to derive ecosystem-scale N2O emissions by lichensand bryophytes from their NPP Alternatively this ratio canbe derived from the independent LiBry estimates of NPP andrespiration The simulated ratio shows a latitudinal patternwith increasing values towards the tropics (Fig 4a) This re-sults from the influence of surface temperature on respira-tion in combination with high nighttime temperatures in thehumid tropics which cause high respiration rates during thenight Note that high respiration relative to NPP of tropicallichens and bryophytes does not necessarily mean high res-piration at the grid cell level since net productivity and cov-erage may be low Respiration by lichens and bryophytes inthe canopy shows a slightly weaker latitudinal gradient thanon the ground which can be explained by efficient evapo-rative cooling in the canopy (Fig 4b) In contrast lichens

N2O release

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

(a)

00 15 30 45 60 75 90 105 120 135 150

[mg mminus2 yrminus1]

N2O release in canopy

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

(b)

00 10 20 30 40 50 60 70 80 90 100

[mg mminus2 yrminus1]

N2O release on ground

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

(c)

00 05 10 15 20 25 30 35 40 45 50

[mg mminus2 yrminus1]

Figure 3 Global patterns of N2O release Nitrous oxide emissionsby lichens and bryophytes estimated by LiBry for (a) all locations ofgrowth (b) the canopy and (c) the ground Note the differing rangesof the colour bars Grey colour denotes regions where no simulatedspecies is able to survive such as ice shields and the driest regionsof deserts

and bryophytes on the ground usually grow within the sur-face boundary layer which reduces cooling by turbulent heattransfer leading to a strong influence of incoming radiationon surface temperature Since radiation input increases to-ward the Equator the ratio of respiration to NPP on theground in the tropics is markedly higher than at high lati-tudes (Fig 4c) The ratio of respiration to NPP varies fromless than 1 to around 2 while most values are around 1 Thismeans that gross primary productivity (GPP) is partitionedroughly equally into NPP and respiration which agrees wellwith the observational data from Lenhart et al (2015)

An overview of global total values of N2O emissions res-piration NPP and the ratio of respiration to NPP estimatedby LiBry is shown in Table 1 Table 2 shows N2O emis-sions by lichens and bryophytes for individual grid cells fromfive different ecosystem classes (see also Table 1) Variationin emissions between species within a grid cell is large itcan exceed 3 orders of magnitude The variation due to cli-

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1598 P Porada et al Nitrous oxide emissions by lichens and bryophytes

Table 1 Annual global total values of N2O emissions NPP respiration and the ratio of respiration to NPP estimated by LiBry and separatedinto lichens and bryophytes living in the canopy and on the ground The values in brackets in the first column show the uncertainty in N2Oemissions due to the conversion of released CO2 to N2O (90 confidence interval from Lenhart et al (2015)) Ecosystem classes shown arebased on the categories made by Olson et al (2001) which were aggregated by us in the same way as in Elbert et al (2012) ldquoGt Crdquo standsfor gigatons of carbon

N2O emissions NPP Respiration Respiration NPP(Tg N2O) yearminus1 (Gt C) yearminus1 (Gt C) yearminus1 [ ]

Canopy+ groundGlobal 027 (019ndash035) 43 45 110Tropical forest 011 (008ndash014) 15 18 133Extratropical forest 011 (008ndash014) 20 18 093Steppe amp savannah 003 (002ndash004) 04 04 121Desert 002 (001ndash003) 04 04 105Tundra 001 (0007ndash0013) 02 02 087

Canopy global 013 (009ndash017) 21 22 101Ground global 014 (010ndash018) 22 23 116

Table 2 Simulated nitrous oxide emissions by lichens and bryophytes in (mg N2O) mminus2 yearminus1 for individual grid cells of the LiBry modelThe values are averages over the last 50 years of a 600-year simulation with 3000 initial species Grid cells are selected from five differentecosystem classes In the two forest classes emissions are separated into canopy and ground In the other classes the model does notrepresent lichens and bryophytes in the canopy The range of N2O emissions based on all surviving artificial species in a grid cell is shownThe average value for all species in a grid cell is derived by an NPP-based weighting scheme (see Sect 2)

Ecosystem class Location Minimum Average Maximum

Tropical forest Central Amazon ground 031 059 088canopy 0081 33 82

Extratropical forest West Siberia ground 0023 17 48canopy 00040 21 62

Steppe amp savannah Central Sahel 00095 0088 032Desert Central Australia 0019 16 55Tundra North Alaska 0012 0095 017

matic conditions is smaller but it still amounts to almost 2orders of magnitude based on the grid cells with the high-est and lowest average emission rates Comparing Table 2 tothe global range of N2O emissions by lichens and bryophytes(Fig 3) shows that the five selected grid cells represent wellthe global variation in emissions due to climatic conditionsThus both functional diversity of the artificial species anddifferent climatic conditions are important for variation ofN2O emissions according to the LiBry simulation

4 Discussion

In this study we estimate nitrous oxide emissions by lichensand bryophytes with the global process-based non-vascularvegetation model LiBry Thereby we derive N2O emissionsfrom respiration fluxes which are together with photosyn-thesis and net primary productivity simulated by LiBry

We use an updated version of LiBry which contains signif-icant modifications with regard to the original version pub-lished in Porada et al (2013) Regarding NPP the new ver-

sion estimates 43 (Gt C) yearminus1 while the original versionof LiBry predicted a range of 034 to 33 (Gt C) yearminus1 Theincrease in predicted NPP is mainly attributed to a highersimulated productivity in the tropical forest canopy sincea new disturbance scheme allows for a higher surface cov-erage of lichens and bryophytes there An empirical globalestimate of NPP by lichens bryophytes free-living terres-trial cyanobacteria and algae (Elbert et al 2012) amountsto 39 (Gt C) yearminus1 Our new estimate is higher than thatby Elbert et al (2012) although LiBry does not considerfree-living cyanobacteria and algae This may be explainedby the small contribution of cyanobacteria and algae to theoverall global carbon uptake which can be compensatedfor by minor relative changes in productivity of lichensand bryophytes (Darrouzet-Nardi et al 2015 Sancho et al2016) It is not straightforward to determine which numberis closest to reality since both the process-based estimate byLiBry and the empirical one by Elbert et al (2012) are sub-ject to uncertainty In the study of Elbert et al (2012) forinstance it is assumed that productivity and active time areuniform within a biome Furthermore Elbert et al (2012) use

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1599

Ratio respiration to NPP

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

a)(

00 03 06 09 12 15 18 21 24

[ ]

Ratio respiration to NPP in canopy

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

b)(

00 03 06 09 12 15 18 21 24

[ ]

Ratio respiration to NPP on ground

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

c)(

00 03 06 09 12 15 18 21 24

[ ]

Figure 4 Global patterns of the ratio of respiration to NPP Ratioof respiration to NPP of lichens and bryophytes estimated by LiBryfor (a) all locations of growth (b) the canopy and (c) the ground

a globally uniform value of surface cover fraction to scaleup local field measurements of productivity to the globalscale However values of surface coverage by lichens andbryophytes compiled by Elbert et al (2012) vary greatly atthe small scale which makes upscaling to larger scales chal-lenging

While productivity estimated by LiBry is evaluated in thisstudy large-scale surface coverage of lichens and bryophytessimulated by LiBry has been evaluated for regions north of50 N in Porada et al (2016a) It was shown that LiBry pre-dicts realistic values of cover fraction Moreover values ofsurface cover predicted by LiBry for other regions of theworld (Porada et al 2016b) are in agreement with the es-timate of Elbert et al (2012) In spite of uncertainties regard-ing productivity and abundance of lichens and bryophytescomparing the empirical and process-based approaches givesconfidence in the order of magnitude of the LiBry simulationresults

As a 50-year steady-state average value we estimate to-tal N2O emissions by lichens and bryophytes of 027 (019ndash035) (Tg N2O) yearminus1 which is at the lower end of the rangeof 032 to 059 (Tg N2O) yearminus1 calculated by Lenhart et al(2015) The evaluation of LiBry regarding simulated NPPshows that our global patterns and total values of NPP arevery similar to the empirical estimate by Elbert et al (2012)Since Lenhart et al (2015) use this NPP estimate by Elbertet al (2012) to derive N2O emissions differences in NPPare most likely not the reason for our lower estimate of N2Oemissions compared to Lenhart et al (2015) Instead thismay be explained by differing methods to compute respira-tion while Lenhart et al (2015) assume a globally uniformratio of respiration to NPP of a value of 2 to estimate respira-tion LiBry simulates respiration independently as a species-specific function of temperature and water status This re-sults in a lower global average value of around 1 for the ra-tio of respiration to NPP predicted by LiBry Our estimatedratio of respiration to NPP agrees well with laboratory mea-surements but it is in general difficult to compare a globalecosystem-scale value to small-scale and short-term obser-vations

Our simulated global pattern of N2O emissions is slightlydifferent than that shown in Lenhart et al (2015) who esti-mate highest values in the boreal zone and only intermedi-ate values in the humid tropics This can be explained bytheir assumed constant ratio of respiration to NPP whichmakes their global pattern of N2O emissions identical to thatof NPP which is shown in Elbert et al (2012) In LiBryhowever the simulated ratio of respiration to NPP increasestowards higher surface temperatures in the tropics (Fig 4)Furthermore the ratio shows large spatial variation Evaluat-ing this simulated pattern is difficult since estimates whichare extrapolated to the large scale such as the NPP estimateby Elbert et al (2012) are not available for respiration bylichens and bryophytes However observed ratios of respira-tion to NPP of lichens and bryophytes vary considerably atthe species level as shown by for example Lenhart et al(2015) Using a constant ratio of respiration to NPP maytherefore introduce a bias in the estimated spatial distribu-tion of N2O emissions

Small-scale measurements of N2O emissions by lichensand bryophytes may show considerable variation Thesources of this variation may be physiological differencesbetween species variation of associated microbial communi-ties andor heterogeneity in climatic conditions We examinethe relative importance for respiration of differences betweenspecies compared to climatic differences with LiBry sincethe model simulates various physiological strategies and rep-resents variation in climatic conditions at the global scaleThereby we assume that the relationship between respira-tion and N2O emissions is relatively insensitive to climaticconditions and physiological differences between species assuggested by the experiments by Lenhart et al (2015) Ta-ble 2 shows that both differences between artificial species

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1600 P Porada et al Nitrous oxide emissions by lichens and bryophytes

and different climatic conditions are important for variationof N2O emissions Upscaling of N2O emission rates mea-sured in the field may therefore be subject to considerableuncertainty Modelling approaches in this direction shouldprobably account for both interspecific variation in processesassociated with N2O release by lichens and bryophytes aswell as variation in climatic conditions

Although our approach considers the most importantsources of variation in N2O emissions by lichens andbryophytes it is associated with uncertainties that should bediscussed further These uncertainties mainly result from ourmethod to estimate respiration and from assumptions con-cerning the empirical relationship between respiration andN2O emissions

Respiration and the ratio of respiration to NPP simulatedby LiBry are difficult to validate since the number of labo-ratory or field studies which measure not only NPP but alsoGPP and respiration is not very high Moreover long-termmeasurements of respiration would be required to determinethe ratio of respiration to NPP Otherwise assumptions aboutthe contribution of respiration in the dark to total respirationare necessary

To obtain N2O emissions from respiration our resultsrely on the laboratory incubation measurements and the cal-culated ratio of N2O emissions to respiration presented inLenhart et al (2015) Furthermore our approach considerseffects of variation in climatic conditions on N2O emissionsby lichens and bryophytes Hence it is necessary to dis-cuss the sensitivity of the relationship between respirationand N2O emissions to a range of climatic conditions Asshown in Lenhart et al (2015 Fig 3) the relationship be-tween respiration and N2O emissions seems to be insensi-tive to temperature changes for the tested species Likewisevariations in water content have no clear effect on the rela-tionship between N2O release and respiration (Lenhart et al2015 Fig S3) Although the sensitivities of N2O release totemperature and water content are similar to those of respira-tion across species the relationship between N2O release andrespiration shows interspecific variation However in spite ofa large number of around 40 sampled species the relation-ship shows a relatively narrow 90 confidence interval of113 to 207 ng N2O (mg CO2)minus1 (Lenhart et al 2015) Thissuggests that the mechanism of N2O release by lichens andbryophytes is similar between different species

To analyse the relation between the production of N2Oand respiratory CO2 in greater detail measurements of bothfluxes by means of online gas exchange measurements wouldbe needed which then could be linked to the observed wa-ter status of the organisms Since LiBry explicitly representsthe dynamic water saturation of lichens and bryophytes thiswould allow a more process-based prediction of the durationand magnitude of N2O emissions In this way the uncertaintyassociated with our approach would be reduced facilitatingan improved estimate of global N2O emissions by lichensand bryophytes

In order to assess model-based estimates of N2O emis-sions by lichens and bryophytes a relatively large number offield measurements are necessary Currently most N2O mea-surements independent of the substrate or organisms mea-sured generally suffer from major uncertainties additionallyto variation from functional diversity and differing climaticconditions first the majority of these studies have been con-ducted using the acetylene inhibition technique The idea ofthis method is to inhibit the last denitrification step so thatthe measured N2O amounts should reveal the sum of N2Oand N2 release during denitrification under natural condi-tions It has however been shown quite a while ago thatthis method leads to an underestimation of denitrificationunder oxic conditions (Bollmann and Conrad 1997) Sec-ondly the most widely used measuring technique has beenthe closed-chamber method which is inexpensive and easyto use This however has major shortcomings as environ-mental conditions are hard to control and only limited surfaceareas can be measured (Butterbach-Bahl et al 2013 Groff-man 2012) Furthermore the limited temporal resolution ofchamber measurements may affect estimated N2O emissions(Barton et al 2015) Thirdly depending on the environmen-tal conditions under which the experiment is performed ndashparticularly water temperature and nutrient conditions ndash theobtained N2O emission rates could differ widely Thus it isindispensable to report and consider the exact environmen-tal conditions under which the measurements were made andto restrict natural emission data to those assessed under typi-cally occurring natural conditions

Respiration by lichens and bryophytes is not the only pro-cess which can be used to estimate their N2O emissionsBarger et al (2013) report a relationship between nitrogenfixation and N2O release in biological soil crusts which in-clude lichens and bryophytes as well as soil bacteria andalgae For this approach however reliable nitrogen fixationdata are sparse It is also possible to estimate the demand fornitrogen by lichens and bryophytes with LiBry with an un-certainty range of around 1 order of magnitude (Porada et al2014) However it is not straightforward to derive realizednitrogen uptake or nitrogen fixation from this since LiBrydoes not yet include processes related to nitrogen uptake ormetabolization of nitrogen species Therefore for this studywe chose the relation between respiration and N2O release toquantify N2O emissions by lichens and bryophytes

Our simulated global N2O emissions by lichens andbryophytes of 027 (019ndash035) (Tg N2O) yearminus1 amount toaround 3 of global N2O emissions from natural sources onland (Ciais et al 2013) This value may sound low at firstglance but it equals about 50 of the atmospheric deposi-tion of N2O into the oceans or 25 of the deposition on land(Ciais et al 2013) Considering that N2O has a strong nega-tive effect on stratospheric ozone and a significant warmingpotential as a greenhouse gas even relatively small emissionsshould not be neglected in global budgets

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1601

The study by Zhuang et al (2012) estimates global pat-terns of N2O emission from soils and finds that the hu-mid tropics contribute most to global N2O emission due tohigh temperature and precipitation Our simulated pattern ofglobal N2O emissions by lichens and bryophytes also showsa hotspot in the humid tropics but the relative contributionof the boreal zone to the global flux seems to be higher thanin Zhuang et al (2012) This probably results from the highsimulated NPP in the boreal zone particularly on the groundwhich compensates for the lower respiration and thereforeN2O emission per productivity due to low temperatures Rel-ative contributions of lichens and bryophytes to N2O emis-sions are highest for ecosystems in desert regions and athigh latitudes which agrees with the results by Lenhart et al(2015)

5 Conclusions

We estimate large-scale spatial patterns and global values ofN2O emissions by lichens and bryophytes from a process-based model of their productivity and respiration Our resultssuggest a significant contribution of lichens and bryophytesto global N2O emissions albeit at the lower end of the rangeof a previous empirical estimate Since both approachesuse respiration to derive N2O emissions our lower estimatelikely results from a different method to predict respirationcompared to the empirical approach Hence while estimatesof productivity are relatively well constrained evaluatingmodels with regard to estimated respiration may improvepredictions of N2O emissions by lichens and bryophytesOne important finding derived from our simulation is that theratio of respiration to NPP by lichens and bryophytes showsspatial variation and a latitudinal gradient at the global scaleThis means that productivity and N2O emissions by the or-ganisms are not necessarily correlated and that tropical re-gions may show higher emissions than polar regions giventhe same NPP Furthermore we show that both physiologicalvariation among species and variation in climatic conditionsare relevant for variation in respiration and consequentlyN2O emissions Ecosystem-scale estimates of N2O emis-sions by lichens and bryophytes should therefore include suf-ficient ranges of species and climatic conditions to avoid bi-ased results Our results build on the empirical finding thatN2O emissions by lichens and bryophytes are linearly re-lated to their respiration This relationship is relatively in-sensitive to climatic conditions and shows no large varia-tion between species However the relationship is based onclosed-chamber measurements Therefore it would be use-ful to perform online gas exchange measurements of N2Oemissions and respiration to test the effect of climatic con-ditions on the relationship between N2O release and respi-ration Furthermore using alternative approaches to estimateN2O emissions by lichens and bryophytes may be helpful toconstrain our approach

Code and data availability The non-vascular vegetation model Li-Bry used here is combined with an interface for parallel computingwhich was developed at the Max Planck Institute for Biogeochem-istry Jena Germany LiBry without the interface is freely availableas long as the names of the copyright holders and a disclaimer aredistributed along with the code in source or binary form The codeis available from the corresponding author upon requestModel output data which are presented as maps in this study areavailable as netCDF files from the authors on request

Author contributions Philipp Porada Bettina Weber andAxel Kleidon designed the model simulations and Philipp Poradacarried them out Philipp Porada prepared the manuscript withcontributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This work has been supported by the PAGE21project grant agreement number 282700 funded by the EC SeventhFramework Programme theme FP7-ENV-2011 and the CARBOP-ERM project grant agreement number 03G0836B funded by theBMBF (German Ministry for Science and Education) The authorsthank the Max Planck Society (Nobel Laureate Fellowship forBettina Weber) for financial support The Max Planck Institute forBiogeochemistry provided computational resources

Edited by V BrovkinReviewed by two anonymous referees

References

Abed R Lam P de Beer D and Stief P High rates of den-itrification and nitrous oxide emission in arid biological soilcrusts from the Sultanate of Oman ISME J 7 1862ndash1875doi101038ismej201355 2013

Barger N Castle S and Dean G Denitrification from nitrogen-fixing biologically crusted soils in a cool desert environ-ment southeast Utah USA Ecological Processes 2 1ndash9doi1011862192-1709-2-16 2013

Barger N Weber B Garcia-Pichel F Zaady E and Belnap JPatterns and Controls on Nitrogen Cycling of Biological SoilCrusts in Biological Soil Crusts An Organizing Principle inDrylands edited by Weber B Buumldel B and Belnap J 257ndash285 Springer International Publishing Cham doi101007978-3-319-30214-0_14 2016

Barton L Wolf B Rowlings D Scheer C Kiese R Grace PStefanova K and Butterbach-Bahl K Sampling frequency af-fects estimates of annual nitrous oxide fluxes Scientific Reports5 15912 doi101038srep15912 2015

Bollmann A and Conrad R Acetylene blockage technique leadsto underestimation of denitrification rates in oxic soils due toscavenging of intermediate nitric oxide Soil Biol Biochem 291067ndash1077 doi101016S0038-0717(97)00007-2 1997

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1598 P Porada et al Nitrous oxide emissions by lichens and bryophytes

Table 1 Annual global total values of N2O emissions NPP respiration and the ratio of respiration to NPP estimated by LiBry and separatedinto lichens and bryophytes living in the canopy and on the ground The values in brackets in the first column show the uncertainty in N2Oemissions due to the conversion of released CO2 to N2O (90 confidence interval from Lenhart et al (2015)) Ecosystem classes shown arebased on the categories made by Olson et al (2001) which were aggregated by us in the same way as in Elbert et al (2012) ldquoGt Crdquo standsfor gigatons of carbon

N2O emissions NPP Respiration Respiration NPP(Tg N2O) yearminus1 (Gt C) yearminus1 (Gt C) yearminus1 [ ]

Canopy+ groundGlobal 027 (019ndash035) 43 45 110Tropical forest 011 (008ndash014) 15 18 133Extratropical forest 011 (008ndash014) 20 18 093Steppe amp savannah 003 (002ndash004) 04 04 121Desert 002 (001ndash003) 04 04 105Tundra 001 (0007ndash0013) 02 02 087

Canopy global 013 (009ndash017) 21 22 101Ground global 014 (010ndash018) 22 23 116

Table 2 Simulated nitrous oxide emissions by lichens and bryophytes in (mg N2O) mminus2 yearminus1 for individual grid cells of the LiBry modelThe values are averages over the last 50 years of a 600-year simulation with 3000 initial species Grid cells are selected from five differentecosystem classes In the two forest classes emissions are separated into canopy and ground In the other classes the model does notrepresent lichens and bryophytes in the canopy The range of N2O emissions based on all surviving artificial species in a grid cell is shownThe average value for all species in a grid cell is derived by an NPP-based weighting scheme (see Sect 2)

Ecosystem class Location Minimum Average Maximum

Tropical forest Central Amazon ground 031 059 088canopy 0081 33 82

Extratropical forest West Siberia ground 0023 17 48canopy 00040 21 62

Steppe amp savannah Central Sahel 00095 0088 032Desert Central Australia 0019 16 55Tundra North Alaska 0012 0095 017

matic conditions is smaller but it still amounts to almost 2orders of magnitude based on the grid cells with the high-est and lowest average emission rates Comparing Table 2 tothe global range of N2O emissions by lichens and bryophytes(Fig 3) shows that the five selected grid cells represent wellthe global variation in emissions due to climatic conditionsThus both functional diversity of the artificial species anddifferent climatic conditions are important for variation ofN2O emissions according to the LiBry simulation

4 Discussion

In this study we estimate nitrous oxide emissions by lichensand bryophytes with the global process-based non-vascularvegetation model LiBry Thereby we derive N2O emissionsfrom respiration fluxes which are together with photosyn-thesis and net primary productivity simulated by LiBry

We use an updated version of LiBry which contains signif-icant modifications with regard to the original version pub-lished in Porada et al (2013) Regarding NPP the new ver-

sion estimates 43 (Gt C) yearminus1 while the original versionof LiBry predicted a range of 034 to 33 (Gt C) yearminus1 Theincrease in predicted NPP is mainly attributed to a highersimulated productivity in the tropical forest canopy sincea new disturbance scheme allows for a higher surface cov-erage of lichens and bryophytes there An empirical globalestimate of NPP by lichens bryophytes free-living terres-trial cyanobacteria and algae (Elbert et al 2012) amountsto 39 (Gt C) yearminus1 Our new estimate is higher than thatby Elbert et al (2012) although LiBry does not considerfree-living cyanobacteria and algae This may be explainedby the small contribution of cyanobacteria and algae to theoverall global carbon uptake which can be compensatedfor by minor relative changes in productivity of lichensand bryophytes (Darrouzet-Nardi et al 2015 Sancho et al2016) It is not straightforward to determine which numberis closest to reality since both the process-based estimate byLiBry and the empirical one by Elbert et al (2012) are sub-ject to uncertainty In the study of Elbert et al (2012) forinstance it is assumed that productivity and active time areuniform within a biome Furthermore Elbert et al (2012) use

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1599

Ratio respiration to NPP

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

a)(

00 03 06 09 12 15 18 21 24

[ ]

Ratio respiration to NPP in canopy

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

b)(

00 03 06 09 12 15 18 21 24

[ ]

Ratio respiration to NPP on ground

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

c)(

00 03 06 09 12 15 18 21 24

[ ]

Figure 4 Global patterns of the ratio of respiration to NPP Ratioof respiration to NPP of lichens and bryophytes estimated by LiBryfor (a) all locations of growth (b) the canopy and (c) the ground

a globally uniform value of surface cover fraction to scaleup local field measurements of productivity to the globalscale However values of surface coverage by lichens andbryophytes compiled by Elbert et al (2012) vary greatly atthe small scale which makes upscaling to larger scales chal-lenging

While productivity estimated by LiBry is evaluated in thisstudy large-scale surface coverage of lichens and bryophytessimulated by LiBry has been evaluated for regions north of50 N in Porada et al (2016a) It was shown that LiBry pre-dicts realistic values of cover fraction Moreover values ofsurface cover predicted by LiBry for other regions of theworld (Porada et al 2016b) are in agreement with the es-timate of Elbert et al (2012) In spite of uncertainties regard-ing productivity and abundance of lichens and bryophytescomparing the empirical and process-based approaches givesconfidence in the order of magnitude of the LiBry simulationresults

As a 50-year steady-state average value we estimate to-tal N2O emissions by lichens and bryophytes of 027 (019ndash035) (Tg N2O) yearminus1 which is at the lower end of the rangeof 032 to 059 (Tg N2O) yearminus1 calculated by Lenhart et al(2015) The evaluation of LiBry regarding simulated NPPshows that our global patterns and total values of NPP arevery similar to the empirical estimate by Elbert et al (2012)Since Lenhart et al (2015) use this NPP estimate by Elbertet al (2012) to derive N2O emissions differences in NPPare most likely not the reason for our lower estimate of N2Oemissions compared to Lenhart et al (2015) Instead thismay be explained by differing methods to compute respira-tion while Lenhart et al (2015) assume a globally uniformratio of respiration to NPP of a value of 2 to estimate respira-tion LiBry simulates respiration independently as a species-specific function of temperature and water status This re-sults in a lower global average value of around 1 for the ra-tio of respiration to NPP predicted by LiBry Our estimatedratio of respiration to NPP agrees well with laboratory mea-surements but it is in general difficult to compare a globalecosystem-scale value to small-scale and short-term obser-vations

Our simulated global pattern of N2O emissions is slightlydifferent than that shown in Lenhart et al (2015) who esti-mate highest values in the boreal zone and only intermedi-ate values in the humid tropics This can be explained bytheir assumed constant ratio of respiration to NPP whichmakes their global pattern of N2O emissions identical to thatof NPP which is shown in Elbert et al (2012) In LiBryhowever the simulated ratio of respiration to NPP increasestowards higher surface temperatures in the tropics (Fig 4)Furthermore the ratio shows large spatial variation Evaluat-ing this simulated pattern is difficult since estimates whichare extrapolated to the large scale such as the NPP estimateby Elbert et al (2012) are not available for respiration bylichens and bryophytes However observed ratios of respira-tion to NPP of lichens and bryophytes vary considerably atthe species level as shown by for example Lenhart et al(2015) Using a constant ratio of respiration to NPP maytherefore introduce a bias in the estimated spatial distribu-tion of N2O emissions

Small-scale measurements of N2O emissions by lichensand bryophytes may show considerable variation Thesources of this variation may be physiological differencesbetween species variation of associated microbial communi-ties andor heterogeneity in climatic conditions We examinethe relative importance for respiration of differences betweenspecies compared to climatic differences with LiBry sincethe model simulates various physiological strategies and rep-resents variation in climatic conditions at the global scaleThereby we assume that the relationship between respira-tion and N2O emissions is relatively insensitive to climaticconditions and physiological differences between species assuggested by the experiments by Lenhart et al (2015) Ta-ble 2 shows that both differences between artificial species

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1600 P Porada et al Nitrous oxide emissions by lichens and bryophytes

and different climatic conditions are important for variationof N2O emissions Upscaling of N2O emission rates mea-sured in the field may therefore be subject to considerableuncertainty Modelling approaches in this direction shouldprobably account for both interspecific variation in processesassociated with N2O release by lichens and bryophytes aswell as variation in climatic conditions

Although our approach considers the most importantsources of variation in N2O emissions by lichens andbryophytes it is associated with uncertainties that should bediscussed further These uncertainties mainly result from ourmethod to estimate respiration and from assumptions con-cerning the empirical relationship between respiration andN2O emissions

Respiration and the ratio of respiration to NPP simulatedby LiBry are difficult to validate since the number of labo-ratory or field studies which measure not only NPP but alsoGPP and respiration is not very high Moreover long-termmeasurements of respiration would be required to determinethe ratio of respiration to NPP Otherwise assumptions aboutthe contribution of respiration in the dark to total respirationare necessary

To obtain N2O emissions from respiration our resultsrely on the laboratory incubation measurements and the cal-culated ratio of N2O emissions to respiration presented inLenhart et al (2015) Furthermore our approach considerseffects of variation in climatic conditions on N2O emissionsby lichens and bryophytes Hence it is necessary to dis-cuss the sensitivity of the relationship between respirationand N2O emissions to a range of climatic conditions Asshown in Lenhart et al (2015 Fig 3) the relationship be-tween respiration and N2O emissions seems to be insensi-tive to temperature changes for the tested species Likewisevariations in water content have no clear effect on the rela-tionship between N2O release and respiration (Lenhart et al2015 Fig S3) Although the sensitivities of N2O release totemperature and water content are similar to those of respira-tion across species the relationship between N2O release andrespiration shows interspecific variation However in spite ofa large number of around 40 sampled species the relation-ship shows a relatively narrow 90 confidence interval of113 to 207 ng N2O (mg CO2)minus1 (Lenhart et al 2015) Thissuggests that the mechanism of N2O release by lichens andbryophytes is similar between different species

To analyse the relation between the production of N2Oand respiratory CO2 in greater detail measurements of bothfluxes by means of online gas exchange measurements wouldbe needed which then could be linked to the observed wa-ter status of the organisms Since LiBry explicitly representsthe dynamic water saturation of lichens and bryophytes thiswould allow a more process-based prediction of the durationand magnitude of N2O emissions In this way the uncertaintyassociated with our approach would be reduced facilitatingan improved estimate of global N2O emissions by lichensand bryophytes

In order to assess model-based estimates of N2O emis-sions by lichens and bryophytes a relatively large number offield measurements are necessary Currently most N2O mea-surements independent of the substrate or organisms mea-sured generally suffer from major uncertainties additionallyto variation from functional diversity and differing climaticconditions first the majority of these studies have been con-ducted using the acetylene inhibition technique The idea ofthis method is to inhibit the last denitrification step so thatthe measured N2O amounts should reveal the sum of N2Oand N2 release during denitrification under natural condi-tions It has however been shown quite a while ago thatthis method leads to an underestimation of denitrificationunder oxic conditions (Bollmann and Conrad 1997) Sec-ondly the most widely used measuring technique has beenthe closed-chamber method which is inexpensive and easyto use This however has major shortcomings as environ-mental conditions are hard to control and only limited surfaceareas can be measured (Butterbach-Bahl et al 2013 Groff-man 2012) Furthermore the limited temporal resolution ofchamber measurements may affect estimated N2O emissions(Barton et al 2015) Thirdly depending on the environmen-tal conditions under which the experiment is performed ndashparticularly water temperature and nutrient conditions ndash theobtained N2O emission rates could differ widely Thus it isindispensable to report and consider the exact environmen-tal conditions under which the measurements were made andto restrict natural emission data to those assessed under typi-cally occurring natural conditions

Respiration by lichens and bryophytes is not the only pro-cess which can be used to estimate their N2O emissionsBarger et al (2013) report a relationship between nitrogenfixation and N2O release in biological soil crusts which in-clude lichens and bryophytes as well as soil bacteria andalgae For this approach however reliable nitrogen fixationdata are sparse It is also possible to estimate the demand fornitrogen by lichens and bryophytes with LiBry with an un-certainty range of around 1 order of magnitude (Porada et al2014) However it is not straightforward to derive realizednitrogen uptake or nitrogen fixation from this since LiBrydoes not yet include processes related to nitrogen uptake ormetabolization of nitrogen species Therefore for this studywe chose the relation between respiration and N2O release toquantify N2O emissions by lichens and bryophytes

Our simulated global N2O emissions by lichens andbryophytes of 027 (019ndash035) (Tg N2O) yearminus1 amount toaround 3 of global N2O emissions from natural sources onland (Ciais et al 2013) This value may sound low at firstglance but it equals about 50 of the atmospheric deposi-tion of N2O into the oceans or 25 of the deposition on land(Ciais et al 2013) Considering that N2O has a strong nega-tive effect on stratospheric ozone and a significant warmingpotential as a greenhouse gas even relatively small emissionsshould not be neglected in global budgets

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1601

The study by Zhuang et al (2012) estimates global pat-terns of N2O emission from soils and finds that the hu-mid tropics contribute most to global N2O emission due tohigh temperature and precipitation Our simulated pattern ofglobal N2O emissions by lichens and bryophytes also showsa hotspot in the humid tropics but the relative contributionof the boreal zone to the global flux seems to be higher thanin Zhuang et al (2012) This probably results from the highsimulated NPP in the boreal zone particularly on the groundwhich compensates for the lower respiration and thereforeN2O emission per productivity due to low temperatures Rel-ative contributions of lichens and bryophytes to N2O emis-sions are highest for ecosystems in desert regions and athigh latitudes which agrees with the results by Lenhart et al(2015)

5 Conclusions

We estimate large-scale spatial patterns and global values ofN2O emissions by lichens and bryophytes from a process-based model of their productivity and respiration Our resultssuggest a significant contribution of lichens and bryophytesto global N2O emissions albeit at the lower end of the rangeof a previous empirical estimate Since both approachesuse respiration to derive N2O emissions our lower estimatelikely results from a different method to predict respirationcompared to the empirical approach Hence while estimatesof productivity are relatively well constrained evaluatingmodels with regard to estimated respiration may improvepredictions of N2O emissions by lichens and bryophytesOne important finding derived from our simulation is that theratio of respiration to NPP by lichens and bryophytes showsspatial variation and a latitudinal gradient at the global scaleThis means that productivity and N2O emissions by the or-ganisms are not necessarily correlated and that tropical re-gions may show higher emissions than polar regions giventhe same NPP Furthermore we show that both physiologicalvariation among species and variation in climatic conditionsare relevant for variation in respiration and consequentlyN2O emissions Ecosystem-scale estimates of N2O emis-sions by lichens and bryophytes should therefore include suf-ficient ranges of species and climatic conditions to avoid bi-ased results Our results build on the empirical finding thatN2O emissions by lichens and bryophytes are linearly re-lated to their respiration This relationship is relatively in-sensitive to climatic conditions and shows no large varia-tion between species However the relationship is based onclosed-chamber measurements Therefore it would be use-ful to perform online gas exchange measurements of N2Oemissions and respiration to test the effect of climatic con-ditions on the relationship between N2O release and respi-ration Furthermore using alternative approaches to estimateN2O emissions by lichens and bryophytes may be helpful toconstrain our approach

Code and data availability The non-vascular vegetation model Li-Bry used here is combined with an interface for parallel computingwhich was developed at the Max Planck Institute for Biogeochem-istry Jena Germany LiBry without the interface is freely availableas long as the names of the copyright holders and a disclaimer aredistributed along with the code in source or binary form The codeis available from the corresponding author upon requestModel output data which are presented as maps in this study areavailable as netCDF files from the authors on request

Author contributions Philipp Porada Bettina Weber andAxel Kleidon designed the model simulations and Philipp Poradacarried them out Philipp Porada prepared the manuscript withcontributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This work has been supported by the PAGE21project grant agreement number 282700 funded by the EC SeventhFramework Programme theme FP7-ENV-2011 and the CARBOP-ERM project grant agreement number 03G0836B funded by theBMBF (German Ministry for Science and Education) The authorsthank the Max Planck Society (Nobel Laureate Fellowship forBettina Weber) for financial support The Max Planck Institute forBiogeochemistry provided computational resources

Edited by V BrovkinReviewed by two anonymous referees

References

Abed R Lam P de Beer D and Stief P High rates of den-itrification and nitrous oxide emission in arid biological soilcrusts from the Sultanate of Oman ISME J 7 1862ndash1875doi101038ismej201355 2013

Barger N Castle S and Dean G Denitrification from nitrogen-fixing biologically crusted soils in a cool desert environ-ment southeast Utah USA Ecological Processes 2 1ndash9doi1011862192-1709-2-16 2013

Barger N Weber B Garcia-Pichel F Zaady E and Belnap JPatterns and Controls on Nitrogen Cycling of Biological SoilCrusts in Biological Soil Crusts An Organizing Principle inDrylands edited by Weber B Buumldel B and Belnap J 257ndash285 Springer International Publishing Cham doi101007978-3-319-30214-0_14 2016

Barton L Wolf B Rowlings D Scheer C Kiese R Grace PStefanova K and Butterbach-Bahl K Sampling frequency af-fects estimates of annual nitrous oxide fluxes Scientific Reports5 15912 doi101038srep15912 2015

Bollmann A and Conrad R Acetylene blockage technique leadsto underestimation of denitrification rates in oxic soils due toscavenging of intermediate nitric oxide Soil Biol Biochem 291067ndash1077 doi101016S0038-0717(97)00007-2 1997

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1599

Ratio respiration to NPP

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

a)(

00 03 06 09 12 15 18 21 24

[ ]

Ratio respiration to NPP in canopy

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

b)(

00 03 06 09 12 15 18 21 24

[ ]

Ratio respiration to NPP on ground

minus180˚ minus90˚ 0˚ 90˚ 180˚minus60˚

minus30˚

0˚

30˚

60˚

90˚

c)(

00 03 06 09 12 15 18 21 24

[ ]

Figure 4 Global patterns of the ratio of respiration to NPP Ratioof respiration to NPP of lichens and bryophytes estimated by LiBryfor (a) all locations of growth (b) the canopy and (c) the ground

a globally uniform value of surface cover fraction to scaleup local field measurements of productivity to the globalscale However values of surface coverage by lichens andbryophytes compiled by Elbert et al (2012) vary greatly atthe small scale which makes upscaling to larger scales chal-lenging

While productivity estimated by LiBry is evaluated in thisstudy large-scale surface coverage of lichens and bryophytessimulated by LiBry has been evaluated for regions north of50 N in Porada et al (2016a) It was shown that LiBry pre-dicts realistic values of cover fraction Moreover values ofsurface cover predicted by LiBry for other regions of theworld (Porada et al 2016b) are in agreement with the es-timate of Elbert et al (2012) In spite of uncertainties regard-ing productivity and abundance of lichens and bryophytescomparing the empirical and process-based approaches givesconfidence in the order of magnitude of the LiBry simulationresults

As a 50-year steady-state average value we estimate to-tal N2O emissions by lichens and bryophytes of 027 (019ndash035) (Tg N2O) yearminus1 which is at the lower end of the rangeof 032 to 059 (Tg N2O) yearminus1 calculated by Lenhart et al(2015) The evaluation of LiBry regarding simulated NPPshows that our global patterns and total values of NPP arevery similar to the empirical estimate by Elbert et al (2012)Since Lenhart et al (2015) use this NPP estimate by Elbertet al (2012) to derive N2O emissions differences in NPPare most likely not the reason for our lower estimate of N2Oemissions compared to Lenhart et al (2015) Instead thismay be explained by differing methods to compute respira-tion while Lenhart et al (2015) assume a globally uniformratio of respiration to NPP of a value of 2 to estimate respira-tion LiBry simulates respiration independently as a species-specific function of temperature and water status This re-sults in a lower global average value of around 1 for the ra-tio of respiration to NPP predicted by LiBry Our estimatedratio of respiration to NPP agrees well with laboratory mea-surements but it is in general difficult to compare a globalecosystem-scale value to small-scale and short-term obser-vations

Our simulated global pattern of N2O emissions is slightlydifferent than that shown in Lenhart et al (2015) who esti-mate highest values in the boreal zone and only intermedi-ate values in the humid tropics This can be explained bytheir assumed constant ratio of respiration to NPP whichmakes their global pattern of N2O emissions identical to thatof NPP which is shown in Elbert et al (2012) In LiBryhowever the simulated ratio of respiration to NPP increasestowards higher surface temperatures in the tropics (Fig 4)Furthermore the ratio shows large spatial variation Evaluat-ing this simulated pattern is difficult since estimates whichare extrapolated to the large scale such as the NPP estimateby Elbert et al (2012) are not available for respiration bylichens and bryophytes However observed ratios of respira-tion to NPP of lichens and bryophytes vary considerably atthe species level as shown by for example Lenhart et al(2015) Using a constant ratio of respiration to NPP maytherefore introduce a bias in the estimated spatial distribu-tion of N2O emissions

Small-scale measurements of N2O emissions by lichensand bryophytes may show considerable variation Thesources of this variation may be physiological differencesbetween species variation of associated microbial communi-ties andor heterogeneity in climatic conditions We examinethe relative importance for respiration of differences betweenspecies compared to climatic differences with LiBry sincethe model simulates various physiological strategies and rep-resents variation in climatic conditions at the global scaleThereby we assume that the relationship between respira-tion and N2O emissions is relatively insensitive to climaticconditions and physiological differences between species assuggested by the experiments by Lenhart et al (2015) Ta-ble 2 shows that both differences between artificial species

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1600 P Porada et al Nitrous oxide emissions by lichens and bryophytes

and different climatic conditions are important for variationof N2O emissions Upscaling of N2O emission rates mea-sured in the field may therefore be subject to considerableuncertainty Modelling approaches in this direction shouldprobably account for both interspecific variation in processesassociated with N2O release by lichens and bryophytes aswell as variation in climatic conditions

Although our approach considers the most importantsources of variation in N2O emissions by lichens andbryophytes it is associated with uncertainties that should bediscussed further These uncertainties mainly result from ourmethod to estimate respiration and from assumptions con-cerning the empirical relationship between respiration andN2O emissions

Respiration and the ratio of respiration to NPP simulatedby LiBry are difficult to validate since the number of labo-ratory or field studies which measure not only NPP but alsoGPP and respiration is not very high Moreover long-termmeasurements of respiration would be required to determinethe ratio of respiration to NPP Otherwise assumptions aboutthe contribution of respiration in the dark to total respirationare necessary

To obtain N2O emissions from respiration our resultsrely on the laboratory incubation measurements and the cal-culated ratio of N2O emissions to respiration presented inLenhart et al (2015) Furthermore our approach considerseffects of variation in climatic conditions on N2O emissionsby lichens and bryophytes Hence it is necessary to dis-cuss the sensitivity of the relationship between respirationand N2O emissions to a range of climatic conditions Asshown in Lenhart et al (2015 Fig 3) the relationship be-tween respiration and N2O emissions seems to be insensi-tive to temperature changes for the tested species Likewisevariations in water content have no clear effect on the rela-tionship between N2O release and respiration (Lenhart et al2015 Fig S3) Although the sensitivities of N2O release totemperature and water content are similar to those of respira-tion across species the relationship between N2O release andrespiration shows interspecific variation However in spite ofa large number of around 40 sampled species the relation-ship shows a relatively narrow 90 confidence interval of113 to 207 ng N2O (mg CO2)minus1 (Lenhart et al 2015) Thissuggests that the mechanism of N2O release by lichens andbryophytes is similar between different species

To analyse the relation between the production of N2Oand respiratory CO2 in greater detail measurements of bothfluxes by means of online gas exchange measurements wouldbe needed which then could be linked to the observed wa-ter status of the organisms Since LiBry explicitly representsthe dynamic water saturation of lichens and bryophytes thiswould allow a more process-based prediction of the durationand magnitude of N2O emissions In this way the uncertaintyassociated with our approach would be reduced facilitatingan improved estimate of global N2O emissions by lichensand bryophytes

In order to assess model-based estimates of N2O emis-sions by lichens and bryophytes a relatively large number offield measurements are necessary Currently most N2O mea-surements independent of the substrate or organisms mea-sured generally suffer from major uncertainties additionallyto variation from functional diversity and differing climaticconditions first the majority of these studies have been con-ducted using the acetylene inhibition technique The idea ofthis method is to inhibit the last denitrification step so thatthe measured N2O amounts should reveal the sum of N2Oand N2 release during denitrification under natural condi-tions It has however been shown quite a while ago thatthis method leads to an underestimation of denitrificationunder oxic conditions (Bollmann and Conrad 1997) Sec-ondly the most widely used measuring technique has beenthe closed-chamber method which is inexpensive and easyto use This however has major shortcomings as environ-mental conditions are hard to control and only limited surfaceareas can be measured (Butterbach-Bahl et al 2013 Groff-man 2012) Furthermore the limited temporal resolution ofchamber measurements may affect estimated N2O emissions(Barton et al 2015) Thirdly depending on the environmen-tal conditions under which the experiment is performed ndashparticularly water temperature and nutrient conditions ndash theobtained N2O emission rates could differ widely Thus it isindispensable to report and consider the exact environmen-tal conditions under which the measurements were made andto restrict natural emission data to those assessed under typi-cally occurring natural conditions

Respiration by lichens and bryophytes is not the only pro-cess which can be used to estimate their N2O emissionsBarger et al (2013) report a relationship between nitrogenfixation and N2O release in biological soil crusts which in-clude lichens and bryophytes as well as soil bacteria andalgae For this approach however reliable nitrogen fixationdata are sparse It is also possible to estimate the demand fornitrogen by lichens and bryophytes with LiBry with an un-certainty range of around 1 order of magnitude (Porada et al2014) However it is not straightforward to derive realizednitrogen uptake or nitrogen fixation from this since LiBrydoes not yet include processes related to nitrogen uptake ormetabolization of nitrogen species Therefore for this studywe chose the relation between respiration and N2O release toquantify N2O emissions by lichens and bryophytes

Our simulated global N2O emissions by lichens andbryophytes of 027 (019ndash035) (Tg N2O) yearminus1 amount toaround 3 of global N2O emissions from natural sources onland (Ciais et al 2013) This value may sound low at firstglance but it equals about 50 of the atmospheric deposi-tion of N2O into the oceans or 25 of the deposition on land(Ciais et al 2013) Considering that N2O has a strong nega-tive effect on stratospheric ozone and a significant warmingpotential as a greenhouse gas even relatively small emissionsshould not be neglected in global budgets

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1601

The study by Zhuang et al (2012) estimates global pat-terns of N2O emission from soils and finds that the hu-mid tropics contribute most to global N2O emission due tohigh temperature and precipitation Our simulated pattern ofglobal N2O emissions by lichens and bryophytes also showsa hotspot in the humid tropics but the relative contributionof the boreal zone to the global flux seems to be higher thanin Zhuang et al (2012) This probably results from the highsimulated NPP in the boreal zone particularly on the groundwhich compensates for the lower respiration and thereforeN2O emission per productivity due to low temperatures Rel-ative contributions of lichens and bryophytes to N2O emis-sions are highest for ecosystems in desert regions and athigh latitudes which agrees with the results by Lenhart et al(2015)

5 Conclusions

We estimate large-scale spatial patterns and global values ofN2O emissions by lichens and bryophytes from a process-based model of their productivity and respiration Our resultssuggest a significant contribution of lichens and bryophytesto global N2O emissions albeit at the lower end of the rangeof a previous empirical estimate Since both approachesuse respiration to derive N2O emissions our lower estimatelikely results from a different method to predict respirationcompared to the empirical approach Hence while estimatesof productivity are relatively well constrained evaluatingmodels with regard to estimated respiration may improvepredictions of N2O emissions by lichens and bryophytesOne important finding derived from our simulation is that theratio of respiration to NPP by lichens and bryophytes showsspatial variation and a latitudinal gradient at the global scaleThis means that productivity and N2O emissions by the or-ganisms are not necessarily correlated and that tropical re-gions may show higher emissions than polar regions giventhe same NPP Furthermore we show that both physiologicalvariation among species and variation in climatic conditionsare relevant for variation in respiration and consequentlyN2O emissions Ecosystem-scale estimates of N2O emis-sions by lichens and bryophytes should therefore include suf-ficient ranges of species and climatic conditions to avoid bi-ased results Our results build on the empirical finding thatN2O emissions by lichens and bryophytes are linearly re-lated to their respiration This relationship is relatively in-sensitive to climatic conditions and shows no large varia-tion between species However the relationship is based onclosed-chamber measurements Therefore it would be use-ful to perform online gas exchange measurements of N2Oemissions and respiration to test the effect of climatic con-ditions on the relationship between N2O release and respi-ration Furthermore using alternative approaches to estimateN2O emissions by lichens and bryophytes may be helpful toconstrain our approach

Code and data availability The non-vascular vegetation model Li-Bry used here is combined with an interface for parallel computingwhich was developed at the Max Planck Institute for Biogeochem-istry Jena Germany LiBry without the interface is freely availableas long as the names of the copyright holders and a disclaimer aredistributed along with the code in source or binary form The codeis available from the corresponding author upon requestModel output data which are presented as maps in this study areavailable as netCDF files from the authors on request

Author contributions Philipp Porada Bettina Weber andAxel Kleidon designed the model simulations and Philipp Poradacarried them out Philipp Porada prepared the manuscript withcontributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This work has been supported by the PAGE21project grant agreement number 282700 funded by the EC SeventhFramework Programme theme FP7-ENV-2011 and the CARBOP-ERM project grant agreement number 03G0836B funded by theBMBF (German Ministry for Science and Education) The authorsthank the Max Planck Society (Nobel Laureate Fellowship forBettina Weber) for financial support The Max Planck Institute forBiogeochemistry provided computational resources

Edited by V BrovkinReviewed by two anonymous referees

References

Abed R Lam P de Beer D and Stief P High rates of den-itrification and nitrous oxide emission in arid biological soilcrusts from the Sultanate of Oman ISME J 7 1862ndash1875doi101038ismej201355 2013

Barger N Castle S and Dean G Denitrification from nitrogen-fixing biologically crusted soils in a cool desert environ-ment southeast Utah USA Ecological Processes 2 1ndash9doi1011862192-1709-2-16 2013

Barger N Weber B Garcia-Pichel F Zaady E and Belnap JPatterns and Controls on Nitrogen Cycling of Biological SoilCrusts in Biological Soil Crusts An Organizing Principle inDrylands edited by Weber B Buumldel B and Belnap J 257ndash285 Springer International Publishing Cham doi101007978-3-319-30214-0_14 2016

Barton L Wolf B Rowlings D Scheer C Kiese R Grace PStefanova K and Butterbach-Bahl K Sampling frequency af-fects estimates of annual nitrous oxide fluxes Scientific Reports5 15912 doi101038srep15912 2015

Bollmann A and Conrad R Acetylene blockage technique leadsto underestimation of denitrification rates in oxic soils due toscavenging of intermediate nitric oxide Soil Biol Biochem 291067ndash1077 doi101016S0038-0717(97)00007-2 1997

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1600 P Porada et al Nitrous oxide emissions by lichens and bryophytes

and different climatic conditions are important for variationof N2O emissions Upscaling of N2O emission rates mea-sured in the field may therefore be subject to considerableuncertainty Modelling approaches in this direction shouldprobably account for both interspecific variation in processesassociated with N2O release by lichens and bryophytes aswell as variation in climatic conditions

Although our approach considers the most importantsources of variation in N2O emissions by lichens andbryophytes it is associated with uncertainties that should bediscussed further These uncertainties mainly result from ourmethod to estimate respiration and from assumptions con-cerning the empirical relationship between respiration andN2O emissions

Respiration and the ratio of respiration to NPP simulatedby LiBry are difficult to validate since the number of labo-ratory or field studies which measure not only NPP but alsoGPP and respiration is not very high Moreover long-termmeasurements of respiration would be required to determinethe ratio of respiration to NPP Otherwise assumptions aboutthe contribution of respiration in the dark to total respirationare necessary

To obtain N2O emissions from respiration our resultsrely on the laboratory incubation measurements and the cal-culated ratio of N2O emissions to respiration presented inLenhart et al (2015) Furthermore our approach considerseffects of variation in climatic conditions on N2O emissionsby lichens and bryophytes Hence it is necessary to dis-cuss the sensitivity of the relationship between respirationand N2O emissions to a range of climatic conditions Asshown in Lenhart et al (2015 Fig 3) the relationship be-tween respiration and N2O emissions seems to be insensi-tive to temperature changes for the tested species Likewisevariations in water content have no clear effect on the rela-tionship between N2O release and respiration (Lenhart et al2015 Fig S3) Although the sensitivities of N2O release totemperature and water content are similar to those of respira-tion across species the relationship between N2O release andrespiration shows interspecific variation However in spite ofa large number of around 40 sampled species the relation-ship shows a relatively narrow 90 confidence interval of113 to 207 ng N2O (mg CO2)minus1 (Lenhart et al 2015) Thissuggests that the mechanism of N2O release by lichens andbryophytes is similar between different species

To analyse the relation between the production of N2Oand respiratory CO2 in greater detail measurements of bothfluxes by means of online gas exchange measurements wouldbe needed which then could be linked to the observed wa-ter status of the organisms Since LiBry explicitly representsthe dynamic water saturation of lichens and bryophytes thiswould allow a more process-based prediction of the durationand magnitude of N2O emissions In this way the uncertaintyassociated with our approach would be reduced facilitatingan improved estimate of global N2O emissions by lichensand bryophytes

In order to assess model-based estimates of N2O emis-sions by lichens and bryophytes a relatively large number offield measurements are necessary Currently most N2O mea-surements independent of the substrate or organisms mea-sured generally suffer from major uncertainties additionallyto variation from functional diversity and differing climaticconditions first the majority of these studies have been con-ducted using the acetylene inhibition technique The idea ofthis method is to inhibit the last denitrification step so thatthe measured N2O amounts should reveal the sum of N2Oand N2 release during denitrification under natural condi-tions It has however been shown quite a while ago thatthis method leads to an underestimation of denitrificationunder oxic conditions (Bollmann and Conrad 1997) Sec-ondly the most widely used measuring technique has beenthe closed-chamber method which is inexpensive and easyto use This however has major shortcomings as environ-mental conditions are hard to control and only limited surfaceareas can be measured (Butterbach-Bahl et al 2013 Groff-man 2012) Furthermore the limited temporal resolution ofchamber measurements may affect estimated N2O emissions(Barton et al 2015) Thirdly depending on the environmen-tal conditions under which the experiment is performed ndashparticularly water temperature and nutrient conditions ndash theobtained N2O emission rates could differ widely Thus it isindispensable to report and consider the exact environmen-tal conditions under which the measurements were made andto restrict natural emission data to those assessed under typi-cally occurring natural conditions

Respiration by lichens and bryophytes is not the only pro-cess which can be used to estimate their N2O emissionsBarger et al (2013) report a relationship between nitrogenfixation and N2O release in biological soil crusts which in-clude lichens and bryophytes as well as soil bacteria andalgae For this approach however reliable nitrogen fixationdata are sparse It is also possible to estimate the demand fornitrogen by lichens and bryophytes with LiBry with an un-certainty range of around 1 order of magnitude (Porada et al2014) However it is not straightforward to derive realizednitrogen uptake or nitrogen fixation from this since LiBrydoes not yet include processes related to nitrogen uptake ormetabolization of nitrogen species Therefore for this studywe chose the relation between respiration and N2O release toquantify N2O emissions by lichens and bryophytes

Our simulated global N2O emissions by lichens andbryophytes of 027 (019ndash035) (Tg N2O) yearminus1 amount toaround 3 of global N2O emissions from natural sources onland (Ciais et al 2013) This value may sound low at firstglance but it equals about 50 of the atmospheric deposi-tion of N2O into the oceans or 25 of the deposition on land(Ciais et al 2013) Considering that N2O has a strong nega-tive effect on stratospheric ozone and a significant warmingpotential as a greenhouse gas even relatively small emissionsshould not be neglected in global budgets

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1601

The study by Zhuang et al (2012) estimates global pat-terns of N2O emission from soils and finds that the hu-mid tropics contribute most to global N2O emission due tohigh temperature and precipitation Our simulated pattern ofglobal N2O emissions by lichens and bryophytes also showsa hotspot in the humid tropics but the relative contributionof the boreal zone to the global flux seems to be higher thanin Zhuang et al (2012) This probably results from the highsimulated NPP in the boreal zone particularly on the groundwhich compensates for the lower respiration and thereforeN2O emission per productivity due to low temperatures Rel-ative contributions of lichens and bryophytes to N2O emis-sions are highest for ecosystems in desert regions and athigh latitudes which agrees with the results by Lenhart et al(2015)

5 Conclusions

We estimate large-scale spatial patterns and global values ofN2O emissions by lichens and bryophytes from a process-based model of their productivity and respiration Our resultssuggest a significant contribution of lichens and bryophytesto global N2O emissions albeit at the lower end of the rangeof a previous empirical estimate Since both approachesuse respiration to derive N2O emissions our lower estimatelikely results from a different method to predict respirationcompared to the empirical approach Hence while estimatesof productivity are relatively well constrained evaluatingmodels with regard to estimated respiration may improvepredictions of N2O emissions by lichens and bryophytesOne important finding derived from our simulation is that theratio of respiration to NPP by lichens and bryophytes showsspatial variation and a latitudinal gradient at the global scaleThis means that productivity and N2O emissions by the or-ganisms are not necessarily correlated and that tropical re-gions may show higher emissions than polar regions giventhe same NPP Furthermore we show that both physiologicalvariation among species and variation in climatic conditionsare relevant for variation in respiration and consequentlyN2O emissions Ecosystem-scale estimates of N2O emis-sions by lichens and bryophytes should therefore include suf-ficient ranges of species and climatic conditions to avoid bi-ased results Our results build on the empirical finding thatN2O emissions by lichens and bryophytes are linearly re-lated to their respiration This relationship is relatively in-sensitive to climatic conditions and shows no large varia-tion between species However the relationship is based onclosed-chamber measurements Therefore it would be use-ful to perform online gas exchange measurements of N2Oemissions and respiration to test the effect of climatic con-ditions on the relationship between N2O release and respi-ration Furthermore using alternative approaches to estimateN2O emissions by lichens and bryophytes may be helpful toconstrain our approach

Code and data availability The non-vascular vegetation model Li-Bry used here is combined with an interface for parallel computingwhich was developed at the Max Planck Institute for Biogeochem-istry Jena Germany LiBry without the interface is freely availableas long as the names of the copyright holders and a disclaimer aredistributed along with the code in source or binary form The codeis available from the corresponding author upon requestModel output data which are presented as maps in this study areavailable as netCDF files from the authors on request

Author contributions Philipp Porada Bettina Weber andAxel Kleidon designed the model simulations and Philipp Poradacarried them out Philipp Porada prepared the manuscript withcontributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This work has been supported by the PAGE21project grant agreement number 282700 funded by the EC SeventhFramework Programme theme FP7-ENV-2011 and the CARBOP-ERM project grant agreement number 03G0836B funded by theBMBF (German Ministry for Science and Education) The authorsthank the Max Planck Society (Nobel Laureate Fellowship forBettina Weber) for financial support The Max Planck Institute forBiogeochemistry provided computational resources

Edited by V BrovkinReviewed by two anonymous referees

References

Abed R Lam P de Beer D and Stief P High rates of den-itrification and nitrous oxide emission in arid biological soilcrusts from the Sultanate of Oman ISME J 7 1862ndash1875doi101038ismej201355 2013

Barger N Castle S and Dean G Denitrification from nitrogen-fixing biologically crusted soils in a cool desert environ-ment southeast Utah USA Ecological Processes 2 1ndash9doi1011862192-1709-2-16 2013

Barger N Weber B Garcia-Pichel F Zaady E and Belnap JPatterns and Controls on Nitrogen Cycling of Biological SoilCrusts in Biological Soil Crusts An Organizing Principle inDrylands edited by Weber B Buumldel B and Belnap J 257ndash285 Springer International Publishing Cham doi101007978-3-319-30214-0_14 2016

Barton L Wolf B Rowlings D Scheer C Kiese R Grace PStefanova K and Butterbach-Bahl K Sampling frequency af-fects estimates of annual nitrous oxide fluxes Scientific Reports5 15912 doi101038srep15912 2015

Bollmann A and Conrad R Acetylene blockage technique leadsto underestimation of denitrification rates in oxic soils due toscavenging of intermediate nitric oxide Soil Biol Biochem 291067ndash1077 doi101016S0038-0717(97)00007-2 1997

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

P Porada et al Nitrous oxide emissions by lichens and bryophytes 1601

The study by Zhuang et al (2012) estimates global pat-terns of N2O emission from soils and finds that the hu-mid tropics contribute most to global N2O emission due tohigh temperature and precipitation Our simulated pattern ofglobal N2O emissions by lichens and bryophytes also showsa hotspot in the humid tropics but the relative contributionof the boreal zone to the global flux seems to be higher thanin Zhuang et al (2012) This probably results from the highsimulated NPP in the boreal zone particularly on the groundwhich compensates for the lower respiration and thereforeN2O emission per productivity due to low temperatures Rel-ative contributions of lichens and bryophytes to N2O emis-sions are highest for ecosystems in desert regions and athigh latitudes which agrees with the results by Lenhart et al(2015)

5 Conclusions

We estimate large-scale spatial patterns and global values ofN2O emissions by lichens and bryophytes from a process-based model of their productivity and respiration Our resultssuggest a significant contribution of lichens and bryophytesto global N2O emissions albeit at the lower end of the rangeof a previous empirical estimate Since both approachesuse respiration to derive N2O emissions our lower estimatelikely results from a different method to predict respirationcompared to the empirical approach Hence while estimatesof productivity are relatively well constrained evaluatingmodels with regard to estimated respiration may improvepredictions of N2O emissions by lichens and bryophytesOne important finding derived from our simulation is that theratio of respiration to NPP by lichens and bryophytes showsspatial variation and a latitudinal gradient at the global scaleThis means that productivity and N2O emissions by the or-ganisms are not necessarily correlated and that tropical re-gions may show higher emissions than polar regions giventhe same NPP Furthermore we show that both physiologicalvariation among species and variation in climatic conditionsare relevant for variation in respiration and consequentlyN2O emissions Ecosystem-scale estimates of N2O emis-sions by lichens and bryophytes should therefore include suf-ficient ranges of species and climatic conditions to avoid bi-ased results Our results build on the empirical finding thatN2O emissions by lichens and bryophytes are linearly re-lated to their respiration This relationship is relatively in-sensitive to climatic conditions and shows no large varia-tion between species However the relationship is based onclosed-chamber measurements Therefore it would be use-ful to perform online gas exchange measurements of N2Oemissions and respiration to test the effect of climatic con-ditions on the relationship between N2O release and respi-ration Furthermore using alternative approaches to estimateN2O emissions by lichens and bryophytes may be helpful toconstrain our approach

Code and data availability The non-vascular vegetation model Li-Bry used here is combined with an interface for parallel computingwhich was developed at the Max Planck Institute for Biogeochem-istry Jena Germany LiBry without the interface is freely availableas long as the names of the copyright holders and a disclaimer aredistributed along with the code in source or binary form The codeis available from the corresponding author upon requestModel output data which are presented as maps in this study areavailable as netCDF files from the authors on request

Author contributions Philipp Porada Bettina Weber andAxel Kleidon designed the model simulations and Philipp Poradacarried them out Philipp Porada prepared the manuscript withcontributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This work has been supported by the PAGE21project grant agreement number 282700 funded by the EC SeventhFramework Programme theme FP7-ENV-2011 and the CARBOP-ERM project grant agreement number 03G0836B funded by theBMBF (German Ministry for Science and Education) The authorsthank the Max Planck Society (Nobel Laureate Fellowship forBettina Weber) for financial support The Max Planck Institute forBiogeochemistry provided computational resources

Edited by V BrovkinReviewed by two anonymous referees

References

Abed R Lam P de Beer D and Stief P High rates of den-itrification and nitrous oxide emission in arid biological soilcrusts from the Sultanate of Oman ISME J 7 1862ndash1875doi101038ismej201355 2013

Barger N Castle S and Dean G Denitrification from nitrogen-fixing biologically crusted soils in a cool desert environ-ment southeast Utah USA Ecological Processes 2 1ndash9doi1011862192-1709-2-16 2013

Barger N Weber B Garcia-Pichel F Zaady E and Belnap JPatterns and Controls on Nitrogen Cycling of Biological SoilCrusts in Biological Soil Crusts An Organizing Principle inDrylands edited by Weber B Buumldel B and Belnap J 257ndash285 Springer International Publishing Cham doi101007978-3-319-30214-0_14 2016

Barton L Wolf B Rowlings D Scheer C Kiese R Grace PStefanova K and Butterbach-Bahl K Sampling frequency af-fects estimates of annual nitrous oxide fluxes Scientific Reports5 15912 doi101038srep15912 2015

Bollmann A and Conrad R Acetylene blockage technique leadsto underestimation of denitrification rates in oxic soils due toscavenging of intermediate nitric oxide Soil Biol Biochem 291067ndash1077 doi101016S0038-0717(97)00007-2 1997

wwwbiogeosciencesnet1415932017 Biogeosciences 14 1593ndash1602 2017

1602 P Porada et al Nitrous oxide emissions by lichens and bryophytes

Brankatschk R Fischer T Veste M and Zeyer J Successionof N cycling processes in biological soil crusts on a CentralEuropean inland dune FEMS Microbiol Ecol 83 149ndash160doi101111j1574-6941201201459x 2013

Butterbach-Bahl K Baggs E Dannenmann M Kiese R andZechmeister-Boltenstern S Nitrous oxide emissions from soilshow well do we understand the processes and their controlsPhilos T R Soc B 368 23713120 doi101098rstb201301222013

Ciais P Sabine C Bala G Bopp L Brovkin V Canadell JChhabra A DeFries R Galloway J Heimann M Jones CLe Queacutereacute C Myneni R Piao S and Thornton P Carbonand Other Biogeochemical Cycles in Climate Change 2013The Physical Science Basis Contribution of Working Group Ito the Fifth Assessment Report of the Intergovernmental Panelon Climate Change edited by Stocker T Qin D Plattner G-K Tignor M Allen S Boschung J Nauels A Xia Y BexV and Midgley P Cambridge University Press Cambridge UKand New York NY USA 2013

Darrouzet-Nardi A Reed S Grote E and Belnap J Observa-tions of net soil exchange of CO2 in a dryland show experimentalwarming increases carbon losses in biocrust soils Biogeochem-istry 126 363ndash378 doi101007s10533-015-0163-7 2015

DeLuca T Zackrisson O Nilsson M and Sellstedt A Quanti-fying nitrogen-fixation in feather moss carpets of boreal forestsNature 419 917ndash920 doi101038nature01051 2002

Elbert W Weber B Burrows S Steinkamp J Buumldel B An-dreae M and Poumlschl U Contribution of cryptogamic covers tothe global cycles of carbon and nitrogen Nat Geosci 5 459ndash462 doi101038ngeo1486 2012

Farquhar G and von Caemmerer S Modelling of Photosyn-thetic Response to Environmental Conditions in Encyclopediaof Plant Physiology edited by Lange O Nobel P OsmondC and Ziegler H Vol 12B Springer Heidelberg 1982

Gaumlrdenaumls A Aringgren G Bird J Clarholm M Hallin S In-eson P Kaumltterer T Knicker H Nilsson S Naumlsholm TOgle S Paustian K Persson T and Stendahl J Knowl-edge gaps in soil carbon and nitrogen interactions ndash Frommolecular to global scale Soil Biol Biochem 43 702ndash717doi101016jsoilbio201004006 2011

Groffman P Terrestrial denitrification challenges and opportuni-ties Ecological Processes 1 1ndash11 doi1011862192-1709-1-11 2012

Hawkes C Nitrogen cycling mediated by biological soil crustsand arbuscular mycorrhizal fungi Ecology 84 1553ndash1562doi1018900012-9658(2003)084[1553NCMBBS]20CO22003

Johnson S Neuer S and Garcia-Pichel F Export of nitroge-nous compounds due to incomplete cycling within biologi-cal soil crusts of arid lands Environ Microbiol 9 680ndash689doi101111j1462-2920200601187x 2007

Lenhart K Weber B Elbert W Steinkamp J Clough TCrutzen P Poumlschl U and Keppler F Nitrous oxide andmethane emissions from cryptogamic covers Glob ChangeBiol 21 3889ndash3900 doi101111gcb12995 2015

Nie Y Wang M Tahvanainen T and Hashidoko Y Nitrous ox-ide emission potentials of Burkholderia species isolated from theleaves of a boreal peat moss Sphagnum fuscum Biosci Biotech

Bioch 79 2086ndash2095 doi10108009168451201510614202015

Olson D Dinerstein E Wikramanayake E Burgess N PowellG Underwood E DrsquoAmico J Itoua I Strand H MorrisonJ Loucks C Allnutt T Ricketts T Kura Y Lamoreux JWettengel W Hedao P and Kassem K Terrestrial Ecoregionsof the World A New Map of Life on Earth BioScience 51 933ndash938 2001

Porada P Weber B Elbert W Poumlschl U and Kleidon AEstimating global carbon uptake by lichens and bryophyteswith a process-based model Biogeosciences 10 6989ndash7033doi105194bg-10-6989-2013 2013

Porada P Weber B Elbert W Poumlschl U and Kleidon AEstimating Impacts of Lichens and Bryophytes on GlobalBiogeochemical Cycles of Nitrogen and Phosphorus and onChemical Weathering Global Biogeochem Cy 28 71ndash85doi1010022013GB004705 2014

Porada P Ekici A and Beer C Effects of bryophyte andlichen cover on permafrost soil temperature at large scaleThe Cryosphere 10 2291ndash2315 doi105194tc-10-2291-20162016a

Porada P Lenton T Pohl A Weber B Mander L Don-nadieu Y Beer C Poumlschl U and Kleidon A High po-tential for weathering and climate effects of non-vascular veg-etation in the Late Ordovician Nat Commun 7 12113doi101038ncomms12113 2016b

Ravishankara A Daniel J and Portmann R Nitrous Ox-ide (N2O) The Dominant Ozone-Depleting SubstanceEmitted in the 21st Century Science 326 123ndash125doi101126science1176985 2009

Sancho L Belnap J Colesie C Raggio J and Weber BCarbon Budgets of Biological Soil Crusts at Micro- Meso-and Global Scales in Biological Soil Crusts An OrganizingPrinciple in Drylands edited by Weber B Buumldel B andBelnap J 287ndash304 Springer International Publishing Chamdoi101007978-3-319-30214-0_15 2016

Smart D and Bloom A Wheat leaves emit nitrous oxide dur-ing nitrate assimilation P Natl Acad Sciences USA 98 7875ndash7878 doi101073pnas131572798 2001

Stewart W Transfer of Biologically Fixed Nitrogen ina Sand Dune Slack Region Nature 214 603ndash604doi101038214603a0 1967

Strauss S Day T and Garcia-Pichel F Nitrogen cycling indesert biological soil crusts across biogeographic regions in theSouthwestern United States Biogeochemistry 108 171ndash182doi101007s10533-011-9587-x 2012

Weber B Wu D Tamm A Ruckteschler N Rodriacuteguez-Caballero E Steinkamp J Meusel H Elbert W BehrendtT Soumlrgel M Cheng Y Crutzen P J Su H and Poumlschl UBiological soil crusts accelerate the nitrogen cycle through largeNO and HONO emissions in drylands P Natl Acad Sci USA112 15384ndash15389 doi101073pnas1515818112 2015

Zhuang Q Lu Y and Chen M An inventory of global N2Oemissions from the soils of natural terrestrial ecosystems AtmosEnviron 47 66ndash75 doi101016jatmosenv201111036 2012

Biogeosciences 14 1593ndash1602 2017 wwwbiogeosciencesnet1415932017