Carbon–nitrogen interactions on land at global scales: current understanding in modelling climate biosphere feedbacks

Available online at www.sciencedirect.com

Carbon–nitrogen interactions on land at global scales: current understanding in modelling climate biosphere feedbacks So¨nke Zaehle and Daniela Dalmonech Interactions between the terrestrial carbon (C) and nitrogen (N) cycles shape the response of ecosystems to global change. The limitation of ecosystem C storage due to N availability, and the response of N2O emissions to environmental conditions and N addition have been intensively studied at the site level. However, their contribution to biosphere–climate interactions at regional to global scales remains unclear. A growing number of global terrestrial biogeochemical models provide a means to scale ecological understanding of the nitrogen cycle to regional and global scales with the ultimate aim to investigate the magnitude of nitrogen cycling effects on global biogeochemistry, as well as their indirect consequences for biogeophysical landatmosphere interactions. Key challenges to modelling the coupled terrestrial carbon–nitrogen cycles arise from the need to account for micro-scale processes to represent and quantify important N fluxes, uncertainties in the representation of key carbon–nitrogen cycle couplings at the ecosystem scale, and vagaries in the available observations to constrain global models. The new generation of carbon–nitrogen cycle models suggests that reactive nitrogen deposition is associated with a moderate increase in current terrestrial carbon sequestration, providing a small climate cooling effect. The models further unanimously demonstrate that nitrogen cycling reduces both global carbon sequestration due to CO2 fertilisation and the carbon losses associated with climate change on land, in sum leading to an acceleration of carbon accumulation in the atmosphere relative to C-cycle only models. A recent study furthermore suggests a moderate positive interaction between terrestrial N2O emissions and recent climatic changes, although the atmospheric increase in N2O over the last few decades appears to be mostly associated with anthropogenic Nr additions to the terrestrial biosphere. At least some of these models can be used to assess the biogeophysical effects of N cycling through altered albedo and changed sensible and latent heat fluxes, but no study so far has assessed these consequences explicitly. Address Max Planck Institute for Biogeochemistry, Biogeochemical Systems Department, Hans-Kno¨ll-Str. 10, D-07745 Jena, Germany Corresponding author: Zaehle, So¨nke ([email protected])

Current Opinion in Environmental Sustainability 2011, 3:311–320 This review comes from a themed issue on Carbon and nitrogen cycles Edited by Carolien Kroeze and Lex Bouwman Received 5 May 2011; Accepted 12 August 2011 Available online 8th September 2011 www.sciencedirect.com

1877-3435/$ – see front matter # 2011 Elsevier B.V. All rights reserved. DOI 10.1016/j.cosust.2011.08.008

Introduction Nitrogen is a fundamental component of living organisms. Ecosystem available forms of nitrogen (ammonium, nitrate amongst other nitrogen-oxides), hereafter reactive N (Nr), are scarce in unperturbed ecosystems (Figure 1). The (carbon) productivity of plants and soil organisms strongly depends on nitrogen, imposing stoichiometric constraints at the level of an individual organism. These two facts lead to a tight coupling of the terrestrial nitrogen and carbon cycles through the competition of plants and soil organisms for nitrogen, as evidenced by the constrained flexibility of ecosystem C:N stoichiometry. Ecosystem N availability thereby has an important role in controlling the productivity, structure and dynamics of terrestrial ecosystems [1] with potentially strong effects on responses of the terrestrial net carbon balance to changes in atmospheric CO2 and climate. Past anthropogenic activities have about doubled the input of reactive N into the land biosphere between 1860 and today, either voluntary by land management and the creation and application of industrial fertilisers, or involuntary through the atmospheric transport and deposition of reactive Nr from, for example, fossil fuel burning [2,3]. Through the so-called nitrogen cascade [2], the anthropogenic Nr has caused widespread environmental consequences including the eutrophication of terrestrial and aquatic ecosystems, contributing to global acidification, the increase of atmospheric [N2O] and the loss of stratospheric ozone [4]. While the potential consequences of the terrestrial N limitation and the Nr perturbation are easily enlisted, the magnitudes of these effects on global biogeochemistry and interactions with the climate system are less easily quantified [5]. The last years have seen the emergence of several global coupled carbon–nitrogen cycle models [6,7–14] that have been developed with the objective to investigate the role of the natural and perturbed terrestrial nitrogen cycles in shaping the terrestrial net carbon balance and land-climate feedbacks. In this review, we give an overview of these recent developments, summarise their key findings with respect to the interactions of Current Opinion in Environmental Sustainability 2011, 3:311–320

312 Carbon and nitrogen cycles

Figure 1

Atmospheric CO2

Nr deposition

hν

Canopy

N2fixation Albedo & sensible heat flux Transpiration Nr volatilisation

Litter Pool Soil Organic Matter Pool

Roots

Mineral Nr pool

Soil moisture Nr leaching Current Opinion in Environmental Sustainability

The coupled cycles of carbon (green), nitrogen (red) and water (blue) in terrestrial ecosystems as considered by the current generation of terrestrial coupled carbon–nitrogen cycle models. Nitrogen enters the terrestrial biosphere due to biological fixation and reactive nitrogen (Nr) deposition. Vegetation and soil organisms compete for mineral soil N, which itself is prone to leaching and biologically mediated gaseous losses, for example, in the form of the greenhouse gas N2O. By its effect on tissue N concentrations, leaf level photosynthetic capacity, and leaf area, N availability affects the terrestrial albedo, the energy and water fluxes, as well as plant C production, respiration, and soil organic matter decay. Carbon and nitrogen flow in parallel between vegetation, litter and soil organic matter respecting the stoichiometry of the various organic matter pools. N limitation occurs where the ecosystem carbon productivity is limited stronger by N availability than by other limiting factors (light, water, phosphorous, etc.).

the terrestrial biosphere and the climate system and outline some of the key challenges to global carbon– nitrogen cycle modelling. The questions at hand evolve around two topics: 1. How do the coupled terrestrial carbon–nitrogen dynamics respond to past (and future) global changes and what are the implications (and potential feedbacks) for the build-up of greenhouse gases such as CO2 and N2O in the atmosphere (and thus climate change)? 2. What are the effects of anthropogenic Nr additions to the atmosphere and land ecosystems in terms of climate forcing? These effects may either be positive (in terms of increased climate forcing) due to the enhanced emission of N2O or increased production of tropospheric ozone, or negative (in terms of decreased climate forcing) through decreased stratospheric ozone and increased aerosol formation. Increased atmospheric Nr loading may also indirectly influence climate through by its effect on terrestrial carbon sequestration, with N deposition generally stimulating growth and thus carbon uptake, and increased tropospheric ozone impairing plant health and growth.

One further needs to distinguish these ‘direct’ biogeochemical consequences of N cycling from the ‘indirect’ biogeophysical climate effects of N dynamics through terrestrial energy and water fluxes which can only be assessed in Earth system models. Current Opinion in Environmental Sustainability 2011, 3:311–320

Key features of current global C–N models Many of the new generation of coupled carbon–nitrogen cycle models are either further developments of previously published ecosystem dynamics models from the early-mid 1990s [15,16] and/or based on joining model components from different existing ecosystem and biogeochemical model concepts [17–19]. A distinct feature of some of the recent models is that they provide a more detailed representation of environmental physics, as they are designed to be coupled to a comprehensive Earth system model (Table 1). Through the strong coupling of biogeochemical and biogeophysical processes these models attempt to assess the full complexity of the C– N climate interactions, and feedbacks to the physical climate system. Figure 1 illustrates the carbon and nitrogen flows typically represented in global carbon–nitrogen cycle models. Because of the physiological dependence of C growth on nitrogen, the C cycle is constrained by the N availability of the ecosystem, affecting amongst others plant production, C allocation to different plant organs and the decomposition of dead organic material (litter) and soil organic matter. Ecosystem N availability is controlled by the balance between N inputs from biological nitrogen fixation and atmospheric deposition and biologically controlled N losses through leaching and denitrification. Implicit to the formulation of (many of) these models are the consequences of plant N availability on water and energy fluxes: N availability affects community structure and foliar area, thereby altering the surface albedo and turbulent energy fluxes, as well as foliar photosynthetic www.sciencedirect.com

www.sciencedirect.com Table 1 Overview on key features of the current generation of coupled carbon–nitrogen cycle climate models and recent global land carbon–nitrogen cycle models f(...) describes that the modelled process is a function of the term(s) in the brackets. Soil T = Soil temperature. Coupled carbon–nitrogen cycle climate models Model name

CLM-CN

CLIMBER-CN

LM3V-CN

O-CN

ISAM-CN

LPJ-DyN

CASA-CNP

NCIM

Current Opinion in Environmental Sustainability 2011, 3:311–320

Reference N effect on photosynthesis N limitation on C growth Prognostic leaf area N effect on allocation C:N stoichiometry

[6 ] f(N deficit)

[7] Implicit

[8] f(N deficit)

[10] f(N deficit)

[11] f(leaf N)

[9] No

[12] No

[13] f(leaf N)

[14] No

GPP #

NPP #

GPP #

GPP #

GPP # & NPP/GPP "

NPP #

NPP #

GPP #

NPP #

Yes No Fixed

No No Fixed

Yes No Fixed

Yes Yes Floating

Yes No Fixed

Yes No Floating

No No Floating

No Yes Fixed

N fixation

f(NPP)

No

Fixed

Yes No Fixed, buffered by storage pool f(N deficit, vegetation type)

f(climate, soil N)

f(climate)

Mass-balance based

f(N deficit, vegetation type)

N uptake a

Litter N immobilisation 1st order decay

Litter N immobilisation 1st order decay

Root mass, transpiration 1st order decay

Root mass, soil T

Soil T & moisture

No No

No No

DNDC-type [19] Yes

Root mass, soil T 1st order decay f(gross N min.) No

DNDC-type [19] Yes

Input from a process-based model Minimal uptake rate 1st order decay No No

Depending on microbial growth f(net N min.) N2O

Soil organic matter decomposition Denitrification N2O/NOx/NH3 production N effect on stomatal conductance Biogeochemistry effect on biogeophysics a

f(excess soil N) No

1st order decay No No

No

No

No

Yes

Yes

No

No

Yes

No

Yes

No

No

Yes

Yes

Water cycle only

Water cycle only

Water cycle only

No

Factors in addition to plant N demand and mineral soil N availability.

1st order decay

1st order decay

Soil T, leaf C

Carbon–nitrogen interactions on land biogeochemistry Zaehle and Dalmonech 313



Coupled carbon–nitrogen cycle models

IGSM-CN

314 Carbon and nitrogen cycles

activity, which through the coupling of photosynthesis and stomatal conductance affects evapotranspiration. Despite their common design, decisive carbon–nitrogen linkages are treated differently and in varying degrees of detail between the models, illustrating the limited consensus on how to scale process knowledge to the ecosystem and biome level (Table 1). A comprehensive review of these differences easily fills an entire book chapter. Only the main points are summarised here. Nitrogen limitation on plant C uptake

The available approaches range from a reduction of the gross or net primary C production to conform with the limits set by the plant available N and the stoichiometric N requirements for new tissue [6,8] or semi-empirical formulations linking production to plant available N [9,10,13], to an explicit simulation of the acclimation process of photosynthesis and plant growth to nitrogen stress by simulating changes in allocation patterns (foliar mass) and foliar N concentration, as well as accounting for sink limitation based on the C:N stoichiometry of newly grown tissue [11]. The first approach provides a direct, but non-observable, link between plant N status and (carbon) growth that can only be evaluated as emergent property by comparing plant nitrogen and carbon growth to observations. The latter approach links plant available N only indirectly to C production and requires more (and uncertain) parameters, at the advantage of simulating observable quantities of N limitation such as seasonally varying leaf photosynthetic activity which can help to quantify seasonal and longer-term changes in N limitation for instance in ecosystem manipulation experiments (e.g. [20]). In few cases, where photosynthetic rates are directly modified [10,11], the ensuing effects on canopy conductance allow for a more comprehensive estimation of the effects of N limitation in coupled biosphere-climate models through the effects on energy and water fluxes. In many of the models, foliage area adjusts to N availability indirectly due to changed net primary production, but only two models explicitly simulate the observed additional effect of N availability of plant allocation to below-ground biomass (directly affecting the C available for leaf area production) [11,14], which is important to understand the effects observed in ecosystem manipulation experiments (see Section ‘Plant nitrogen uptake and competition with soil microbes’). C:N stoichiometry

Most models assume constant C:N ratios for each tissue pools, that is, changes in ecosystem stoichiometry occur only due to changing carbon allocation patterns (i.e. differential investments into foliar, fine root and woody biomass) or changing the partitioning between phytomass and soil organic matter. Only few models allow tissue C:N ratios to acclimate to prevailing N availability [11,12,18]. Flexible C:N ratios allow to reproduce observed changes Current Opinion in Environmental Sustainability 2011, 3:311–320

in stoichiometry from ecosystem manipulation experiments and environmental gradient studies, for instance free-air CO2 enrichment experiments [20], and buffer N effects on C storage (e.g. increases in C:N under nutrient stress allow to store more C per unit N availability). However, such model dynamics may also be reproduced by a temporary N storage pool to buffer short-term fluctuations in N availability [10]. Flexible C:N stoichiometry permits further to explicitly simulate changing foliar N (and thus photosynthesis), as well as litter quality effects on the decay of soil organic matter, allowing for a wider range of feedback processes to be considered, and, importantly, of data sources to evaluate the simulated nutrient limitation. While flexibility in stoichiometry seems the preferable model approach, none of the current models incorporates a mechanistic understanding of the causes and limits of this flexibility, leading to uncertainty in the parameterisation of this plasticity. Biological nitrogen fixation

In its most simple form, biological nitrogen fixation is represented based on empirical relationships with climate parameters to estimate the geographic distribution of the N fixation independent of the actual plant N demand [21,22]. More sophisticated models assume that biological fixation is proportional to plant productivity [6], or apply physiological models of N fixation to simulate the nitrogen fixation capacity based on resource limitation [10,14,23]. Introducing an N-status-dependent N fixation term reduces the extent of simulated N limitation when C availability is increased (e.g. due to CO2 fertilisation), and may synergistically lead to more C storage [14]. Interactions of N fixation with climate change and CO2 fertilisation may substantially alter N fixation in the future, with potentially large effects on global C sequestration [24]. These effects are yet to be explored in a dynamic carbon–nitrogen cycle model. A caveat in modelling N fixation is the lack of accessible data on long-term biological N fixation to evaluate model projections, and the co-limitation of N fixation by phosphorous (P) and micronutrients such as molybdenum, requiring either the explicit modelling of these cycles, or at least some representation of the limits induced by them. Plant nitrogen uptake and competition with soil microbes

The models differ with respect to whether or not plants are assumed to be able to access all mineral soil nitrogen, whether or not the kinetics of N uptake are considered, whether or not N uptake is actively controlled by plants and whether or not carbon costs of N acquisition (in terms of respiration costs or C allocation to fine roots) are considered. Fisher et al. [23] provide a comprehensive model of active and passive plant nitrogen uptake mechanisms, but this model has yet to be coupled to a complete vegetation-soil nitrogen cycle model. To our knowledge, none of the global models explicitly simulates the uptake www.sciencedirect.com

Carbon–nitrogen interactions on land biogeochemistry Zaehle and Dalmonech 315

of organic form of N or the effects and dynamics of mycorrhizal symbiosis, despite its primary importance in boreal soils [25]. One frequently made criticism of carbon–nitrogen cycle models (and carbon cycle models in general) is their treatment of soil organic matter as first order decay kinetics. This simplification ignores the potential effects on long-term soil organic matter decay due to vertical and temporal dynamics of the microbial community and its adaptations in community structure and activity to changes in litter quality [26]. Therefore, much of the intricate complexity of soil processes, such as lagged responses of microbes and plants to sudden changes in N availability, is not captured. The potential of these more complex approaches remains to be fully explored, but their applicability at global scales has been recently demonstrated [14]. A systematic comparison as to the effect of this enhanced representation is urgently needed. The current generation of C–N models treats the vertical distribution of soil C and N at best implicitly, thereby not allowing for an assessment of the differential response of soil organic matter decomposition to temperature and moisture changes with soil depth [27]. Plant N uptake modelling primarily determines the competitiveness of plants for N compared with soil microbes, which is typically only treated implicitly as the outcome of soil organic matter decomposition [17]. This competition is usually not explicitly described, but an emergent property of the time-evolution of the seasonal and longterm changes net mineralisation from soil organic matter decay and plant N demand, rendering comparisons between model concepts difficult. Yet it is the most critical component of a C–N model, as the plant–microbe competition determines the coupled C–N model’s long-term dynamics and response to perturbations. Despite a general understanding of the processes involved, there are only few observations to constrain this interaction at the ecosystem level. One potential way forward is to use detailed studies of N fertilisation effects including the fate of added Nr isotope tracers [28]. Denitrification

N2 from denitrification is thought to be the largest loss term of nitrogen from terrestrial ecosystems [2]. Quantifying this loss term is not only important to understand the global terrestrial nitrogen budget (and therefore also the carbon budget) but also because nitrification and denitrification are associated with the production of NOx and the greenhouse gas N2O. Most models accounting for nitrification and denitrification processes go back to the hole-in-the-pipe idea [29]. Boyer et al. [30] provide an excellent review on the process-based modelling approaches at a site-level scale that have been developed since the early 1990s. Early global carbon–nitrogen cycle models ignored denitrification or subsumed it into a www.sciencedirect.com

generic N loss term [15,18], a practice still employed by some of the recent models [10,13]. The earliest global scale N cycle models to simulate nitrification–denitrification related N emissions assumed that gaseous N losses were proportional to gross soil N mineralisation, with the fraction of N2O and N2 produced depending on soil moisture [31,32]. Some of the current global carbon– nitrogen cycle models simulate explicitly nitrogen leaching and denitrification either as a fraction of gross mineralisation or mineral soil N [6,9]. Only few attempts have been made to mimic the behaviour of site-scale models at larger scales to estimate nitrification and denitrification rates [11,12]. The fundamental difference between these two latter approaches is that in the first gaseous N losses are coupled to nitrogen turnover, which causes ecosystems to constantly loose a fraction of its N capital, whereas in the second approach, in which loss rates depend on the actual mineral nitrogen concentrations of the soils, the loss rate may be substantially reduced in nutrient stressed ecosystems. This difference may turn out to be decisive for estimating the sign of the land N2O–climate interaction. The intricate complexity of trace gas production, relying on reliable representations of the time–space dynamics of soil redox potential, and labile carbon and mineral nitrogen availability makes modelling N trace gas emissions the most uncertain projection of carbon–nitrogen cycle models, which is unfortunately also only poorly constrained by suitable long-term observations.

Key results of the current generation of global carbon–nitrogen cycle models Implications of coupled C–N dynamics for the presentday terrestrial C cycle

Consistent with expectations [1], carbon–nitrogen cycle models show an attenuation of boreal and mid-latitude productivity due to limiting nitrogen availability. The models diverge with respect to the extent of nutrient limitation in the moist tropics. Some models suggest little N limitation, as high N fixation rates compensate for large N losses [10,21,22], while others suggest N limitation also in these ecosystems [7,33]. Although there are studies showing N limitation in some tropical forests [34], P limitation is thought to prevail in most of these systems [35]. N limitation may potentially be taken as a surrogate for P limitation in the case of the terrestrial response to CO2 fertilisation [33]. However, this argument fails when perturbations of the biogeochemical cycles of N and P themselves are considered. Wang et al. [13] have recently introduced a coupled C:N:P model that explicitly treats P dynamics, suggesting that the tropics would be generally more limited by P than by N, mainly due to the assumed distribution of plant available soil P. More work on quantifying the geographical changes in N:P limitation is needed. However, global modelling of P cycling effects on C is hampered by the lack of an adequate quantitative understanding of the spatial distribution of plant available P. Current Opinion in Environmental Sustainability 2011, 3:311–320

316 Carbon and nitrogen cycles

Table 2

Figure 2

Estimates of the terrestrial net carbon uptake [Pg C yrS1] in terrestrial ecosystems due to deposition of reactive nitrogen by the current generation of carbon–nitrogen cycle models.

Land Carbon-Cycle Climate feedbacks [W m-2 K-1] −2.0

Approach

Source

Reference period Net terrestrial C uptake

CLIMBER-CN ISAM-CN CLM-CN O-CN

[8] [22] [6,43] [21,38]

1990s 1990s 1976–2000 1990s

0.4 (0.6) a 0.26 0.2–0.3 0.2–0.5

−1.5

−1.0

−0.5

0.0

0.5

1.0

Carbon-concentration interaction Carbon-climate interaction

(a)

C4MIP ensemble CLM-CN

a

Value in brackets accounting for the synergistic interactions with atm. CO2 and climate.

IGSM-CN O-CN

Early analyses of the effect of the anthropogenic Nr perturbation on the terrestrial carbon balance have suggested a large range (0.2–1.2 Pg C yr1) of N deposition induced carbon storage [36]. The recent models, considering a multitude of model forcings and more complete representation of N cycle processes, generally suggest N sequestration at the lower end of this range (0.2–0.4 Pg C yr1) [6,8,21,22], located in accordance with the spatial patterns of nitrogen deposition mainly in Eastern North America, Central Europe and temperate/subtropical East Asia (Table 2). This range not only partly results from the above-mentioned differences in process representations embodied by the different models but potentially also from uncertainty in the magnitude and spatial patterns of nitrogen deposition [37]. Because of the non-linear synergistic effects between changes in carbon cycling and nitrogen availability, reported differences in the magnitude of the N deposition effect may also be due to the way the attribution to a particular forcing has been made [8,21]. Zaehle et al. [38] have recently quantified the radiative climate forcing of this effect in the year 2005 relative to pre-industrial times as 96  14 mW m2 based on propagating the simulated pre-industrial to present-day C fluxes from one particular model into atmospheric concentrations. Given the wider range of estimates on the effect from the entire model ensemble estimates to date, the effect is likely within the range of 96 to 192 mW m2. Changes in the N cycle also induce changes in biogeophysical parameters (evapotranspiration and turbulent exchanges), which themselves will also have an effect on climate. However, no study so far has explicitly integrated these effects into studies of the effect of N on climate. Effects of the natural nitrogen cycle on carbon cycle responses to global change

As a direct result of nitrogen limitation, all coupled carbon–nitrogen cycle models show a much lower CO2 fertilisation response, that is, a lower C sequestration as a result of increased plant productivity in a CO2 richer atmosphere, than their carbon only homologues [7,33,39] (see Figure 2a). This reduced CO2 fertilisation effect is already apparent for projected net land C uptake Current Opinion in Environmental Sustainability 2011, 3:311–320

(b)

C cycle C-N cycle C4MIP ensemble C-N model ensemble IGSM-CN O-CN

−2.0

−1.5

−1.0

−0.5

0.0

0.5

1.0

-2 -1 Land Carbon-Cycle Climate feedbacks [W m K ] Current Opinion in Environmental Sustainability

Climate feedbacks associated with the terrestrial carbon cycle and the effect of considering N dynamics evaluated using comprehensive coupled carbon–(nitrogen)-cycle climate models in the year 2100. (a) Effects due to CO2 fertilisation (carbon-concentration interaction) and climate change (carbon–climate interaction), compared between an ensemble of 11 carbon cycle only models [58], and the three projections from carbon–nitrogen cycle models available to date (CLM-CN [33], IGSM-CN [7] and O-CN [39]). (b) The net climate feedback from the terrestrial carbon cycle (sum of carbon-concentration and carbon– climate interaction), between the ensemble of carbon–climate and carbon–nitrogen-climate models. To illustrate the effect of N dynamics more clearly, the change net climatic feedbacks due to terrestrial carbon feedbacks are illustrated for IGSM-C(N) and O-C(N), for which both carbon only and carbon–nitrogen projections are available.

during the 20th century, but becomes more pronounced during the 21st century. Thornton et al. [6] have shown for their model that this reduction is not a result of the globally lower vegetation productivity simulated by C–N cycle models, but a consequence of the nitrogen dynamics acting on long-term carbon cycling. The N limitation on 21st century C sequestration is strongest in the boreal zone and decreases towards the temperate and tropical latitudes, but its magnitude and geographical distribution varies between the models. N deposition has only a small effect on these trends in future terrestrial carbon storage, when compared with the changes induced by changes in atmospheric CO2 and climate [6,21]. The second generic feature of C:N models is a reduced global carbon loss as a result of climate change [7,33,39] (Figure 2a): Resulting from temperature-enhanced soil organic matter decomposition, ecosystem N availability www.sciencedirect.com

Carbon–nitrogen interactions on land biogeochemistry Zaehle and Dalmonech 317

increases and thereby potentially enhances plant growth. Despite the rather small fraction of N recovered in vegetation, this may result in increased ecosystem C storage, as increased C sequestration in biomass compensates for the enhanced soil carbon loss because of the wider C:N ratio in vegetation when compared with soil organic matter. In two models this effect is sufficiently large to swap the net carbon balance effect of climate change from negative to positive until the year 2100, implying more terrestrial carbon storage with climate warming and thus a negative carbon–climate interaction [7,33]. However, at least for IGSM-CN the risk of a positive carbon–climate interaction beyond 2100 is not eliminated at high rates of climate change [7]. One other CN-model still shows a—albeit reduced—positive carbon–climate interaction as its carbon cycle only homologue (Figure 2a). These findings are generally consistent with the results of CO2 fertilisation and soil warming experiments that have demonstrated a strong interaction between N availability and vegetation growth [40,41]. Only few models have demonstrated the degree of agreement with the mean responses of free-air CO2 enrichment experiments based on generic model experiments [10,39,42]. Only one of these studies has provided a detailed assessment against ecosystem manipulation experiments both with respect to elevated CO2 and soil warming and demonstrated a reasonable agreement [39]. Irrespective of the magnitude of nitrogen limitations on the carbon-concentration and carbon–climate interaction, the simulated atmospheric CO2 concentrations in all coupled carbon–nitrogen cycle climate simulations suggested that the N effect on the carbon-concentration interaction dominates the response for the next century, implying higher projected CO2 concentrations (and thus rates of climate change) than projected with the C-only homologues [5,7,33,39,43] (see Figure 2b). Nitrous oxide and climate

Two carbon–nitrogen cycle models accounting for nitriprocesses have recently fication–denitrification attempted to estimate the terrestrial N2O source [12,38]. Both models generate spatial patterns of N2O emissions, and global levels of terrestrial N2O emissions (6.7–8.3 Tg N2O-N yr1) generally commensurate with current understanding (Xu-Ri, personal communication). One of these models has been applied to analyse trends in terrestrial N2O emissions from pre-industrial times to present day. The simulated changes in terrestrial N2O correspond to a change in radiative forcing (for the year 2005) of about 16  5 (19  9) mW m2 for climate change (CO2 fertilisation) [38]. The negative effect from CO2 fertilisation due to the increased N demand from plants requires further investigation, as it is conflicting with the results of a recent meta-analysis of ecosystem manipulation experiments suggesting a small positive www.sciencedirect.com

effect [44]. There are yet no global projections of terrestrial N2O emission changes under future climate changes from these models, but if current trends continue, a slightly positive interaction between terrestrial N2O emissions and climate might be expected, in accordance with suggestions from ice-core records [45]. In accordance with budget-oriented methods [46,47] C–N model projections suggest, however, that fertiliser and manure additions dominate the trends in terrestrial N2O emissions, and are responsible for most of the positive radiative forcing (in the year 2005) to the climate system (125  20 mW m2) [38].

Key challenges The undoubtedly greatest challenge to the development of reliable global models of the coupled terrestrial carbon–nitrogen cycles is the intrinsic difficulty in evaluating these complex and non-linear models, and thus identifying an adequate parameterisation of the important C–N processes. The extent of nitrogen limitation is quantified differently in the models. This makes it difficult to compare the simulated levels of N limitation on land carbon cycling, unless, for instance, common model experiments are performed that manipulate N availability. A further complicating issue is the high number of degrees of freedom in a modelling system with two or more coupled element cycles and with large e-folding times due to the recalcitrant soil pools, which strongly influence long-term dynamics, requiring observations at multiple time-scales to ascertain model behaviour. Lastly, the difficulty in evaluation of these models resides in the high degree of N recycling in N-deficient ecosystems and the importance of small and sporadic N losses, such as that of NOx and N2O; the observational difficulties in measuring important N fluxes at ecosystem or higher scales; and the large background concentration of N2, swamping signals from terrestrial emissions. Given the scarcity of relevant global data on terrestrial nitrogen cycling, current practice—though regrettably not exploited rigorously and to a full extent—is to infer the validity of carbon–nitrogen models from their capacity to reproduce local and global features of the (much more ‘easily’ and better observed) global carbon cycle [13,21,42]. These comparisons are most useful, when done with the carbon and carbon–nitrogen cycle version of the model to attribute model-data mismatches to imperfections in the representation of either cycle [21,43]. Key data sources include observed productivities at monitoring stations, spatial gradients in vegetation greenness and the atmospheric CO2 measurement network, specifically the interhemispheric gradient in the seasonal amplitude of net C exchanges, and the interannual variability in the atmospheric growth rates. These contain information on the spatial distribution of global terrestrial productivity and the short-term climate sensitivity of the global carbon cycle. Unlike for CO2 and methane, atmospheric N Current Opinion in Environmental Sustainability 2011, 3:311–320

318 Carbon and nitrogen cycles

observations offer little constraint on the climate sensitivity of terrestrial nitrogen cycle because variations in the stratospheric destruction of N2O dominate the seasonal and likely also interannual variability [48], and the short-lived nature of other atmospheric Nr. There is an urgent need to include alternative new data streams that provide indicators of nitrogen cycling and nitrogen stress, such as observed variations of foliar nitrogen content, giving estimates on the nitrogen available for photosynthesis [49,50], estimates of the large scale export of N from ecosystems through riverine transport [51], or the difference in the isotopic signatures of soil and vegetation N [52], both indicators of the N recycling and loss in/from ecosystems. However, the use of these data is not straightforward: Methodological issues hampering the comparability of model and observations need to be addressed. A stronger check on the validity of the carbon–nitrogen cycle coupling is to perform dedicated model perturbation experiments. Most can be learned from detailed studies at sites for which many ecosystem characteristics are known. However, vagaries in such comparisons are numerous, including: 1. The difficulty to disentangle the effects of the initial conditions of a particular ecosystem from the generic ecosystem response. In particular, local factors such as the site history determine the initial site conditions that may impact on the results, hampering their potential to identify failures in global models. 2. The short time-scale of many ecosystem manipulation experiments, since the extent of the (simulated) nutrient limitation depends on the time-scale at which the effect is evaluated [7]. Furthermore, step increases induce rapid shifts in nutrient demand and availability, whereas global changes are occurring gradually. 3. The unknown representativeness of the available ecosystem manipulation experiments for regional scale responses, most notably resulting from the lack of information from tropical ecosystems [53]. One way forward is to subject global models to perturbations in the form of sensitivity studies [39,42] and evaluate the simulated effects based on meta-analyses [20,54–56]. There is a double dividend of performing such a model-data confrontation systematically with the majority of the existing carbon–nitrogen cycle models: The first is a gain in understanding of the plausibility and response range of C:N model projections, the second the identification of geographical regions in which new experimental data or better process-understanding would help most to reduce projected uncertainties.

two important effects to be taken account of: The first is the direct effect of Nr in the atmosphere on radiative forcing through its impact on tropospheric aerosol and ozone. Secondly, increased tropospheric ozone burden has a detrimental effect on plant (and human) health, and therefore likely reduces the growth and carbon storage capacity of terrestrial ecosystems, that is, reducing the current climate cooling due to terrestrial carbon sequestration [57].

Summary Despite a common framework, the current generation of coupled land carbon-nitrogen cycle models differs in important details on how C–N linkages and in particular, N losses and trace gas emissions are represented. The strength of these models lies in modelling the C–N coupling effects rather than trace gas emissions. However, first estimates of terrestrial N2O fluxes generally confer with observational understanding of the space–time trends over recent decades. Irrespective of the model differences, a range of common qualitative features of these models emerge: 1. N deposition contributes to a moderate extent to the current global carbon uptake, with likely diminishing importance in the future. 2. N limitation strongly reduces C sequestration in boreal and temperate ecosystems, while model projections diverge in tropical regions. All model simulations to date suggest that this carbon-concentration effect dominates the overall importance of the C–N linkages. In conclusion, accounting for C–N interactions leads to a more rapid build-up of CO2 in the atmosphere under future scenarios, and hence higher rates of climate change, that projected with C cycle only models. These trends are further modified depending on whether ecosystem-scale biological N fixation is assumed to increase with plant N demand or not. 3. Current trends in N2O emissions are dominated by the effects of anthropogenic Nr input to the terrestrial biosphere. However, there is a potential for a moderate positive climate feedback. 4. Future challenges are the systematic use of ecosystem manipulation experiments to evaluate the responses of the coupled carbon–nitrogen cycle models to perturbations in climate, atmospheric CO2 concentrations, land-use and N additions; the quantification of the biogeophysical consequences of N cycling effects on present-day climate and future climate change projection; and the integration of missing components of N effects related to reactive N exchanges with the atmosphere and the effects of atmospheric pollution.

Acknowledgements An aspect currently not covered by models of the terrestrial nitrogen cycle is the climate effect of anthropogenic Nr through its effect on atmospheric chemistry. There are Current Opinion in Environmental Sustainability 2011, 3:311–320

SZ was supported by the Marie Curie Reintegration Grant JULIA (PERG02-GA-2007-224775), the NitroEurope-IP (grant agreement no. 017841) and SZ and DD by the European Community’s Seventh Framework Programme under grant agreement no. 238366. www.sciencedirect.com

Carbon–nitrogen interactions on land biogeochemistry Zaehle and Dalmonech 319

References and recommended reading Papers of particular interest, published within the period of review, have been highlighted as:  of special interest  of outstanding interest

1.

Vitousek PM, Howarth RW: Nitrogen limitation on land and in the sea — how it can occur. Biogeochemistry 1991, 13:87-115.

2. 

Galloway JN, Dentener FJ, Capone DG, Boyer EW, Howarth RW, Seitzinger SP, Asner GP, Cleveland C, Green P, Holland E et al.: Nitrogen cycles: past, present and future. Biogeochemistry 2004, 70(2):153-226. The most comprehensive review of global nitrogen fluxes to date. A rather tough read due to all the information contained. Some flux estimates have been updated since, quoted in this paper.

3. Gruber N, Galloway JN: An Earth-system perspective of the  global nitrogen cycle. Nature 2008, 451(7176):293-296. A very well-written overview on the importance of the global N cycle in the Earth system. 4.

Denman KL, Brasseur G, Chidthaisong A, Ciais P, Cox PM, Dickinson RE, Hauglustaine D, Heinze C, Holland E, Jacob D et al.: Couplings Between Changes in the Climate System and Biogeochemistry. In Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Edited by Solomon S, D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.Tignor, Miller HL: Cambridge, United Kingdom and New York, NY, USA: Cambridge University Press; 2007:511–539.

5. 

Arneth A, Harrison SP, Zaehle S, Tsigaridis K, Menon S, Bartlein PJ, Feichte J, Korhola A, Kulmala M, O’Donnell D et al.: Terrestrial biogeochemical feedbacks in the climate system. Nat Geosci 2010, 3(8):525-532 doi: 10.1038/ngeo905. A concise overview on the current state of modelling feedbacks between terrestrial biogeochemistry and climate.

6. 

Thornton PE, Lamarque JF, Rosenbloom NA, Mahowald NM: Influence of carbon–nitrogen cycle coupling on land model response to CO2 fertilization and climate variability. Glob Biogeochem Cycles 2007, 21: doi: 10.1029/2006GB002868. The first paper to directly assess the consequences of carbon–nitrogen coupling on carbon-cycle climate feedbacks due to the effect of N limitation on terrestrial carbon storage.

7.

Sokolov AP, Kicklighter DW, Melillo JM, Felzer BS, Schlosser CA, Cronin TW: Consequences of considering carbon–nitrogen interactions on the feedbacks between climate and the terrestrial carbon cycle. J Clim 2008, 21(15):3776-3796.

8.

Churkina G, Brovkin V, von Bloh W, Trusilova K, Jung M, Dentener F: Synergy of rising nitrogen depositions and atmospheric CO2 on land carbon uptake moderately offsets global warming. Glob Biogeochem Cycles 2009, 23(4): GB4027.

9.

Yang X, Wittig V, Jain A, Post W: The integration of nitrogen cycle dynamics into the Integrated Science Assessment Model (ISAM) for the study of terrestrial ecosystem responses to global change. Glob Biogeochem Cycles 2009 doi: 10.1029/ 2009GB003474.

10. Gerber S, Hedin LO, Oppenheimer M, Pacala SW, Shevliakova E: Nitrogen cycling and feedbacks in a global dynamic land model. Glob Biogeochem Cycles 2010 doi: 10.1029/ 2008GB003336. 11. Zaehle S, Friend AD: Carbon and nitrogen cycle dynamics in the O-CN land surface model: 1. Model description, site-scale evaluation and sensitivity to parameter estimates. Glob Biogeochem Cycles 2010, 24:GB1005 doi: 10.1029/ 2009GB003521. 12. Xu-Ri, Prentice IC: Terrestrial nitrogen cycle simulation with a dynamic global vegetation model. Glob Change Biol 2008, 14: doi: 10.1111/j.1365-2486.2008.01625.x. 13. Wang YP, Law RM, Pak B: A global model of carbon, nitrogen and phosphorus cycles for the terrestrial biosphere. Biogeosciences 2010, 7(7):2261-2282 doi: 10.5194/bg-7-2261-2010. www.sciencedirect.com

14. Esser G, Kattge J, Sakalli A: Feedback of carbon and nitrogen cycles enhances carbon sequestration in the terrestrial biosphere. Glob Change Biol 2011, 17(2):819-842. 15. Melillo JM, McGuire AD, Kicklighter DW, Moore B, Vorosmarty CJ, Schloss AL: Global climate-change and terrestrial net primary production. Nature 1993, 363(6426):234-240. 16. Running SW, Hunt ER: Generalization of a forest ecosystem process model for other biomes, BIOME-BGC, and an application for global-scale models. In Scaling Physiological Processes: Leaf to Globe. Edited by Ehleringer JR, Field CB. New York: Academic Press, Inc.; 1993:141-158. 17. Parton WJ, Scurlock JMO, Ojima DS, Gilmanov TG, Scholes RJ, Schimmel DS, Kirchner T, Menaut JC, Seastedt T, Moya EG et al.: Observations and modelling of biomass and soil organic matter dynamics for the grassland biome worldwide. Glob Biogeochem Cycles 1993, 7(4):785-809. 18. Friend AD, Stevens AK, Knox RG, Cannell MGR: A processbased, terrestrial biosphere model of ecosystem dynamics (Hybrid v3.0). Ecol Model 1997, 95:249-287. 19. Li CS, Aber J, Stange F, Butterbach-Bahl K, Papen H: A process-oriented model of N2O and NO emissions from forest soils: 1. Model development. J Geophys Res Atmos 2000, 105:4369-4384. 20. Norby RJ, DeLucia EH, Gielen B, Calfapietra C, Giardina CP, King JS, Ledford J, McCarthy HR, Moore DJP, Ceulemans R et al.: Forest response to elevated CO2 is conserved across a broad range of productivity. Proc Natl Acad Sci U S A 2005, 102:18052-18056. 21. Zaehle S, Friend AD, Dentener F, Friedlingstein P, Peylin P, Schulz M: Carbon and nitrogen cycle dynamics in the O-CN land surface model: 2. The role of the nitrogen cycle in the historical terrestrial carbon balance. Glob Biogeochem Cycles 2010, 24:GB1006 doi: 10.1029/2009GB003522. 22. Jain A, Yang X, Kheshgi H, McGuire AD, Post W, Kicklighter D: Nitrogen attenuation of terrestrial carbon cycle response to global environmental factors. Glob Biogeochem Cycles 2009, 23:GB4028 doi: 10.1029/2009GB003519. 23. Fisher JB, Sitch S, Malhi Y, Fisher RA, Huntingford C, Tan SY: Carbon cost of plant nitrogen acquisition: a mechanistic, globally applicable model of plant nitrogen uptake, retranslocation, and fixation. Glob Biogeochem Cycles 2010, 24:. 24. Wang YP, Houlton BZ: Nitrogen constraints on terrestrial carbon update: implications for the global carbon-climate  feedback. Geophys Res Lett 2009, 36:L24403 doi: 10.1029/ 2009GL041009. A first assessment of the effect of varying biological N fixation on terrestrial carbon cycle projections, but unfortunately not assessed within the framework of a coupled carbon–nitrogen cycle model. 25. Na¨sholm T, Kielland K, Ganeteg U: Uptake of organic nitrogen by plants. New Phytol 2009, 182:31-48. 26. Schimel J, Bennett J: Nitrogen mineralization: challenges of a changing paradigm. Ecology 2004, 85(3):591-602. 27. Koven C, Friedlingstein P, Ciais P, Khvorostyanov D, Krinner G, Tarnocai C: On the formation of high-latitude soil carbon stocks: Effects of cryoturbation and insulation by organic matter in a land surface model. Geophys Res Lett 2009, 36:pL21501. 28. Nadelhoffer KJ, Colman BP, Currie WS, Magill A, Aber JD: Decadal-scale fates of N-15 tracers added to oak and pine stands under ambient and elevated N inputs at the Harvard Forest (USA). Forest Ecol Manage 2004, 196(1):89-107. 29. Firestone MK, Davidson EA: Microbial basis of NO and N2O production and consumption in soils. In Exchange of Trace Gases Between Terrestrial Ecosystems and the Atmosphere. Edited by Andreae MO, Schimel DS. New York, USA: John Wiley; 1989:7-21. 30. Boyer EW, Alexander RB, Parton WJ, Li CS, Butterbach-Bahl K, Donner SD, Skaggs RW, Del Gross SJ: Modeling denitrification  in terrestrial and aquatic ecosystems at regional scales. Ecol Appl 2006, 16(6):2123-2142. Current Opinion in Environmental Sustainability 2011, 3:311–320

320 Carbon and nitrogen cycles

A very comprehensive review on existing denitrification models—understand that still needs to find its way into most of the global carbon– nitrogen cycle models. 31. Potter CS, Matson PA, Vitousek PM, Davidson EA: Process modeling of controls on nitrogen trace gas emissions from soils worldwide. J Geophys Res Atmos 1996, 101(D1):1361-1377. 32. Nevison CD, Esser G, Holland EA: A global model of changing N2O emissions from natural and perturbed soils. Clim Change 1996, 32(3):327-378. 33. Thornton PE, Doney SC, Lindsay K, Moore JK, Mahowald N, Randerson JT, Fung I, Lamarque J-F, Feddema JJ, Lee Y-H: Carbon–nitrogen interactions regulate climate–carbon cycle feedbacks: results from an atmosphere–ocean general circulation model. Biogeosciences 2009, 6:2099-2120. 34. LeBauer DS, Treseder KK: Nitrogen limitation of net primary productivity in terrestrial ecosystems is globally distributed. Ecology 2008, 89(2):371-379. 35. Townsend AR, Cleveland CC, Houlton BZ, Alden CB, White JWC: Multi-element regulation of the tropical forest carbon cycle. Front Ecol Environ 2011, 9(1):9-17.

44. van Groenigen KJ, Osenberg C, Hungate B: Increased soil emissions of potent greenhouse gases under increased atmospheric CO2. Nature 2011, 475:214-216 doi: 10.1038/ nature10176. 45. Flu¨ckiger J, Da¨llenbach A, Blunier T, Stauffer B, Stocker TF, Raynaud D, Barnola JM: Variations in atmospheric N2O concentration during abrupt climatic changes. Science 1999, 285(5425):227-230. 46. Davidson EA: The contribution of manure and fertilizer nitrogen to atmospheric nitrous oxide since 1860. Nat Geosci 2009, 2(9):659-662. 47. Syakila A, Kroeze C: The global nitrous oxide budget revisited. Greenhouse Gas Meas Manage 2011, 1:17-26. 48. Nevison CD, Mahowald NM, Weiss RF, Prinn RG: Interannual and seasonal variability in atmospheric N2O. Glob Biogeochem Cycles 2007, 21(3):pGB3017. 49. Kattge J, Wirth C, Prentice IC, Rammig A, Fisher RA, Zaehle S: TRY — a global database of plant traits. Glob Change Biol 2011, 17(9):2905-2935 doi: 10.1111/j.1365-2486.2011.02451.x.

36. Townsend A, Braswell BH, Holland E, Penner J: Spatial and temporal patterns in potential terrestrial carbon storage due to deposition of fossil fuel derived nitrogen. Ecol Appl 1996, 6:806-814.

50. Ollinger S, Frolking S, Richardson A, Martin M, Hollinger D, Reich P, Plourde L: Nitrogen-albedo relationship in forests remains robust and thought-provoking Reply. Proc Natl Acad Sci U S A 2009, 106:E17.

37. Dentener F, Drevet J, Lamarque JF, Bey I, Eickhout B, Fiore AM, Hauglustaine D, Horowitz LW, Krol M, Kulshrestha UC et al.: Nitrogen and sulfur deposition on regional and global scales: a multimodel evaluation. Glob Biogeochem Cycles 2006, 20(4) doi: 10.1029/2005GB002672.

51. Seitzinger SP, Kroeze C: Global distribution of nitrous oxide production and N inputs in freshwater and coastal marine ecosystems. Glob Biogeochem Cycles 1998, 12(1): 93-113.

38. Zaehle S, Ciais P, Friend AD, Prieur V: Carbon benefits of  anthropogenic reactive nitrogen offset by nitrous oxide emissions. Nat Geosci 2011 doi: 10.1038/NGEO1207. A first attempt to assess the consequences of anthropogenic Nr addition on terrestrial greenhouse fluxes using a coupled carbon–nitrogen cycle model and a consistent set of historic model driver data. 39. Zaehle S, Friedlingstein P, Friend A: Terrestrial nitrogen  feedbacks may accelerate future climate change. Geophys Res Lett 2010, 37:L01401 doi: 10.1029/2009GL041345. A study bringing together the results of ecosystem manipulation and the projections of a coupled carbon–nitrogen cycle model to assess the carbon-climate feedbacks under nitrogen constraints. 40. Melillo JM, Butler S, Johnson J, Mohan J, Steudler P, Lux H, Burrows E, Bowles F, Smith R, Scott L et al.: Soil warming, carbon–nitrogen interactions, and forest carbon budgets. Proc Natl Acad Sci U S A 2011, 108:9508-9512. 41. Norby RJ, Warren JM, Iversen CM, Medlyn BE, McMurtrie RE: CO2 enhancement of forest productivity constrained by limited nitrogen availability. Proc Natl Acad Sci U S A 2009, 107:19368-19373. 42. Randerson JT, Hoffman FM, Thornton PE, Mahowald NM, Lindsay K, Lee YH, Nevison CD, Doney SC, Bonan G, Stockli R et al.: Systematic assessment of terrestrial biogeochemistry in coupled climate–carbon models. Glob Change Biol 2009, 15(10):2462-2484. 43. Bonan GB, Levis S: Quantifying carbon–nitrogen feedbacks in the Community Land Model (CLM4). Geophys Res Lett 2010, 37:.

Current Opinion in Environmental Sustainability 2011, 3:311–320

52. Amundson R, Austin AT, Schuur EAG, Yoo K, Matzek V, Kendall C, Uebersax A, Brenner D, Baisden WT: Global patterns of the isotopic composition of soil and plant nitrogen. Glob Biogeochem Cycles 2003, 17(1). 53. Hickler T, Smith B, Prentice IC, Mjofors K, Miller P, Arneth A, Sykes MT: CO2 fertilization in temperate FACE experiments not representative of boreal and tropical forests. Glob Change Biol 2008, 14(7):1531-1542. 54. Rustad LE, Campbell JL, Marion GM, Norby RJ, Mitchell MJ, Hartley AE, Cornelissen JHC, Gurevitch J: A meta-analysis of the response of soil respiration, net nitrogen mineralization, and aboveground plant growth to experimental ecosystem warming. Oecologia 2001, 126(4):543-562. 55. Sutton MA, Simpson D, Levy PE, Smith RI, Reis S, van Oijen M, de Vries W: Uncertainties in the relationship between atmospheric nitrogen deposition and forest carbon sequestration. Glob Change Biol 2008, 14:2057-2063. 56. Thomas RQ, Canham CD, Weathers KC, Goodale CL: Increased tree carbon storage in response to nitrogen deposition in the US. Nat Geosci 2010, 3(1):13-17. 57. Sitch S, Cox PM, Collins WJ, Huntingford C: Indirect radiative forcing of climate change through ozone effects on the landcarbon sink. Nature 2007, 448(7155):791-794. 58. Friedlingstein P, Cox P, Betts R, Bopp L, Von Bloh W, Brovkin V, Cadule P, Doney S, Eby M, Fung I et al.: Climate–carbon cycle feedback analysis: results from the C4MIP model intercomparison. J Clim 2006, 19(14):3337-3353.

www.sciencedirect.com