ALBERT

Description: Carbon–nitrogen (C–N) interactions regulate N availability for plant growth and for emissions of nitrous oxide (N2O) and the uptake of carbon dioxide. Future projections of these terrestrial greenhouse gas fluxes are strikingly divergent, leading to major uncertainties in projected global warming. Here we analyse the large increase in terrestrial N2O emissions over the past 21 000 years as reconstructed from ice-core isotopic data and presented in part 1 of this study. Remarkably, the increase occurred in two steps, each realized over decades and within a maximum of 2 centuries, at the onsets of the major deglacial Northern Hemisphere warming events. The data suggest a highly dynamic and responsive global N cycle. The increase may be explained by an increase in the flux of reactive N entering and leaving ecosystems or by an increase in N2O yield per unit N converted. We applied the LPX-Bern dynamic global vegetation model in deglacial simulations forced with Earth system model climate data to investigate N2O emission patterns, mechanisms, and C–N coupling. The N2O emission changes are mainly attributed to changes in temperature and precipitation and the loss of land due to sea-level rise. LPX-Bern simulates a deglacial increase in N2O emissions but underestimates the reconstructed increase by 47 %. Assuming time-independent N sources in the model to mimic progressive N limitation of plant growth results in a decrease in N2O emissions in contrast to the reconstruction. Our results appear consistent with suggestions of (a) biological controls on ecosystem N acquisition and (b) flexibility in the coupling of the C and N cycles during periods of rapid environmental change. A dominant uncertainty in the explanation of the reconstructed N2O emissions is the poorly known N2O yield per N lost through gaseous pathways and its sensitivity to soil conditions. The deglacial N2O record provides a constraint for future studies.

Description: We present a new global reconstruction of seasonal climates at the Last Glacial Maximum (LGM, 21 000 years BP) made using 3-D variational data assimilation with pollen-based site reconstructions of six climate variables and the ensemble average of the PMIP3—CMIP5 simulations as a prior (initial estimate of LGM climate). We assume that the correlation matrix of the uncertainties in the prior is both spatially and temporally Gaussian, in order to produce a climate reconstruction that is smoothed both from month to month and from grid cell to grid cell. The pollen-based reconstructions include mean annual temperature (MAT), mean temperature of the coldest month (MTCO), mean temperature of the warmest month (MTWA), growing season warmth as measured by growing degree days above a baseline of 5 ∘C (GDD5), mean annual precipitation (MAP), and a moisture index (MI), which is the ratio of MAP to mean annual potential evapotranspiration. Different variables are reconstructed at different sites, but our approach both preserves seasonal relationships and allows a more complete set of seasonal climate variables to be derived at each location. We further account for the ecophysiological effects of low atmospheric carbon dioxide concentration on vegetation in making reconstructions of MAP and MI. This adjustment results in the reconstruction of wetter climates than might otherwise be inferred from the vegetation composition. Finally, by comparing the uncertainty contribution to the final reconstruction, we provide confidence intervals on these reconstructions and delimit geographical regions for which the palaeodata provide no information to constrain the climate reconstructions. The new reconstructions will provide a benchmark created using clear and defined mathematical procedures that can be used for evaluation of the PMIP4–CMIP6 entry-card LGM simulations and are available at https://doi.org/10.17864/1947.244 (Cleator et al., 2020b).

Description: Global fire-vegetation models are widely used to assess impacts of environmental change on fire regimes and the carbon cycle and to infer relationships between climate, land use and fire. However, differences in model structure and parameterizations, in both the vegetation and fire components of these models, could influence overall model performance, and to date there has been limited evaluation of how well different models represent various aspects of fire regimes. The Fire Model Intercomparison Project (FireMIP) is coordinating the evaluation of state-of-the-art global fire models, in order to improve projections of fire characteristics and fire impacts on ecosystems and human societies in the context of global environmental change. Here we perform a systematic evaluation of historical simulations made by nine FireMIP models to quantify their ability to reproduce a range of fire and vegetation benchmarks. The FireMIP models simulate a wide range in global annual total burnt area (39–536 Mha) and global annual fire carbon emission (0.91–4.75 Pg C yr−1) for modern conditions (2002–2012), but most of the range in burnt area is within observational uncertainty (345–468 Mha). Benchmarking scores indicate that seven out of nine FireMIP models are able to represent the spatial pattern in burnt area. The models also reproduce the seasonality in burnt area reasonably well but struggle to simulate fire season length and are largely unable to represent interannual variations in burnt area. However, models that represent cropland fires see improved simulation of fire seasonality in the Northern Hemisphere. The three FireMIP models which explicitly simulate individual fires are able to reproduce the spatial pattern in number of fires, but fire sizes are too small in key regions, and this results in an underestimation of burnt area. The correct representation of spatial and seasonal patterns in vegetation appears to correlate with a better representation of burnt area. The two older fire models included in the FireMIP ensemble (LPJ–GUESS–GlobFIRM, MC2) clearly perform less well globally than other models, but it is difficult to distinguish between the remaining ensemble members; some of these models are better at representing certain aspects of the fire regime; none clearly outperforms all other models across the full range of variables assessed.

Description: Nitrogen content per unit leaf area (Narea) is a key variable in plant functional ecology and biogeochemistry. Narea comprises a structural component, which scales with leaf mass per area (LMA), and a metabolic component, which scales with Rubisco capacity. The co-ordination hypothesis, as implemented in LPJ and related global vegetation models, predicts that Rubisco capacity should be directly proportional to irradiance but should decrease with increases in ci : ca and temperature because the amount of Rubisco required to achieve a given assimilation rate declines with increases in both. We tested these predictions using LMA, leaf δ13C, and leaf N measurements on complete species assemblages sampled at sites on a north–south transect from tropical to temperate Australia. Partial effects of mean canopy irradiance, mean annual temperature, and ci : ca (from δ13C) on Narea were all significant and their directions and magnitudes were in line with predictions. Over 80 % of the variance in community-mean (ln) Narea was accounted for by these predictors plus LMA. Moreover, Narea could be decomposed into two components, one proportional to LMA (slightly steeper in N-fixers), and the other to Rubisco capacity as predicted by the co-ordination hypothesis. Trait gradient analysis revealed ci : ca to be perfectly plastic, while species turnover contributed about half the variation in LMA and Narea. Interest has surged in methods to predict continuous leaf-trait variation from environmental factors, in order to improve ecosystem models. Coupled carbon–nitrogen models require a method to predict Narea that is more realistic than the widespread assumptions that Narea is proportional to photosynthetic capacity, and/or that Narea (and photosynthetic capacity) are determined by N supply from the soil. Our results indicate that Narea has a useful degree of predictability, from a combination of LMA and ci : ca – themselves in part environmentally determined – with Rubisco activity, as predicted from local growing conditions. This finding is consistent with a plant-centred approach to modelling, emphasizing the adaptive regulation of traits. Models that account for biodiversity will also need to partition community-level trait variation into components due to phenotypic plasticity and/or genotypic differentiation within species vs. progressive species replacement, along environmental gradients. Our analysis suggests that variation in Narea is about evenly split between these two modes.

Description: Continual input of reactive nitrogen (N) is required to support the natural turnover of N in terrestrial ecosystems. This N demand can be satisfied in various ways, including biological N fixation (BNF) (the dominant pathway under natural conditions), lightning-induced abiotic N fixation, N uptake from sedimentary substrates, and N deposition from natural and anthropogenic sources. We estimated the global new N fixation demand (NNF), i.e. the total new N input required to sustain net primary production (NPP) in non-agricultural terrestrial ecosystems regardless of its origin, using a N-enabled global dynamic vegetation model (DyN-LPJ). DyN-LPJ does not explicitly simulate BNF; rather, it estimates total NNF using a mass balance criterion and assumes that this demand is met from one source or another. The model was run in steady state and then in transient mode driven by recent changes in CO2 concentration and climate. A range of values for key stoichiometric parameters was considered, based on recently published analyses. Modelled NPP and C : N ratios of litter and soil organic matter were consistent with independent estimates. Modelled geographic patterns of ecosystem NNF were similar to other analyses, but actual estimated values exceeded recent estimates of global BNF. The results were sensitive to a few key parameters: the fraction of litter carbon respired to CO2 during decomposition and plant-type-specific C : N ratios of litter and soil. The modelled annual NNF increased by about 15 % during the course of the transient run, mainly due to increasing CO2 concentration. The model did not overestimate recent terrestrial carbon uptake, suggesting that the increase in NNF demand has so far been met. Rising CO2 is further increasing the NNF demand, while the future capacity of N sources to support this is unknown.

Description: Recent advances in the representation of soil carbon decomposition and carbon–nitrogen interactions implemented previously into separate versions of the land surface scheme JSBACH are here combined in a single version, which is set to be used in the upcoming 6th phase of coupled model intercomparison project (CMIP6).Here we demonstrate that the new version of JSBACH is able to reproduce the spatial variability in the reactive nitrogen-loss pathways as derived from a compilation of δ15N data (R = 0. 76, root mean square error (RMSE)  = 0. 2, Taylor score  = 0. 83). The inclusion of carbon–nitrogen interactions leads to a moderate reduction (−10 %) of the carbon-concentration feedback (βL) and has a negligible effect on the sensitivity of the land carbon cycle to warming (γL) compared to the same version of the model without carbon–nitrogen interactions in idealized simulations (1 % increase in atmospheric carbon dioxide per year). In line with evidence from elevated carbon dioxide manipulation experiments, pronounced nitrogen scarcity is alleviated by (1) the accumulation of nitrogen due to enhanced nitrogen inputs by biological nitrogen fixation and reduced losses by leaching and volatilization. Warming stimulated turnover of organic nitrogen further counteracts scarcity.The strengths of the land carbon feedbacks of the recent version of JSBACH, with βL = 0. 61 Pg ppm−1 and γL = −27. 5 Pg °C−1, are 34 and 53 % less than the averages of CMIP5 models, although the CMIP5 version of JSBACH simulated βL and γL, which are 59 and 42 % higher than multi-model average. These changes are primarily due to the new decomposition model, indicating the importance of soil organic matter decomposition for land carbon feedbacks.

Description: Ecosystem models commonly assume that key photosynthetic traits, such as carboxylation capacity measured at a standard temperature, are constant in time. The temperature responses of modelled photosynthetic or respiratory rates then depend entirely on enzyme kinetics. Optimality considerations, however, suggest this assumption may be incorrect. The “coordination hypothesis” (that Rubisco- and electron-transport-limited rates of photosynthesis are co-limiting under typical daytime conditions) predicts, instead, that carboxylation (Vcmax) capacity should acclimate so that it increases somewhat with growth temperature but less steeply than its instantaneous response, implying that Vcmax when normalized to a standard temperature (e.g. 25 ∘C) should decline with growth temperature. With additional assumptions, similar predictions can be made for electron-transport capacity (Jmax) and mitochondrial respiration in the dark (Rdark). To explore these hypotheses, photosynthetic measurements were carried out on woody species during the warm and the cool seasons in the semi-arid Great Western Woodlands, Australia, under broadly similar light environments. A consistent proportionality between Vcmax and Jmax was found across species. Vcmax, Jmax and Rdark increased with temperature in most species, but their values standardized to 25 ∘C declined. The ci:ca ratio increased slightly with temperature. The leaf N : P ratio was lower in the warm season. The slopes of the relationships between log-transformed Vcmax and Jmax and temperature were close to values predicted by the coordination hypothesis but shallower than those predicted by enzyme kinetics.

Description: In recent years evidence has emerged that the amount of isoprene emitted from a leaf is affected by the CO2 growth environment. Many – though not all – laboratory experiments indicate that emissions increase significantly at below-ambient CO2 concentrations and decrease when concentrations are raised to above-ambient. A small number of process-based leaf isoprene emission models can reproduce this CO2 stimulation and inhibition. These models are briefly reviewed, and their performance in standard conditions compared with each other and to an empirical algorithm. One of the models was judged particularly useful for incorporation into a dynamic vegetation model framework, LPJ-GUESS, yielding a tool that allows the interactive effects of climate and increasing CO2 concentration on vegetation distribution, productivity, and leaf and ecosystem isoprene emissions to be explored. The coupled vegetation dynamics-isoprene model is described and used here in a mode particularly suited for the ecosystem scale, but it can be employed at the global level as well. Annual and/or daily isoprene emissions simulated by the model were evaluated against flux measurements (or model estimates that had previously been evaluated with flux data) from a wide range of environments, and agreement between modelled and simulated values was generally good. By using a dynamic vegetation model, effects of canopy composition, disturbance history, or trends in CO2 concentration can be assessed. We show here for five model test sites that the suggested CO2-inhibition of leaf-isoprene metabolism can be large enough to offset increases in emissions due to CO2-stimulation of vegetation productivity and leaf area growth. When effects of climate change are considered atop the effects of atmospheric composition the interactions between the relevant processes will become even more complex. The CO2-isoprene inhibition may have the potential to significantly dampen the expected steep increase of ecosystem isoprene emission in a future, warmer atmosphere with higher CO2 levels; this effect raises important questions for projections of future atmospheric chemistry, and its connection to the terrestrial vegetation and carbon cycle.

Description: We present a benchmark system for global vegetation models. This system provides a quantitative evaluation of multiple simulated vegetation properties, including primary production; seasonal net ecosystem production; vegetation cover, composition and height; fire regime; and runoff. The benchmarks are derived from remotely sensed gridded datasets and site-based observations. The datasets allow comparisons of annual average conditions and seasonal and inter-annual variability, and they allow the impact of spatial and temporal biases in means and variability to be assessed separately. Specifically designed metrics quantify model performance for each process, and are compared to scores based on the temporal or spatial mean value of the observations and a "random" model produced by bootstrap resampling of the observations. The benchmark system is applied to three models: a simple light-use efficiency and water-balance model (the Simple Diagnostic Biosphere Model: SDBM), and the Lund-Potsdam-Jena (LPJ) and Land Processes and eXchanges (LPX) dynamic global vegetation models (DGVMs). SDBM reproduces observed CO2 seasonal cycles, but its simulation of independent measurements of net primary production (NPP) is too high. The two DGVMs show little difference for most benchmarks (including the inter-annual variability in the growth rate and seasonal cycle of atmospheric CO2), but LPX represents burnt fraction demonstrably more accurately. Benchmarking also identified several weaknesses common to both DGVMs. The benchmarking system provides a quantitative approach for evaluating how adequately processes are represented in a model, identifying errors and biases, tracking improvements in performance through model development, and discriminating among models. Adoption of such a system would do much to improve confidence in terrestrial model predictions of climate change impacts and feedbacks.

Description: An extensive data set on net primary production (NPP) in China's forests is analysed with two semi-empirical models based on the light use efficiency (LUE) and water use efficiency (WUE) concepts, respectively. Results are shown to be broadly consistent with other data sets (grassland above-ground NPP; globally extrapolated gross primary production, GPP) and published analyses. But although both models describe the data about equally well, they predict notably different responses to [CO2] and temperature. These are illustrated by sensitivity tests in which [CO2] is kept constant or doubled, temperatures are kept constant or increased by 3.5 K, and precipitation is changed by ±10%. Precipitation changes elicit similar responses in both models. The [CO2] response of the WUE model is much larger but is probably an overestimate for dense vegetation as it assumes no increase in runoff; while the [CO2] response of the LUE model is probably too small for sparse vegetation as it assumes no increase in vegetation cover. In the LUE model warming reduces total NPP with the strongest effect in South China, where the growing season cannot be further extended. In the WUE model warming increases total NPP, again with the strongest effect in South China, where abundant water supply precludes stomatal closure. The qualitative differences between the two formulations illustrate potential causes of the large differences (even in sign) in the global NPP response of dynamic global vegetation models to [CO2] and climate change. As it is not clear which response is more realistic, the issue needs to be resolved by observation and experiment.