Terrestrial Biogeochemical Models Research Articles

Efficient spin-up of Earth System Models using sequence acceleration.

Marine and terrestrial biogeochemical models are key components of the Earth System Models (ESMs) used to project future environmental changes. However, their slow adjustment time also hinders effective use of ESMs because of the enormous computational resources required to integrate them to a pre-industrial equilibrium. Here, a solution to this "spin-up" problem based on "sequence acceleration", is shown to accelerate equilibration of state-of-the-art marine biogeochemical models by over an order of magnitude. The technique can be applied in a "black box" fashion to existing models. Even under the challenging spin-up protocols used for Intergovernmental Panel on Climate Change (IPCC) simulations, this algorithm is 5 times faster. Preliminary results suggest that terrestrial models can be similarly accelerated, enabling a quantification of major parametric uncertainties in ESMs, improved estimates of metrics such as climate sensitivity, and higher model resolution than currently feasible.

Uncertainty propagation in a global biogeochemical model driven by leaf area data

Satellite-observed leaf area index (LAI) is often used to depict vegetation canopy structure and photosynthesis processes in terrestrial biogeochemical models. However, it remains unclear how the uncertainty of LAI among different satellite products propagates to the modeling of carbon (C), nitrogen (N), and phosphorus (P) cycles. Here, we separately drive a global biogeochemical model by three satellite-derived LAI products (i.e., GIMMS LAI3g, GLASS, and GLOBMAP) from 1982 to 2011. Using a traceability analysis, we explored the propagation of LAI-driven uncertainty to modeled C, N, and P storage among different biomes. The results showed that the data uncertainty of LAI was more considerable in the tropics than in non-tropical regions, whereas the modeling uncertainty of C, N, and P stocks showed a contrasting biogeographic pattern. The spread of simulated C, N, and P storage derived by different LAI datasets resulted from assimilation rates of elements in shrubland and C3 grassland but from the element residence time (τ) in deciduous needle leaf forest and tundra regions. Moreover, the assimilation rates of elements are the main contributing factor, with 67.6, 93.2, and 93% of vegetated grids for the modeled uncertainty of C, N, and P storage among the three simulations. We further traced the variations in τ to baseline residence times of different elements and the environmental scalars. These findings indicate that the data uncertainty of plant leaf traits can propagate to ecosystem processes in global biogeochemical models, especially in non-tropical forests.

Soil-related developments of the Biome-BGCMuSo v6.2 terrestrial ecosystem model

Abstract. Terrestrial biogeochemical models are essential tools to quantify climate–carbon cycle feedback and plant–soil relations from local to global scale. In this study, a theoretical basis is provided for the latest version of the Biome-BGCMuSo biogeochemical model (version 6.2). Biome-BGCMuSo is a branch of the original Biome-BGC model with a large number of developments and structural changes. Earlier model versions performed poorly in terms of soil water content (SWC) dynamics in different environments. Moreover, lack of detailed nitrogen cycle representation was a major limitation of the model. Since problems associated with these internal drivers might influence the final results and parameter estimation, additional structural improvements were necessary. In this paper the improved soil hydrology as well as the soil carbon and nitrogen cycle calculation methods are described in detail. Capabilities of the Biome-BGCMuSo v6.2 model are demonstrated via case studies focusing on soil hydrology, soil nitrogen cycle, and soil organic carbon content estimation. Soil-hydrology-related results are compared to observation data from an experimental lysimeter station. The results indicate improved performance for Biome-BGCMuSo v6.2 compared to v4.0 (explained variance increased from 0.121 to 0.8 for SWC and from 0.084 to 0.46 for soil evaporation; bias changed from −0.047 to −0.007 m3 m−3 for SWC and from −0.68 to −0.2 mm d−1 for soil evaporation). Simulations related to nitrogen balance and soil CO2 efflux were evaluated based on observations made in a long-term field experiment under crop rotation. The results indicated that the model is able to provide realistic nitrate content estimation for the topsoil. Soil nitrous oxide (N2O) efflux and soil respiration simulations were also realistic, with overall correspondence with the observations (for the N2O efflux simulation bias was between −0.13 and −0.1 mgNm-2d-1, and normalized root mean squared error (NRMSE) was 32.4 %–37.6 %; for CO2 efflux simulations bias was 0.04–0.17 gCm-2d-1, while NRMSE was 34.1 %–40.1 %). Sensitivity analysis and optimization of the decomposition scheme are presented to support practical application of the model. The improved version of Biome-BGCMuSo has the ability to provide more realistic soil hydrology representation as well as nitrification and denitrification process estimation, which represents a major milestone.

On the modeling paradigm of plant root nutrient acquisition

Plant root nutrient acquisition, and to a lesser extent foliar nutrient uptake, maintain plant metabolism and strongly regulate terrestrial biogeochemistry and carbon-climate feedbacks. However, terrestrial biogeochemical models differ in their representations of plant root nutrient acquisition, leading to significantly different, and uncertain, carbon cycle and future climate projections. Here we first review biogeochemical principles and observations relevant to three essential plant root nutrient acquisition mechanisms: activity of nutrient acquiring proteins, maintenance of nutrient stoichiometry, and energy expenditure for these processes. We next examine how these mechanisms are considered in three existing modeling paradigms, and conclude by recommending the capacity-based approach, the need for observations, and necessary modeling developments of plant root nutrient acquisition to improve carbon-climate feedback projections.

Spatial and temporal patterns of global soil heterotrophic respiration in terrestrial ecosystems

Abstract. Soil heterotrophic respiration (RH) is one of the largest and most uncertain components of the terrestrial carbon cycle, directly reflecting carbon loss from soils to the atmosphere. However, high variations and uncertainties of RH existing in global carbon cycling models require RH estimates from different angles, e.g., a data-driven angle. To fill this knowledge gap, this study applied a Random Forest (RF) algorithm (a machine learning approach) to (1) develop a globally gridded RH dataset and (2) investigate its spatial and temporal patterns from 1980 to 2016 at the global scale by linking field observations from the Global Soil Respiration Database and global environmental drivers (temperature, precipitation, soil water content, etc.). Finally, a globally gridded RH dataset was developed covering from 1980 to 2016 with a spatial resolution of half a degree and a temporal resolution of 1 year. Globally, the average annual RH was 57.2±0.6 Pg C a−1 from 1980 to 2016, with a significantly increasing trend of 0.036±0.007 Pg C a−2. However, the temporal trend of the carbon loss from RH varied in climate zones, and RH showed a significant and increasing trend in boreal and temperate areas. In contrast, such a trend was absent in tropical regions. Temperature-driven RH dominated 39 % of global land and was primarily distributed at high-latitude areas. The areas dominated by precipitation and soil water content were mainly semiarid and tropical areas, accounting for 36 % and 25 % of global land area, respectively, suggesting variations in the dominance of environmental controls on the spatial patterns of RH. The developed globally gridded RH dataset will further aid in the understanding of the mechanisms of global soil carbon dynamics, serving as a benchmark to constrain terrestrial biogeochemical models. The dataset is publicly available at https://doi.org/10.6084/m9.figshare.8882567 (Tang et al., 2019a).

Data-driven ENZYme (DENZY) model represents soil organic carbon dynamics in forests impacted by nitrogen deposition

Soil microorganisms participate in almost all soil organic carbon (SOC) transformations, but they are not represented explicitly in the current generation of earth system models. This study used a data-driven approach to incorporate extracellular enzyme activity into the Terrestrial ECOsystem (TECO) model, and the updated version was named the Data-driven ENZYme (DENZY) model. DENZY is based on results from an extensive data synthesis, which show that the CN ratio is positively correlated with ligninase activity (R2 = 0.50). The latter is inversely correlated to soil organic carbon storage. The DENZY model was parameterized using the revise database to information from a recent meta-analysis and tested for its ability to simulate SOC dynamics at Duke Forest (North Carolina, USA) from 1996 to 2007. DENZY can well simulate the observed negative relationship between ligninase activity and SOC under N deposition conditions (R2 ranges from 0.61 to 0.89). Moreover, outputs from DENZY better matched the observed SOC than its prototype model with the same parameterization. This study provides a simple and straightforward approach to effectively use real-world observations to improve SOC projections in terrestrial biogeochemical models.

An integrated method for high-resolution definition of new diameter-based fine root sub-classes of Fagus sylvatica L.

Compared to the traditional approach, applying micrometric image analysis to fine root samples of Fagus sylvatica with subsequent data treatment through principal component and cluster analysis yielded specific diameter sizes for fine root sub-classes having better resolution of the corresponding branching orders, and a more coherent relationship with the values of annual production and turnover rate. Fine root traits are poorly understood, impeding an accurate representation of terrestrial biogeochemical models. Traditionally used, arbitrary diameter thresholds lead to a misestimation of fine root traits such as branching order, environmental relationship, annual production, and turnover rate. Here, we present, as modification of the traditional method, an integrated approach to segregate, at high-resolution, fine root populations of Fagus sylvatica into new diameter sub-classes that better correspond with the traits mentioned above. Samples, collected with a sequential soil coring method, were subjected to a micrometric image analysis, and resultant data were treated with principal component and cluster analysis. Results showed that fine roots were distributed into diameter-size sub-classes (0–0.3 mm, 0.3–1 mm, and 1–2 mm) different from those determined by traditional methods (0–0.5 mm, 0.5–1 mm, and 1–2 mm). New sub-classes provided a better resolution of the corresponding branching-orders, and the values of annual production and turnover rate were more coherent with diameter class and soil depth. Moreover, new sub-classes provided a more precise match with soil temperature than traditional methods. Our method may help to unveil fine root dynamics and development, reduce data analysis time, and make the diameter-based classification more precise and trustworthy even in the case of non-intact samples.

Water response of ecosystem respiration regulates future projection of net ecosystem productivity in a semiarid grassland

Recent evidences show that terrestrial biogeochemical models have large uncertainty in estimating climate-change effect on grassland net ecosystem productivity (NEP), which is defined as the difference between gross ecosystem photosynthesis (GEP) and respiration (ER). It remains unclear that whether GEP or ER limits the model capability to simulate NEP responses to climate change in semiarid grasslands. Given the surrogate CENTURY-type model is widely used for Earth system modeling, we investigated two of them (i.e., DAYCENT and TECO models) and examined which processes dominate their ability to capture the responses of NEP to experimental climate changes in a temperate steppe of northern China. During the simulation from 2006 to 2008, the two models captured the observed mean annual NEP in the control plots when they were validated by the observations from an adjacent eddy-flux tower. However, they failed to capture the treatment effects of experimental warming and increased precipitation on NEP because of the poor estimations of ER responses. DAYCENT model simulated a higher precipitation effect on ER (37.83%) and TECO model overestimated the warming effect on ER by 8.18%. The simulation of treatment effects on ER and therefore NEP can be improved by an optimized parameterization of the water-related decay functions for soil organic carbon (C). The simulated cumulative loss of total ecosystem C stock during 2010–2100 were decreased when the TECO model used experiment-fitted parameters (0.72 kg C m−2) instead of using the initial validation with eddy-flux data (0.96 kg C m−2). The ecosystem shifted from C sink to source at threshold of 435 mm of annual total precipitation. Our findings indicate that future projection of C cycle in semiarid grasslands could be improved by better understanding of water response of ecosystem respiratory processes.

Rhizospheric carbon-nitrogen interactions in a mixed-species pasture after 13 years of elevated CO2

Responses of rhizospheric processes to elevated CO2 vary across mono-species studies, yet our understanding of the responses of multiple co-occurring species is still limited. We conducted a field study using the New Zealand free-air CO2 enrichment (FACE) experiment to evaluate rhizospheric carbon (C) and nitrogen (N) dynamics in a low-fertility species (Agrostis capillaris) and a high-fertility species (Lolium perenne) under 13 years of elevated CO2 and at ambient levels. Despite contrasting responses of plant growth, elevated CO2 significantly increased the total C and N (TC and TN, respectively), net N mineralization and nitrification rates (NMR and NR, respectively) as well as the δ15N value of soil organic N in the rhizospheres across species. However, the rhizospheric microbial biomass C (MBC) was significantly enhanced for L. perenne but reduced for A. capillaris under elevated CO2, and was positively corrected with plant biomass. Elevated CO2 did not affect rhizospheric NO3−-N concentrations for either species, while it significantly increased the rhizospheric NH4+-N concentration for A. capillaris. Together, the increased TC as well as the decreased MBC suggest rhizosphere in A. capillaris may be more favorable to long term C sequestration under elevated CO2. The increased NH4+-N concentration may stimulate the growth of A. capillaris under elevated CO2 in the near future, which will have significant implication for the change in species composition. The findings in this study are useful for parameterizing terrestrial biogeochemical models with mixed-species ecosystems, which can be used to analyze the effects of possible future climate change.

The effects of constraining variables on parameter optimization in carbon and water flux modeling over different forest ecosystems

The ability of terrestrial biogeochemical models in predicting land-atmospheric carbon and water exchanges is largely hampered by the insufficient characterization of model parameters. The direct observations of carbon/water fluxes and the associated environmental variables from eddy covariance (EC) flux towers provide a notable opportunity to examine the underlying processes controlling carbon and water exchanges between terrestrial ecosystems and the atmosphere. In this study, we applied the Metropolis simulated annealing technique to conduct parameter optimization analyses of a process-based biogeochemical model, simplified PnET (SIPNET), using a variety of constraining variables from EC observations and leaf area index (LAI) from MODIS at three ChinaFLUX forest sites: a temperate mixed forest (CBS), a subtropical evergreen coniferous plantation (QYZ) and a subtropical evergreen broad-leaved forest (DHS). Our analyses focused on (1) identifying the key model parameters influencing the simulation of carbon and water fluxes with SIPNET; (2) evaluating how different combinations of constraining variables influence parameter estimations and associated uncertainties; and (3) assessing the model performance with the optimized parameterization in predicting carbon and water fluxes in the three forest ecosystems. Our sensitivity analysis indicated that, among three different forest ecosystems, the prediction of carbon and water fluxes was mostly affected by photosynthesis-related parameters. The performances of the model simulations depended on different parameterization schemes, especially the combinations of constraining variables. The parameterization scheme using both net ecosystem exchange (NEE) and evapotranspiration (ET) as constraining variables performed best with most well-constrained parameters. When LAI was added to the optimization, the number of well-constrained model parameters was increased. In addition, we found that the model cannot be well-parameterized with only growing-season observations, especially for those forest ecosystems with distinct seasonal variation. With the optimized parameterization scheme using both NEE and ET observations all year round, the SIPNET were able to simulate the seasonal and inter-annual variations of carbon and water exchanges in three forest ecosystems.

Parameter and prediction uncertainty in an optimized terrestrial carbon cycle model: Effects of constraining variables and data record length

[1] Many parameters in terrestrial biogeochemical models are inherently uncertain, leading to uncertainty in predictions of key carbon cycle variables. At observation sites, this uncertainty can be quantified by applying model-data fusion techniques to estimate model parameters using eddy covariance observations and associated biometric data sets as constraints. Uncertainty is reduced as data records become longer and different types of observations are added. We estimate parametric and associated predictive uncertainty at the Morgan Monroe State Forest in Indiana, USA. Parameters in the Local Terrestrial Ecosystem Carbon (LoTEC) are estimated using both synthetic and actual constraints. These model parameters and uncertainties are then used to make predictions of carbon flux for up to 20 years. We find a strong dependence of both parametric and prediction uncertainty on the length of the data record used in the model-data fusion. In this model framework, this dependence is strongly reduced as the data record length increases beyond 5 years. If synthetic initial biomass pool constraints with realistic uncertainties are included in the model-data fusion, prediction uncertainty is reduced by more than 25% when constraining flux records are less than 3 years. If synthetic annual aboveground woody biomass increment constraints are also included, uncertainty is similarly reduced by an additional 25%. When actual observed eddy covariance data are used as constraints, there is still a strong dependence of parameter and prediction uncertainty on data record length, but the results are harder to interpret because of the inability of LoTEC to reproduce observed interannual variations and the confounding effects of model structural error.

Overlooking the canopy: The importance of canopy structure in scaling isoprenoid emissions from the leaf to the landscape

Isoprene and monoterpenes are highly reactive organic compounds, emitted by most plant species, which play an important role in air chemistry and air pollution. Different leaf-scale isoprenoid emission models are available. These models are scaled to the canopy through coupling them to terrestrial biogeochemical models and thus used to generate regional emissions inventories. Although the leaf scale models have been shown to perform similarly, large unexplained differences exist in regional emissions inventories. This may be explained in part by the complete lack of inter-comparisons of emission model estimates when scaled from the leaf to the canopy. In this paper we address this problem by coupling four different isoprene emission models (Guenther et al. model, Niinemets et al. model, BIM2 and the Martin et al. model) to two terrestrial biogeochemical model platforms (MoBiLE, GOTILWA+) that describe canopy structure differently. Simulations of isoprene emissions for the Puechabon Mediterranean holm oak stand are analysed, with both canopy photosynthesis models constrained using FLUXNET measurements. The results demonstrate that even with constrained canopy level photosynthesis, large model platform dependent within canopy differences can exist in both modelled photosynthesis and emissions. This results in large differences in modelled isoprenoid emissions, due to the relatively higher sensitivity of emissions to canopy microclimate, in particular temperature. This is the first time emission results from two biogeochemical platforms have been compared, and demonstrates that different canopy descriptions can lead to larger differences in modelled emissions than that attributable to the difference between the emission models themselves. This is an important aspect that has not been acknowledged by the emission modelling community.

Effects of processes at the population and community level on carbon dynamics of an ecosystem model

AbstractEcological processes at the population and community level are often ignored in biogeochemical models, however, the effects of excluding these processes at the ecosystem level is uncertain. In this study we analyzed the set of behaviors that emerge after introducing population and community processes into an ecosystem carbon model. We used STANDCARB, a hybrid model that incorporates population, community, and ecosystem processes to predict carbon dynamics over time. Our simulations showed that at the population level, colonization and mortality rates can limit the maximum biomass achieved during a successional sequence. Specifically, colonization rates control temporal lags in the initiation of carbon accumulation, and mortality rates can have important effects on annual variation in live biomass. At the community level, differences in species traits and changes in species composition over time introduced significant changes in carbon dynamics. Species with different set of parameters, such as growth and mortality rates, introduce patterns of carbon accumulation that could not be reproduced using a single species with the average of parameters of multiple species or by simulating the most abundant species (strategies commonly employed in terrestrial biogeochemical models). We conclude that omitting population and community processes from biogeochemical models introduces an important source of uncertainty that can impose important limitations for predictions of future carbon balances.

Contribution of ocean, fossil fuel, land biosphere, and biomass burning carbon fluxes to seasonal and interannual variability in atmospheric CO2

Seasonal and interannual variability in atmospheric carbon dioxide (CO2) concentrations was simulated using fluxes from fossil fuel, ocean and terrestrial biogeochemical models, and a tracer transport model with time‐varying winds. The atmospheric CO2 variability resulting from these surface fluxes was compared to observations from 89 GLOBALVIEW monitoring stations. At northern hemisphere stations, the model simulations captured most of the observed seasonal cycle in atmospheric CO2, with the land tracer accounting for the majority of the signal. The ocean tracer was 3–6 months out of phase with the observed cycle at these stations and had a seasonal amplitude only ∼10% on average of observed. Model and observed interannual CO2 growth anomalies were only moderately well correlated in the northern hemisphere (R ∼ 0.4–0.8), and more poorly correlated in the southern hemisphere (R < 0.6). Land dominated the interannual variability (IAV) in the northern hemisphere, and biomass burning in particular accounted for much of the strong positive CO2 growth anomaly observed during the 1997–1998 El Niño event. The signals in atmospheric CO2 from the terrestrial biosphere extended throughout the southern hemisphere, but oceanic fluxes also exerted a strong influence there, accounting for roughly half of the IAV at many extratropical stations. However, the modeled ocean tracer was generally uncorrelated with observations in either hemisphere from 1979–2004, except during the weak El Niño/post‐Pinatubo period of the early 1990s. During that time, model results suggested that the ocean may have accounted for 20–25% of the observed slowdown in the atmospheric CO2 growth rate.

OptIC project: An intercomparison of optimization techniques for parameter estimation in terrestrial biogeochemical models

We describe results of a project known as OptIC (Optimisation InterComparison) for comparison of parameter estimation methods in terrestrial biogeochemical models. A highly simplified test model was used to generate pseudo‐data to which noise with different characteristics was added. Participants in the OptIC project were asked to estimate the model parameters used to generate this data, and to predict model variables into the future. Ten participants contributed results using one of the following methods: Levenberg‐Marquardt, adjoint, Kalman filter, Markov chain Monte Carlo and genetic algorithm. Methods differed in how they locate the minimum (gradient‐descent or global search), how observations are processed (all at once sequentially), or the number of iterations used, or assumptions about the statistics (some methods assume Gaussian probability density functions; others do not). We found the different methods equally successful at estimating the parameters in our application. The biggest variation in parameter estimates arose from the choice of cost function, not the choice of optimization method. Relatively poor results were obtained when the model‐data mismatch in the cost function included weights that were instantaneously dependent on noisy observations. This was the case even when the magnitude of residuals varied with the magnitude of observations. Missing data caused estimates to be more scattered, and the uncertainty of predictions increased correspondingly. All methods gave biased results when the noise was temporally correlated or non‐Gaussian, or when incorrect model forcing was used. Our results highlight the need for care in choosing the error model in any optimization.

Comparing and evaluating process-based ecosystem model predictions of carbon and water fluxes in major European forest biomes.

Process-based models can be classified into: (a) terrestrial biogeochemical models (TBMs), which simulate fluxes of carbon, water and nitrogen coupled within terrestrial ecosystems, and (b) dynamic global vegetation models (DGVMs), which further couple these processes interactively with changes in slow ecosystem processes depending on resource competition, establishment, growth and mortality of different vegetation types. In this study, four models - RHESSys, GOTILWA+, LPJ-GUESS and ORCHIDEE - representing both modelling approaches were compared and evaluated against benchmarks provided by eddy-covariance measurements of carbon and water fluxes at 15 forest sites within the EUROFLUX project. Overall, model-measurement agreement varied greatly among sites. Both modelling approaches have somewhat different strengths, but there was no model among those tested that universally performed well on the two variables evaluated. Small biases and errors suggest that ORCHIDEE and GOTILWA+ performed better in simulating carbon fluxes while LPJ-GUESS and RHESSys did a better job in simulating water fluxes. In general, the models can be considered as useful tools for studies of climate change impacts on carbon and water cycling in forests. However, the various sources of variation among models simulations and between models simulations and observed data described in this study place some constraints on the results and to some extent reduce their reliability. For example, at most sites in the Mediterranean region all models generally performed poorly most likely because of problems in the representation of water stress effects on both carbon uptake by photosynthesis and carbon release by heterotrophic respiration (Rh ). The use of flux data as a means of assessing key processes in models of this type is an important approach to improving model performance. Our results show that the models have value but that further model development is necessary with regard to the representation of the some of the key ecosystem processes.

Characterization of North American land cover from NOAA‐AVHRR data using the EOS MODIS Land Cover Classification Algorithm

Land cover is a key boundary condition in weather, climate, and terrestrial biogeochemical models. Until recently, such models have used maps depicting potential vegetation, which are known to be of relatively poor quality, to parameterize land surface properties. In this paper we describe the compilation and assessment of a new map of North American land cover produced through the application of advanced pattern recognition techniques to multitemporal satellite data. This map was produced in a fully automated fashion using supervised classification methods that are robust, fully automated, and repeatable. The processing flow described in this paper is a prototype of the algorithm to be used to generate maps of global land cover using data from EOS MODIS. The superior quality and timeliness of these maps should be very useful for a wide array of sub‐continental to global‐scale modeling and analysis activities.

Satellite observation of El Niño effects on Amazon Forest phenology and productivity

Climate variability may affect the functioning of Amazon moist tropical forests, and recent modeling analyses suggest that the carbon dynamics of the region vary interannually in response to precipitation and temperature anomalies. However, due to persistent orbital and atmospheric artifacts in the satellite record, remote sensing observations have not provided quantitative evidence that climate variation affects Amazon forest phenology or productivity. We developed a method to minimize and quantify non‐biological artifacts in NOAA AVHRR satellite data, providing a record of estimated forest phenological variation from 1982–1993. The seasonal NDVI amplitude (a proxy for phenology) increased throughout much of the basin during El Niño periods when rainfall was anomalously low. Wetter La Niña episodes brought consistently smaller NDVI amplitudes. Using radiative transfer and terrestrial biogeochemical models driven by these satellite data, we estimate that canopy energy absorption and net primary production of Amazon forests varied interannually by as much as 21% and 18%, respectively. These results provide large‐scale observational evidence for interannual sensitivity to El Niño of plant phenology and carbon flux in Amazon forests.

Modeling the effects of snowpack on heterotrophic respiration across northern temperate and high latitude regions: Comparison with measurements of atmospheric carbon dioxide in high latitudes

Simulations by global terrestrial biogeochemical models (TBMs) consistently underestimate the concentration of atmospheric carbon dioxide (CO2 at high latitude monitoring stations during the non-growing season. We hypothesized that heterotrophic respiration is underestimated during the nongrowing season primarily because TBMs do not generally consider the insulative effects of snowpack on soil temperature. To evaluate this hypothesis, we compared the performance of baseline and modified versions of three TBMs in simulating the seasonal cycle of atmospheric CO2 at high latitude CO2 monitoring stations; the modified version maintained soil temperature at 0 °C when modeled snowpack was present. The three TBMs include the Carnegie-Ames-Stanford Approach (CASA), Century, and the Terrestrial Ecosystem Model (TEM). In comparison with the baseline simulation of each model, the snowpack simulations caused higher releases of CO2 between November and March and greater uptake of CO2 between June and August for latitudes north of 30° N. We coupled the monthly estimates of CO2 exchange, the seasonal carbon dioxide flux fields generated by the HAMOCC3 seasonal ocean carbon cycle model, and fossil fuel source fields derived from standard sources to the three-dimensional atmospheric transport model TM2 forced by observed winds to simulate the seasonal cycle of atmospheric CO2 at each of seven high latitude monitoring stations. In comparison to the CO2 concentrations simulated with the baseline fluxes of each TBM, concentrations simulated using the snowpack fluxes are generally in better agreement with observed concentrations between August and March at each of the monitoring stations. Thus, representation of the insulative effects of snowpack in TBMs generally improves simulation of atmospheric CO2 concentrations in high latitudes during both the late growing season and nongrowing season. These simulations highlight the global importance of biogeochemical processes during the nongrowing season in estimating carbon balance of ecosystems in northern high and temperate latitudes.

Evaluation of terrestrial carbon cycle models through simulations of the seasonal cycle of atmospheric CO2: First results of a model intercomparison study

Results of an intercomparison among terrestrial biogeochemical models (TBMs) are reported, in which one diagnostic and five prognostic models have been run with the same long‐term climate forcing. Monthly fields of net ecosystem production (NEP), which is the difference between net primary production (NPP) and heterotrophic respiration RH, at 0.5° resolution have been generated for the terrestrial biosphere. The monthly estimates of NEP in conjunction with seasonal CO2 flux fields generated by the seasonal Hamburg Model of the Oceanic Carbon Cycle (HAMOCC3) and fossil fuel source fields were subsequently coupled to the three‐dimensional atmospheric tracer transport model TM2 forced by observed winds. The resulting simulated seasonal signal of the atmospheric CO2 concentration extracted at the grid cells corresponding to the locations of 27 background monitoring stations of the National Oceanic and Atmospheric Administration/Climate Monitoring and Diagnostics Laboratory network is compared with measurements from these sites. The Simple Diagnostic Biosphere Model (SDBM1), which is tuned to the atmospheric CO2 concentration at five monitoring stations in the northern hemisphere, successfully reproduced the seasonal signal of CO2 at the other monitoring stations. The SDBM1 simulations confirm that the north‐south gradient in the amplitude of the atmospheric CO2 signal results from the greater northern hemisphere land area and the more pronounced seasonality of radiation and temperature in higher latitudes. In southern latitudes, ocean‐atmosphere gas exchange plays an important role in determining the seasonal signal of CO2. Most of the five prognostic models (i.e., models driven by climatic inputs) included in the intercomparison predict in the northern hemisphere a reasonably accurate seasonal cycle in terms of amplitude and, to some extent, also with respect to phase. In the tropics, however, the prognostic models generally tend to overpredict the net seasonal exchanges and stronger seasonal cycles than indicated by the diagnostic model and by observations. The differences from the observed seasonal signal of CO2 may be caused by shortcomings in the phenology algorithms of the prognostic models or by not properly considering the effects of land use and vegetation fires on CO2 fluxes between the atmosphere and terrestrial biosphere.