Terrestrial Ecosystem Model Research Articles

Deterministic and stochastic components of atmospheric CO[formula omitted] inside forest canopies and consequences for predicting carbon and water exchange

Atmospheric CO2 concentrations strongly influence the exchange of energy, water and carbon between the atmosphere and the terrestrial biosphere. The CO2 available to plants can be highly variable given the stochastic nature of the phenomena involved in its dynamics. However, most terrestrial ecosystem models consider CO2 concentration as a quasi-deterministic variable or use measurements taken from heights above the canopy where CO2 is usually well mixed. Therefore, in this study, we aimed to evaluate if a stochastic treatment of CO2 concentrations leads to different predictions of carbon assimilation (An) and transpiration (T) compared to a strictly deterministic approach, as well as the use of CO2 measurements from different heights. To address this goal, we used a set of observations taken in forests from the ICOS network and the ATTO research site. We applied the exponential smoothing method to decompose the time series into their trend, seasonal and stochastic (residual) components. We found that the residual component of CO2 (rCO2) follows a Laplace probability density function and its stochastic magnitude is inversely proportional to the height above the ground. To quantify the degree to which predictions of An and T would be affected by stochastic effects, we ran simulations considering the different components of the time series and Farquhar’s and Penman–Monteith’s models. We found significant differences in the predictions of An and T for diverse heights, with larger An (T) fluxes closer (farther) to the ground. Still, this effect is mainly due to the deterministic component that increases with decreasing height. The stochastic component tends to reduce (increase) An (T) compared to the deterministic approach. However, the difference between approaches is not large enough to compensate for the deterministic effect, suggesting that there is no merit for the consideration of rCO2 in future simulations as long as measurements taken inside the canopy are used.

Scale variance in the carbon dynamics of fragmented, mixed-use landscapes estimated using model–data fusion

Abstract. Many terrestrial landscapes are heterogeneous. Mixed land cover and land use generate a complex mosaic of fragmented ecosystems at fine spatial resolutions with contrasting ecosystem stocks, traits, and processes, each differently sensitive to environmental and human factors. Representing spatial complexity within terrestrial ecosystem models is a key challenge for understanding regional carbon dynamics, their sensitivity to environmental gradients, and their resilience in the face of climate change. Heterogeneity underpins this challenge due to the trade-off between the fidelity of ecosystem representation within modelling frameworks and the computational capacity required for fine-scale model calibration and simulation. We directly address this challenge by quantifying the sensitivity of simulated carbon fluxes in a mixed-use landscape in the UK to the spatial resolution of the model analysis. We test two different approaches for combining Earth observation (EO) data into the CARDAMOM model–data fusion (MDF) framework, assimilating time series of satellite-based EO-derived estimates of ecosystem leaf area and biomass stocks to constrain estimates of model parameters and their uncertainty for an intermediate complexity model of the terrestrial C cycle. In the first approach, ecosystems are calibrated and simulated at pixel level, representing a “community average” of the encompassed land cover and management. This represents our baseline approach. In the second, we stratify each pixel based on land cover (e.g. coniferous forest, arable/pasture) and calibrate the model independently using EO data specific to each stratum. We test the scale dependence of these approaches for grid resolutions spanning 1 to 0.05∘ over a mixed-land-use region of the UK. Our analyses indicate that spatial resolution matters for MDF. Under the community average baseline approach biological C fluxes (gross primary productivity, Reco) simulated by CARDAMOM are relatively insensitive to resolution. However, disturbance fluxes exhibit scale variance that increases with greater landscape fragmentation and for coarser model domains. In contrast, stratification of assimilated data based on fine-resolution land use distributions resolved the resolution dependence, leading to disturbance fluxes that were 40 %–100 % higher than the baseline experiments. The differences in the simulated disturbance fluxes result in estimates of the terrestrial carbon balance in the stratified experiment that suggest a weaker C sink compared to the baseline experiment. We also find that stratifying the model domain based on land use leads to differences in the retrieved parameters that reflect variations in ecosystem function between neighbouring areas of contrasting land use. The emergent differences in model parameters between land use strata give rise to divergent responses to future climate change. Accounting for fine-scale structure in heterogeneous landscapes (e.g. stratification) is therefore vital for ensuring the ecological fidelity of large-scale MDF frameworks. The need for stratification arises because land use places strong controls on the spatial distribution of carbon stocks and plant functional traits and on the ecological processes controlling the fluxes of C through landscapes, particularly those related to management and disturbance. Given the importance of disturbance to global terrestrial C fluxes, together with the widespread increase in fragmentation of forest landscapes, these results carry broader significance for the application of MDF frameworks to constrain the terrestrial C balance at regional and national scales.

A boreal forest model benchmarking dataset for North America: a case study with the Canadian Land Surface Scheme Including Biogeochemical Cycles (CLASSIC)

Climate change is rapidly altering composition, structure, and functioning of the boreal biome, across North America often broadly categorized into ecoregions. The resulting complex changes in different ecoregions present a challenge for efforts to accurately simulate carbon dioxide (CO2) and energy exchanges between boreal forests and the atmosphere with terrestrial ecosystem models (TEMs). Eddy covariance measurements provide valuable information for evaluating the performance of TEMs and guiding their development. Here, we compiled a boreal forest model benchmarking dataset for North America by harmonizing eddy covariance and supporting measurements from eight black spruce (Picea mariana)-dominated, mature forest stands. The eight forest stands, located in six boreal ecoregions of North America, differ in stand characteristics, disturbance history, climate, permafrost conditions and soil properties. By compiling various data streams, the benchmarking dataset comprises data to parameterize, force, and evaluate TEMs. Specifically, it includes half-hourly, gap-filled meteorological forcing data, ancillary data essential for model parameterization, and half-hourly, gap-filled or partitioned component flux data on CO2 (net ecosystem production, gross primary production [GPP], and ecosystem respiration [ER]) and energy (latent [LE] and sensible heat [H]) and their daily aggregates screened based on half-hourly gap-filling quality criteria. We present a case study with the Canadian Land Surface Scheme Including Biogeochemical Cycles (CLASSIC) to: (1) demonstrate the utility of our dataset to benchmark TEMs and (2) provide guidance for model development and refinement. Model skill was evaluated using several statistical metrics and further examined through the flux responses to their environmental controls. Our results suggest that CLASSIC tended to overestimate GPP and ER among all stands. Model performance regarding the energy fluxes (i.e., LE and H) varied greatly among the stands and exhibited a moderate correlation with latitude. We identified strong relationships between simulated fluxes and their environmental controls except for H, thus highlighting current strengths and limitations of CLASSIC.

Uncertainty in land use obscures global soil organic carbon stock estimates

The impact of land use and land cover change (LULCC) on soil organic carbon (SOC) stock is one of the most uncertain items in estimating the global C budget. Despite the improvements in satellite monitoring techniques and inventory data in recent decades, the uncertainty in modeled LULCC-induced SOC changes stemming from the choice of land use datasets remains largely unknown. Using a process-based model, the Dynamic Land Ecosystem Model (DLEM), we investigated global SOC changes during 1900–2018 driven by six LULCC datasets (i.e., LUH2-GCB2019, ESA CCI-LC, MODIS, GLASS-GLC, HH, and RF), which were generated by varied data sources and methodologies. The simulated global SOC stock was negatively affected by land conversions in all the LULCC datasets; however, the corresponding SOC loss was highly different ranging from 33.11 Pg C to 106.20 Pg C. Such significant differences in the global SOC stock estimates due to LULCC uncertainty were mainly located in boreal and temperate forests of the northern high latitudes and were most likely attributed to the LULCC-induced changes in vegetation net primary production. Meanwhile, regions exhibiting large divergence in relative changes of LULCC-induced SOC loss were mainly located in the low latitudes. When considering the interactive effects of LULCC with other environmental factors, the simulated SOC showed divergent trends, increasing in MODIS-, ESA CCI-LC-, and GLASS-GLC-based estimations, but decreasing in LUH2-GCB2019-, HH-, and RF-based estimations. These results highlight the importance of the accuracy of LULCC data in determining the global carbon budget. Future efforts are required for harmonizing satellite observations and inventory data both spatially and temporally to better represent the land conversion processes in terrestrial ecosystem modeling.

LandInG 1.0: a toolbox to derive input datasets for terrestrial ecosystem modelling at variable resolutions from heterogeneous sources

Abstract. We present the Land Input Generator (LandInG) version 1.0, a new toolbox for generating input datasets for terrestrial ecosystem models (TEMs) from diverse and partially conflicting data sources. While LandInG 1.0 is applicable to process data for any TEM, it is developed specifically for the open-source dynamic global vegetation, hydrology, and crop growth model LPJmL (Lund–Potsdam–Jena with managed Land). The toolbox documents the sources and processing of data to model inputs and allows for easy changes to the spatial resolution. It is designed to make inconsistencies between different sources of data transparent so that users can make their own decisions on how to resolve these should they not be content with the default assumptions made here. As an example, we use the toolbox to create input datasets at 5 and 30 arcmin spatial resolution covering land, country, and region masks, soil, river networks, freshwater reservoirs, irrigation water distribution networks, crop-specific annual land use, fertilizer, and manure application. We focus on the toolbox describing the data processing rather than only publishing the datasets as users may want to make different choices for reconciling inconsistencies, aggregation, spatial extent, or similar. Also, new data sources or new versions of existing data become available continuously, and the toolbox approach allows for incorporating new data to stay up to date.

Seasonal variation of ecosystem photosynthetic capacity and its environmental drivers in global grasslands

Ecosystem maximum photosynthetic rate (Amax) is an important ecosystem functional property, as it is critical for ecosystem productivity modeling. However, little is known about the mechanisms that regulate the seasonal variation of Amax in grasslands, one of the dominant vegetation types worldwide. In this study, we analyzed the seasonal variability of Amax of grassland sites across the globe and its environmental drivers. We found that grassland Amax had strong seasonal variations, which were influenced by the climate and agricultural management, such as grass cutting and grazing. Second, the seasonal variation of Amax at all arid grasslands [mean annual vapor pressure deficit (VPD) > 10 hPa] was driven more by changes in canopy physiological property (i.e., maximum photosynthetic rate per leaf area Amaxa) than canopy structural property (i.e., leaf area, presented by LAI), because Amaxa had stronger temporal variability than LAI in these ecosystems. Third, temperature and VPD were the most influential factors for the seasonal variability of Amax and LAI, but environmental variables only explained a small proportion of the seasonal variation of Amaxa, which was probably because Amaxa was more related to plant traits. Our findings provide new ideas for better parameterizations of Amax in terrestrial ecosystem models.

Comparing simulated tree biomass from daily, monthly, and seasonal climate input of terrestrial ecosystem model

Terrestrial ecosystem models are driven by climate data, which often have daily, monthly, and seasonal resolutions, but few have examined the effects of different temporal resolutions of climate data on modeled vegetation change. In this study, we investigated the effects of daily, monthly, and seasonal climate data on simulated tree-species biomass in a temperate forest region of northeast China. We conducted the study by using the LINKAGES terrestrial ecosystem model to quantify the relative importance of climate data resolutions (daily, monthly and seasonal) and interspecific competition on predicted tree-species biomass from short-term (<40 years) to long-term (70–100 years) simulations. Results showed that simulated tree-species biomass was sensitive to daily, monthly, and seasonal climate data. The difference in the simulated biomass based on daily vs. monthly climate data was not significant for tree species monocultures, suggesting that monthly climate data can be used as a surrogate for daily climate data in such simulations. However, interspecific competition amplifies the differences in biomass simulations under daily and monthly climate data, thus increasing prediction uncertainties in mixed-species systems. Results based on climate data at the seasonal resolution showed the greatest departure from the daily and monthly climate data, because averaging climate data over a season may result in unrealistic climate estimates for modeling tree biomass growth. Simulated biomass of shade-intolerant tree species was highest when using climate data at the daily resolution, followed by monthly and seasonal climate data. The opposite pattern was true for shade-tolerant tree species. At the early stages of simulated tree growth, inter-tree competition was a more important factor affecting the biomass of late-successional, shade-tolerant tree species while different temporal resolutions of climate data were more important for early- and mid-successional species. This trend reverses at the late stages of stand development. These results can help guide choosing climate input of terrestrial ecosystem models and aiding results interpretation.

Net ecosystem productivity of Panjin Phragmites australis wetland during 1971 to 2020 and its impact factors.

Coastal estuarine wetland ecosystem has strong ability for carbon (C) storage and sequestration. Accurate assessment of C sequestration and its environmental impact factors is the basis of scientific protection and mana-gement of coastal estuarine wetlands. Taking the Panjin reed (Phragmites australis) wetland as the object, we used terrestrial ecosystem model, together with Mann-Kendall mutation test, statistical analysis methods, and scenario simulation experiment, to analyze the temporal characteristics, stability, changing trend of net ecosystem production (NEP) of wetlands and the contribution rate of environmental impact factors to NEP during 1971 to 2020. The results showed that the annual average NEP of Panjin reed wetland was 415.51 g C·m-2·a-1 during 1971 to 2020, with a steady increase rate of 1.7 g C·m-2·a-1, which would still have a continuous increasing trend in the future. The annual average NEP in spring, summer, autumn, and winter was 33.95, 418.05, -18.71, and -17.78 g C·m-2·a-1, with an increase rate of 0.35, 1.26, 0.14 and -0.06 g C·m-2·a-1, respectively. In the future, NEP would show an increasing trend in both spring and summer, but a declining trend in both autumn and winter. The contribution rates of environmental impact factors to NEP of Panjin reed wetland depended on temporal scale. At the interannual scale, the contribution rate of precipitation was the highest (37.1%), followed by CO2 (28.4%), air temperature (25.1%) and photosynthetically active radiation (9.4%). Precipitation mainly affected NEP in both spring and autumn with the contribution rates of 49.5% and 38.8%, while CO2 concentration (36.9%) and air temperature (-86.7%) were dominant in summer and winter, respectively.

Thermal acclimation of plant photosynthesis and autotrophic respiration in a northern peatland

Peatlands contain one-third of global soil carbon (C), but the responses of peatland ecosystems to long-term warming are not well understood. Here, we pursue an emergent understanding of warming effects on ecosystem C fluxes at peatlands by constraining a process-oriented model, the terrestrial ECOsystem model, with observational data from a long-term warming experiment at the Spruce and Peatland Responses Under Changing Environments site. Model-based assessments show that ecosystem-level photosynthesis and autotrophic respiration exhibited significant thermal acclimation, with temperature sensitivities being linearly decreased with warming. Using the thermal-acclimated parameter values, simulated gross primary production, net primary production, and plant autotrophic respiration (R a), were all lower than those simulated with non-thermal acclimated parameter values. In contrast, ecosystem respiration simulated with thermal acclimated parameter values was higher than that simulated with non-thermal acclimated parameter values. Net ecosystem CO2 exchange was much higher after constraining model parameters with observational data from the warming treatments, releasing C at a rate of 28.3 g C m−2 yr−1 °C−1. Our data-model integration study suggests that peatlands are likely to release more C than previously estimated. Earth system models may overestimate C uptake by peatlands under warming if physiological thermal acclimation of plants is not incorporated. Thus, it is critical to consider the long-term physiological thermal acclimation of plants in the models to better predict global C dynamics under future climate and their feedback to climate change.

Underestimated Interannual Variability of Terrestrial Vegetation Production by Terrestrial Ecosystem Models

AbstractVegetation gross primary production (GPP) is the largest terrestrial carbon flux and plays an important role in regulating the carbon sink. Current terrestrial ecosystem models (TEMs) are indispensable tools for evaluating and predicting GPP. However, to which degree the TEMs can capture the interannual variability (IAV) of GPP remains unclear. With large data sets of remote sensing, in situ observations, and predictions of TEMs at a global scale, this study found that the current TEMs substantially underestimate the GPP IAV in comparison to observations at global flux towers. Our results also showed the larger underestimations of IAV in GPP at nonforest ecosystem types than forest types, especially in arid and semiarid grassland and shrubland. One cause of the underestimation is that the IAV in GPP predicted by models is strongly dependent on canopy structure, that is, leaf area index (LAI), and the models underestimate the changes of canopy physiology responding to climate change. On the other hand, the simulated interannual variations of LAI are much less than the observed. Our results highlight the importance of improving TEMs by precisely characterizing the contribution of canopy physiological changes on the IAV in GPP and of clarifying the reason for the underestimated IAV in LAI. With these efforts, it may be possible to accurately predict the IAV in GPP and the stability of the global carbon sink in the context of global climate change.

Applications of CNOP-P Method to Predictability Studies of Terrestrial Ecosystems

In this paper, recent research on terrestrial ecosystem predictability using the conditional nonlinear optimal parameter perturbation (CNOP-P) method is summarized. The main findings include the impacts of uncertainties in climate change on uncertainties in simulated terrestrial ecosystems, the identification of key physical parameters that lead to large uncertainties in terrestrial ecosystem modeling and prediction, and the evaluation of the simulation ability and prediction skill of terrestrial ecosystems by reducing key physical parameter errors. The study areas included the Inner Mongolia region, north–south transect of eastern China, and Qinghai–Tibet Plateau region. The periods of the studies were from 1961 to 1970 for the impacts of uncertainties in climate change on uncertainties in simulated terrestrial ecosystems, and from 1951 to 2000 for the identification of the most sensitive combinations of physical parameters. Climatic Research Unit (CRU) data were employed. The numerical results indicate the important role of nonlinear changes in climate variability due to the occurrences of extreme events characterized by CNOP-P in the abrupt grassland ecosystem equilibrium state and formation of carbon sinks in China. Second, the most sensitive combinations of physical parameters to the uncertainties in simulations and predictions of terrestrial ecosystems identified by the CNOP-P method were more sensitive than those obtained by traditional methods (e.g., one-at-a-time (OAT) and stochastic methods). Furthermore, the improvement extent of the simulation ability and prediction skill of terrestrial ecosystems by reducing the errors of the sensitive physical parameter combinations identified by the CNOP-P method was higher than that by the traditional methods.

Atmospheric dryness thresholds of grassland productivity decline in China

Atmospheric dryness events are bound to have a broad and profound impact on the functions and structures of grassland ecosystems. Current research has confirmed that atmospheric dryness is a key moisture constraint that inhibits grassland productivity, yet the risk threshold for atmospheric dryness to initiate ecosystem productivity loss has not been explored. Based on this, we used four terrestrial ecosystem models to simulate gross primary productivity (GPP) data, analyzed the role of vapor pressure deficit (VPD) in regulating interannual variability in Chinese grasslands by focusing on the dependence structure of VPD and GPP, and then constructed a bivariate linkage function to calculate the conditional probability of ecosystem GPP loss under atmospheric dryness, and further analyzed the risk threshold of ecosystem GPP loss triggered by atmospheric dryness. The main results are as follows: we found that (1) the observed and modeled VPD of Chinese grasslands increases rapidly in both historical and future periods. VPD has a strongly negative regulation on ecosystem GPP, and atmospheric dryness is an important moisture constraint that causes deficit and even death to ecosystem GPP. (2) The probability of the enhanced atmospheric dryness that induced GPP decline in Chinese grasslands in the future period increases significantly. (3) When the VPD is higher than 40.07 and 27.65 percentile of the past and future time series, respectively, the risk threshold of slight ecosystem GPP loss can be easily initiated by atmospheric dryness. (4) When the VPD is higher than 82.57 and 65.09 percentile, respectively, the threshold of moderate ecosystem GPP loss can be exceeded by the benchmark probability. (5) The risk threshold of severe ecosystem GPP loss is not initiated by atmospheric dryness in the historical period, and the threshold of severe ecosystem GPP loss can be initiated when the future VPD is higher than 91.92 percentile. In total, a slight atmospheric dryness event is required to initiate a slight ecosystem GPP loss threshold, and a stronger atmospheric dryness event is required to initiate a severe ecosystem GPP loss. Our study enhances the understandings of past and future atmospheric dryness on grassland ecosystems, and strongly suggests that more attention be invested in improving next-generation models of vegetation dynamics processes with respect to the response of mechanisms of ecosystem to atmospheric dryness.

Embracing fine-root system complexity in terrestrial ecosystem modeling.

Projecting the dynamics and functioning of the biosphere requires a holistic consideration of whole-ecosystem processes. However, biases toward leaf, canopy, and soil modeling since the 1970s have constantly left fine-root systems being rudimentarily treated. As accelerated empirical advances in the last two decades establish clearly functional differentiation conferred by the hierarchical structure of fine-root orders and associations with mycorrhizal fungi, a need emerges to embrace this complexity to bridge the data-model gap in still extremely uncertain models. Here, we propose a three-pool structure comprising transport and absorptive fine roots with mycorrhizal fungi (TAM) to model vertically resolved fine-root systems across organizational and spatial-temporal scales. Emerging from a conceptual shift away from arbitrary homogenization, TAM builds upon theoretical and empirical foundations as an effective and efficient approximation that balances realism and simplicity. A proof-of-concept demonstration of TAM in a big-leaf model both conservatively and radically shows robust impacts of differentiation within fine-root systems on simulating carbon cycling in temperate forests. Theoretical and quantitative support warrants exploiting its rich potentials across ecosystems and models to confront uncertainties and challenges for a predictive understanding of the biosphere. Echoing a broad trend of embracing ecological complexity in integrative ecosystem modeling, TAM may offer a consistent framework where modelers and empiricists can work together toward this grand goal.

Early Evidence That Soil Dryness Causes Widespread Decline in Grassland Productivity in China

The burning of fossil fuels by humans emits large amounts of CO2 into the atmosphere and strongly affects the Earth’s carbon balance, with grassland ecosystems changing from weak carbon sinks that were previously close to equilibrium to core carbon sinks. Chinese grasslands are located in typical arid–semi-arid and semi-arid climatic regions, and drought events in the soil and atmosphere can have strong and irreversible consequences on the function and structure of Chinese grassland ecosystems. Based on this, we investigated the response of the gross primary production (GPP) of Chinese grasslands to land–atmosphere moisture constraints, using GPP data simulated through four terrestrial ecosystem models and introduced copula functions and Bayesian equations. The main results were as follows: (1) Soil moisture trends were not significant, and changes were dominated by interannual variability. The detrended warm-season SM correlated with GPP at 0.48 and 0.63 for the historical and future periods, respectively; thus, soil moisture is the critical water stress that regulates interannual variability in Chinese grassland GPP. (2) The positive correlation between shallow SM (0–50 cm) and GPP was higher (r = 0.62). Shallow-soil moisture is the main soil layer that constrains GPP, and the soil moisture decrease in shallow layers is much more likely to cause GPP decline in Chinese grasslands than that in deep-soil water. (3) The probability of GPP decline in Chinese grasslands caused by drought in shallow soils of 0–20 and 20–50 cm is 32.49% and 27.64%, respectively, which is much higher than the probability of GPP decline in deeper soils. In particular, soil drought was more detrimental to grassland GPP in Xinjiang and the Loess Plateau. (4) The probability of soil drought causing GPP decline was higher than that of atmospheric drought during the historical period (1.78–8.19%), but the probability of an atmospheric drought-induced GPP deficit increases significantly in the future and becomes a key factor inhibiting GPP accumulation in some regions (e.g., the Loess Plateau). Our study highlighted the response of grassland ecosystems after the occurrence of soil drought, especially for the shallow-soil-water indicator, which provides important theoretical references for grassland drought disaster emergency prevention and policy formulation.

Peatlands and their carbon dynamics in northern high latitudes from 1990 to 2300: a process-based biogeochemistry model analysis

Abstract. Northern peatlands have been a large C sink during the Holocene, but whether they will keep being a C sink under future climate change is uncertain. This study simulates the responses of northern peatlands to future climate until 2300 with a Peatland version Terrestrial Ecosystem Model (PTEM). The simulations are driven with two sets of CMIP5 climate data (IPSL-CM5A-LR and bcc-csm1-1) under three warming scenarios (RCPs 2.6, 4.5 and 8.5). Peatland area expansion, shrinkage, and C accumulation and decomposition are modeled. In the 21st century, northern peatlands are projected to be a C source of 1.2–13.3 Pg C under all climate scenarios except for RCP 2.6 of bcc-csm1-1 (a sink of 0.8 Pg C). During 2100–2300, northern peatlands under all scenarios are a C source under IPSL-CM5A-LR scenarios, being larger sources than bcc-csm1-1 scenarios (5.9–118.3 vs. 0.7–87.6 Pg C). C sources are attributed to (1) the peatland water table depth (WTD) becoming deeper and permafrost thaw increasing decomposition rate; (2) net primary production (NPP) not increasing much as climate warms because peat drying suppresses net N mineralization; and (3) as WTD deepens, peatlands switching from moss–herbaceous dominated to moss–woody dominated, while woody plants require more N for productivity. Under IPSL-CM5A-LR scenarios, northern peatlands remain as a C sink until the pan-Arctic annual temperature reaches −2.6 to −2.89 ∘C, while this threshold is −2.09 to −2.35 ∘C under bcc-csm1-1 scenarios. This study predicts a northern peatland sink-to-source shift in around 2050, earlier than previous estimates of after 2100, and emphasizes the vulnerability of northern peatlands to climate change.

A comparative study of 17 phenological models to predict the start of the growing season

Vegetation phenological models play a major role in terrestrial ecosystem modeling. However, substantial uncertainties still occur in phenology models because the mechanisms underlying spring phenological events are unclear. Taking into account the asymmetric effects of daytime and nighttime temperature on spring phenology, we analyzed the performance of 17 spring phenological models by combining the effects of photoperiod and precipitation. The global inventory modeling and mapping study third-generation normalized difference vegetation index data (1982–2014) were used to extract the start of the growing season (SOS) in the North–South Transect of Northeast Asia. The satellite-derived SOS of deciduous needleleaf forest (DNF), mixed forest (MF), open shrublands (OSL), and woody savannas (WS) showed high correlation coefficients (r) with the model-predicted SOS, with most exceeding 0.7. For all vegetation types studied, the models that considered the effect of photoperiod and precipitation did not significantly improve the model performance. For temperature-based models, the model using the growing-degree-day temperature response had a lower root mean square error compared with the models using the sigmoid temperature response Importantly, we found that daily maximum temperature was most suitable for the spring phenology prediction of DNF, OSL, and WS; daily mean temperature for MF; and daily minimum temperature for grasslands. These findings indicate that future spring phenological models should consider the asymmetric effect between daytime and nighttime temperature across different vegetation types.

Improving the Quality of MODIS LAI Products by Exploiting Spatiotemporal Correlation Information

The Moderate Resolution Imaging Spectroradiometer (MODIS) Leaf Area Index (LAI) product is critical for global terrestrial carbon monitoring and ecosystem modeling. However, MODIS LAI is calculated on a pixel-by-pixel and day-by-day basis without using spatial or temporal correlation information, which leads to its high sensitivity of LAI to uncertainties in observed reflectance resulting in an increased noise level in time series. While exploiting prior knowledge is a common practice to fill gaps in observations, little research has been conducted on reducing noisy fluctuations and improving the overall quality of the MODIS LAI product. To address this issue, we proposed a Spatio-Temporal Information Composition Algorithm (STICA), which directly introduces prior Spatio-temporal correlation and Multiple Quality Assessment (MQA) information into the existing MODIS LAI product. STICA reduces the noise level and improves the quality of the product while maintaining the original physically-based (Radiative Transfer Model, RTM) LAI production process. In our analysis, the R2 increased from 0.79 to 0.81, and the RMSE decreased from 0.81 to 0.68 compared to the ground-based LAI reference. The improvement was more pronounced with the degradation of the data quality. STICA reduced noisy fluctuations in the LAI time series to varying degrees among eight biome types. In the Amazon Forest, STICA significantly improved the time-series stability of LAI. Moreover, STICA can effectively eliminate abnormal declines in time series and correct for extreme outliers in LAI. We expect that the MODIS LAI Reanalyzed product generated by this method will better support the application of high-quality LAI datasets.

Machine learning models inaccurately predict current and future high-latitude C balances

The high-latitude carbon (C) cycle is a key feedback to the global climate system, yet because of system complexity and data limitations, there is currently disagreement over whether the region is a source or sink of C. Recent advances in big data analytics and computing power have popularized the use of machine learning (ML) algorithms to upscale site measurements of ecosystem processes, and in some cases forecast the response of these processes to climate change. Due to data limitations, however, ML model predictions of these processes are almost never validated with independent datasets. To better understand and characterize the limitations of these methods, we develop an approach to independently evaluate ML upscaling and forecasting. We mimic data-driven upscaling and forecasting efforts by applying ML algorithms to different subsets of regional process-model simulation gridcells, and then test ML performance using the remaining gridcells. In this study, we simulate C fluxes and environmental data across Alaska using ecosys, a process-rich terrestrial ecosystem model, and then apply boosted regression tree ML algorithms to training data configurations that mirror and expand upon existing AmeriFLUX eddy-covariance data availability. We first show that a ML model trained using ecosys outputs from currently-available Alaska AmeriFLUX sites incorrectly predicts that Alaska is presently a modeled net C source. Increased spatial coverage of the training dataset improves ML predictions, halving the bias when 240 modeled sites are used instead of 15. However, even this more accurate ML model incorrectly predicts Alaska C fluxes under 21st century climate change because of changes in atmospheric CO2, litter inputs, and vegetation composition that have impacts on C fluxes which cannot be inferred from the training data. Our results provide key insights to future C flux upscaling efforts and expose the potential for inaccurate ML upscaling and forecasting of high-latitude C cycle dynamics.

Asymmetric responses of soil bacterial community and soil respiration to precipitation changes: A global meta‐analysis

AbstractChanges in the frequency and intensity of precipitation altered the hydrological features, thereafter greatly affecting carbon (C) cycling in terrestrial ecosystems. Soil bacteria are pivotal drivers of the global C cycle in terrestrial ecosystems; however, the responses of soil bacterial communities and the C‐cycle they regulated to precipitation changes remain unclear. Here, we conducted a global meta‐analysis using 98 paired observations from 58 studies to explore the responses of soil bacteria (abundance and diversity) and their C‐related functions (soil respiration) to increased/decreased precipitation. We found that the response of soil bacterial abundance was linearly correlated with precipitation changes, while the response of soil respiration was negatively asymmetric (concave‐down). The relationship among soil bacterial abundance, alpha diversity, and soil respiration was weakened by decreased precipitation, indicating that decreased precipitation exhibited a stronger effect on soil bacterial community and soil respiration. Our study extends the understanding of soil bacteria and their functions in response to precipitation change and facilitates the prediction of soil carbon cycle‐related functions by using soil microorganisms in terrestrial ecosystem models under future precipitation scenarios.

Environmental controls on the light use efficiency of terrestrial gross primary production.

Gross primary production (GPP) by terrestrial ecosystems is a key quantity in the global carbon cycle. The instantaneous controls of leaf-level photosynthesis are well established, but there is still no consensus on the mechanisms by which canopy-level GPP depends on spatial and temporal variation in the environment. The standard model of photosynthesis provides a robust mechanistic representation for C3 species; however, additional assumptions are required to "scale up" from leaf to canopy. As a consequence, competing models make inconsistent predictions about how GPP will respond to continuing environmental change. This problem is addressed here by means of an empirical analysis of the light use efficiency (LUE) of GPP inferred from eddy covariance carbon dioxide flux measurements, in situ measurements of photosynthetically active radiation (PAR), and remotely sensed estimates of the fraction of PAR (fAPAR) absorbed by the vegetation canopy. Focusing on LUE allows potential drivers of GPP to be separated from its overriding dependence on light. GPP data from over 100 sites, collated over 20 years and located in a range of biomes and climate zones, were extracted from the FLUXNET2015 database and combined with remotely sensed fAPAR data to estimate daily LUE. Daytime air temperature, vapor pressure deficit, diffuse fraction of solar radiation, and soil moisture were shown to be salient predictors of LUE in a generalized linear mixed-effects model. The same model design was fitted to site-based LUE estimates generated by 16 terrestrial ecosystem models. The published models showed wide variation in the shape, the strength, and even the sign of the environmental effects on modeled LUE. These findings highlight important model deficiencies and suggest a need to progress beyond simple "goodness of fit" comparisons of inferred and predicted carbon fluxes toward an approach focused on the functional responses of the underlying dependencies.