Abstract. The terrestrial carbon cycle plays a critical role in modulating the interactions of climate with the Earth system, but different models often make vastly different predictions of its behavior. Efforts to reduce model uncertainty have commonly focused on model structure, namely by introducing additional processes and increasing structural complexity. However, the extent to which increased structural complexity can directly improve predictive skill is unclear. While adding processes may improve realism, the resulting models are often encumbered by a greater number of poorly determined or over-generalized parameters. To guide efficient model development, here we map the theoretical relationship between model complexity and predictive skill. To do so, we developed 16 structurally distinct carbon cycle models spanning an axis of complexity and incorporated them into a model–data fusion system. We calibrated each model at six globally distributed eddy covariance sites with long observation time series and under 42 data scenarios that resulted in different degrees of parameter uncertainty. For each combination of site, data scenario, and model, we then predicted net ecosystem exchange (NEE) and leaf area index (LAI) for validation against independent local site data. Though the maximum model complexity we evaluated is lower than most traditional terrestrial biosphere models, the complexity range we explored provides universal insight into the inter-relationship between structural uncertainty, parametric uncertainty, and model forecast skill. Specifically, increased complexity only improves forecast skill if parameters are adequately informed (e.g., when NEE observations are used for calibration). Otherwise, increased complexity can degrade skill and an intermediate-complexity model is optimal. This finding remains consistent regardless of whether NEE or LAI is predicted. Our COMPLexity EXperiment (COMPLEX) highlights the importance of robust observation-based parameterization for land surface modeling and suggests that data characterizing net carbon fluxes will be key to improving decadal predictions of high-dimensional terrestrial biosphere models.

Abstract. The terrestrial carbon cycle plays a critical role in modulating the interactions of climate with the Earth system, but different models often make vastly different predictions of its behavior. Efforts to reduce model uncertainty have commonly focused on model structure, namely by introducing additional processes and increasing structural complexity. However, the extent to which increased structural complexity can directly improve predictive skill is unclear. While adding processes may improve realism, the resulting models are often encumbered by a greater number of poorly-determined or over-generalized parameters. To guide efficient model development, here we map the theoretical relationship between model complexity and predictive skill. To do so, we developed 16 structurally distinct carbon cycle models spanning an axis of complexity and incorporated them into a model–data fusion system. We calibrated each model at 6 globally-distributed eddy covariance sites with long observation time series and under 42 data scenarios that resulted in different degrees of parameter uncertainty. For each combination of site, data scenario, and model, we then predicted net ecosystem exchange (NEE) and leaf area index (LAI) for validation against independent local site data. Though the maximum model complexity we evaluated is lower than most traditional terrestrial biosphere models, the complexity range we explored provides universal insight into the inter-relationship between structural uncertainty, parametric uncertainty, and model forecast skill. Specifically, increased complexity only improves forecast skill if parameters are adequately informed (e.g., when NEE observations are used for calibration). Otherwise, increased complexity can degrade skill and an intermediate-complexity model is optimal. This finding remains consistent regardless of whether NEE or LAI is predicted. Our COMPLexity EXperiment (COMPLEX) highlights the importance of robust, observation-based parameterization for land surface modeling and suggests that data characterizing net carbon fluxes will be key to improving decadal predictions of high-dimensional terrestrial biosphere models.

Abstract. The capacity of Amazon forests to sequester carbon is threatened by climate-change-induced shifts in precipitation patterns. However, the relative importance of plant physiology, ecosystem structure and trait composition responses in determining variation in gross primary productivity (GPP) remain largely unquantified and vary among models. We evaluate the relative importance of key climate constraints to GPP, comparing direct plant physiological responses to water availability and indirect structural and trait responses (via changes to leaf area index (LAI), roots and photosynthetic capacity). To separate these factors we combined the soil–plant–atmosphere model with forcing and observational data from seven intensively studied forest plots along an Amazon drought stress gradient. We also used machine learning to evaluate the relative importance of individual climate factors across sites. Our model experiments showed that variation in LAI was the principal driver of differences in GPP across the gradient, accounting for 33 % of observed variation. Differences in photosynthetic capacity (Vcmax and Jmax) accounted for 21 % of variance, and climate (which included physiological responses) accounted for 16 %. Sensitivity to differences in climate was highest where a shallow rooting depth was coupled with a high LAI. On sub-annual timescales, the relative importance of LAI in driving GPP increased with drought stress (R2=0.72), coincident with the decreased importance of solar radiation (R2=0.90). Given the role of LAI in driving GPP across Amazon forests, improved mapping of canopy dynamics is critical, opportunities for which are offered by new satellite-based remote sensing missions such as GEDI, Sentinel and FLEX.

Sugarcane production supports the livelihoods of millions of small‐scale farmers in developing countries, and the bioenergy needs of millions of consumers. Yet, future sugarcane yields remain uncertain due to differences in climate projections, and because the sensitivity of sugarcane ecophysiology to individual climate drivers (i.e. temperature, precipitation, shortwave radiation, VPD and CO2) and their interactions is largely unresolved. Here we ask: how sensitive is sugarcane yield to future climate change, including climate extremes, and what are its key climate drivers? We combine the Soil‐Plant‐Atmosphere model with detailed time‐series measurements from experimental plots in Guangxi, China, and São Paulo State, Brazil. We first calibrated and validated modelled carbon and water cycling against field flux and biometric data. Second, we simulated sugarcane growth under the historical climate (1980–2018), and six future climate projections (2015–2100). We computed the ‘yield‐effect’ of each climate driver by generating synthetic climate forcings in which the driver time series was alternated to that of the historical median. In Guangxi, median yield and yield lows (i.e. lower decile) were relatively insensitive to forecast climate change. In São Paulo, median yield and yield lows decreased under all future climates projections (truex¯ = −4% and −12% respectively). At Guangxi, where moisture stress was low, radiation was the principal driver of yield variability (yield‐effect truex¯ = −1.2%). Conversely, high moisture stress at São Paulo raised yield sensitivity to temperature (yield‐effect truex¯ = −7.9%). In contrast, a number of other modelling studies report a positive effect of increased temperatures on sugarcane yield. We ascribe the disparity between model predictions to the representation of key phenological processes, including the link between leaf ageing and thermal time, and the role of ageing in driving leaf senescence. We highlight climate sensitivity of phenological processes as a key focus for future research efforts.

Leaf area is a key parameter underpinning ecosystem carbon, water and energy exchanges via photosynthesis, transpiration and absorption of radiation, from local to global scales. Satellite-based Earth Observation (EO) can provide estimates of leaf area index (LAI) with global coverage and high temporal frequency. However, the error and bias contained within these EO products and their variation in time and across spatial resolutions remain poorly understood. Here, we used nearly 8000 in situ measurements of LAI from six forest environments in southern China to evaluate the magnitude, uncertainty, and dynamics of three widely used EO LAI products. The finer spatial resolution GEOV3 PROBA-V 300 m LAI product best estimates the observed LAI from a multi-site dataset (R2 = 0.45, bias = −0.54 m2 m−2, RMSE = 1.21 m2 m−2) and importantly captures canopy dynamics well, including the amplitude and phase. The GEOV2 PROBA-V 1 km LAI product performed the next best (R2 = 0.36, bias = −2.04 m2 m−2, RMSE = 2.32 m2 m−2) followed by MODIS 500 m LAI (R2 = 0.20, bias = −1.47 m2 m−2, RMSE = 2.29 m2 m−2). The MODIS 500 m product did not capture the temporal dynamics observed in situ across southern China. The uncertainties estimated by each of the EO products are substantially smaller (3–5 times) than the observed bias for EO products against in situ measurements. Thus, reported product uncertainties are substantially underestimated and do not fully account for their total uncertainty. Overall, our analysis indicates that both the retrieval algorithm and spatial resolution play an important role in accurately estimating LAI for the dense canopy forests in Southern China. When constraining models of the carbon cycle and other ecosystem processes are run, studies should assume that current EO product LAI uncertainty estimates underestimate their true uncertainty value.

Leaf area index (LAI) underpins terrestrial ecosystem functioning, yet our ability to predict LAI remains limited. Across Amazon forests, mean LAI, LAI seasonal dynamics and leaf traits vary with soil moisture stress. We hypothesise that LAI variation can be predicted via an optimality‐based approach, using net canopy C export (NCE, photosynthesis minus the C cost of leaf growth and maintenance) as a fitness proxy. We applied a process‐based terrestrial ecosystem model to seven plots across a moisture stress gradient with detailed in situ measurements, to determine nominal plant C budgets. For each plot, we then compared observations and simulations of the nominal (i.e. observed) C budget to simulations of alternative, experimental budgets. Experimental budgets were generated by forcing the model with synthetic LAI timeseries (across a range of mean LAI and LAI seasonality) and different leaf trait combinations (leaf mass per unit area, lifespan, photosynthetic capacity and respiration rate) operating along the leaf economic spectrum. Observed mean LAI and LAI seasonality across the soil moisture stress gradient maximised NCE, and were therefore consistent with optimality‐based predictions. Yet, the predictive power of an optimality‐based approach was limited due to the asymptotic response of simulated NCE to mean LAI and LAI seasonality. Leaf traits fundamentally shaped the C budget, determining simulated optimal LAI and total NCE. Long‐lived leaves with lower maximum photosynthetic capacity maximised simulated NCE under aseasonal high mean LAI, with the reverse found for short‐lived leaves and higher maximum photosynthetic capacity. The simulated leaf trait LAI trade‐offs were consistent with observed distributions. We suggest that a range of LAI strategies could be equally economically viable at local level, though we note several ecological limitations to this interpretation (e.g. between‐plant competition). In addition, we show how leaf trait trade‐offs enable divergence in canopy strategies. Our results also allow an assessment of the usefulness of optimality‐based approaches in simulating primary tropical forest functioning, evaluated against in situ data.

<p><strong>Abstract.</strong> The capacity of Amazon forests to sequester carbon is threatened by climate change-induced shifts in precipitation patterns. However, the relative importance of plant physiology, ecosystem structure, and trait composition responses in determining variation in GPP, remain largely unquantified, and vary among models. We evaluate the relative importance of key climate constraints to gross primary productivity (GPP), comparing direct plant physiological responses to water availability and indirect structural and trait responses (via changes to leaf area index (LAI), roots and photosynthetic capacity). To separate these factors we combined the Soil-Plant-Atmosphere model with forcing and observational data from seven intensively studied forest plots along an Amazon soil moisture-stress gradient. We also used machine learning to evaluate the relative importance of individual climate factors across sites. Our model experiments showed that variation in LAI was the principal driver of differences in GPP across the gradient, accounting for 33 % of observed variation. Differences in photosynthetic capacity (V<sub>cmax</sub> and J<sub>max</sub>) accounted for 21 % of variance, and climate (which included physiological responses) accounted for 16 %. Sensitivity to differences in climate was highest where shallow rooting depth was coupled with high LAI. On sub-annual timescales, the relative importance of LAI in driving GPP increased with soil moisture-stress (R<sup>2</sup> = 0.72), whilst the importance of solar radiation decreased (R<sup>2</sup> = 0.90). Given the role of LAI in driving GPP across Amazon forests, improved mapping of canopy dynamics is critical, opportunities for which are offered by new satellite-based remote sensing missions such as GEDI, Sentinel and FLEX.</p>

Efforts to abate climate change heavily rely on carbon sequestration by trees. However, analyses of tree carbon dynamics often neglect trees outside of forests (TOF) and spatially detailed information about tree carbon sequestration rates are largely missing. Here we describe a new method which combines remote sensing with forest inventory data from 127,358 sites to first estimate tree age and site productivity, which we then used to estimate carbon storage and sequestration rates for all trees inside and outside forests across Great Britain. Our models estimate carbon storage and sequestration rates with R2 values of 0.86 and 0.56 (RMSEs of 70 tCO2e ha-1 and 3.4 tCO2e ha-1 yr-1). They also reveal the important finding that 17 % (165.6 MtCO2e) of the total carbon storage and 21% (3.4 MtCO2e yr-1) of the total carbon sequestration rate of all trees in Great Britain come from TOF, with particularly high contributions in England (24.3 % and 34.1 %), followed by Wales (12.5 % and 17.6 %) and Scotland (2.6 % and 1.8 %). Future estimates of carbon status and fluxes need to account for the significant contributions of TOF because these trees, often found in field margins and hedgerows are potentially an important carbon offset. Our novel approach enables carbon baseline assessments against which changes can be assessed at management relevant scales, improving the means to measure progress towards net zero emissions targets and associated environmental policies.