Bibi S. Naz

Water and carbon exchanges are coupled because carbon dioxide (CO ₂ ) and water vapor loss occur through leaf stomata, yet land surface predictions remain uncertain because vegetation and soil parameters are poorly constrained and often calibrated independently at individual sites. We develop a hierarchical ensemble smoother with multiple data assimilation (H-ES-MDA) for the enCore Community Land Model (eCLM) that pools information across 38 European eddy-covariance sites. The framework estimates shared plant functional type (PFT) vegetation parameters while retaining site-level vegetation deviations and site-specific soil hydraulic parameters. Six assimilation scenarios combined soil moisture (SM), evapotranspiration (ET), net ecosystem exchange (NEE), with performance assessed in independent validation years. Joint assimilation of SM, ET, and NEE produced the most balanced improvements relative to prior simulations, improving median validation performance by 30–46% for SM, 7% for ET, and 18% for NEE, while avoiding the SM degradation observed with flux-only assimilation. Shared PFT posteriors retained substantial NEE skill for grasslands and evergreen needleleaf forests, whereas site-level deviations added little skill. A leave-one-site-out experiment for grasslands confirmed that shared NEE skill persisted at sites excluded from calibration. Transfer was weaker for deciduous broadleaf forest and cropland and did not consistently extend from NEE to unassimilated gross primary production and ecosystem respiration. Kullback–Leibler (KL) divergence and Shapley attribution further showed that strong parameter updating does not necessarily imply predictive skill. Hierarchical multi-site assimilation therefore provides a direct way to test which parameter information can be shared regionally and which processes require local or improved model representation.

Land surface models (LSMs) such as the Community Land Model version 5 (CLM5) are central to climate–carbon assessments but exhibit persistent biases relative to site‐level observations. It remains unclear to what extent model‐data mismatches arise from uncertainty in meteorological forcing, soil properties, or vegetation traits, and to what extent they instead reflect structural deficiencies. We evaluated surface and root‐zone soil moisture (SM, SMr), evapotranspiration (ET), sensible heat flux (H), net ecosystem exchange (NEE), and gross primary production (GPP) at 14 European eddy‐covariance sites using 128‐member CLM5 ensembles in which (a) atmospheric inputs, (b) soil parameters, and (c) vegetation parameters were perturbed. Soil parameters dominate ensemble spread in SM/SMr, whereas vegetation parameters dominate spread in ET, H, NEE, and GPP. Despite this, substantial mean biases remained (0.18 cm ³ cm ⁻³ for SM, −0.22 mm d ⁻¹ for ET, −19.7 W m ⁻² for H, −1.83 gC m ⁻² d ⁻¹ for GPP, and 1.16 gC m ⁻² d ⁻¹ for NEE), and the 99% ensemble envelopes failed to cover many observations. Coverage percentage (CP; fraction of observations within the 99% ensemble envelope) was 74.3% (SM), 74.6% (SMr), 68.8% (ET), 47.7% (H), 65.1% (NEE), and 70.1% (GPP). A companion CLM5‐SP (Satellite Phenology) experiment with prescribed ICOS leaf area index (LAI) confirmed that prognostic LAI errors in CLM5‐BGC (Biogeochemistry) are the primary structural driver of SM/SMr and GPP coverage deficits, while BGC's carbon–nitrogen coupling better captures ET and H variability. CP diagnostics delineate parametric from structural error regimes across variables and plant functional types.

Ecotron facilities are key tools to emulate and monitor environmental conditions for improving the understanding of terrestrial ecosystems. Although their conditions can be precisely modified by cutting-edge controlling systems, the external factors influencing the boundary conditions of such ecotron units remain uncertain under reconstructed scenarios (e.g., climate change). We assessed how an agroecosystem model can be used to simulate the bottom boundary conditions of the newly developed ecotron facility AgraSim. The facility comprises six identical ecotron units composed of a climate-controlled plant chamber on top of a cylindrical lysimeter (1 m2 by 150 cm depth). An automatic suction cup–pumping system installed below each lysimeter is used to control the bottom boundary condition of the system depending on the measured pressure head near its bottom. The first experimental trial of AgraSim is aimed to quantify the key climate responses of a typical agricultural field located in North-Rhine Westphalia (Germany) to transient climate change. A set of four climate scenarios were derived from storyline simulations with the regional atmospheric circulation model ICON, imposing a transient temperature gradient of +1oC to +4oC within the climate chambers. While the atmospheric forcings and the soil texture and hydraulic properties are well-characterized, the bottom boundary condition at 150 cm depth is unknown for the different climate change scenarios. We thus used an agroecosystem model (AgroC) to numerically solve water movement within the soil column (0-150 cm) and investigated the impact of choosing one of the following bottom boundary conditions: fixed pressure heads (FP); free-drainage (FD); and a modified seepage face at h=-100hPa (SP). A simulation that assumes a 500 cm deep soil column and free drainage was taken as a reference for assessing the performance of each of the different bottom boundary conditions. This setup was replicated for two rainfed cropping systems (winter wheat and maize), each parameterized with three types of rooting profile representing shallow, deeper and homogeneous root profiles, respectively. A 1-year spin-up run was performed to minimize the effect of the initial conditions. Our model simulations showed that the different boundary conditions only affected soil moisture below 30 cm, while topsoil moisture was mainly controlled by atmospheric forcings. The FP and FD scenarios tended to underestimate soil water availability in the 150 cm column, especially during critical summer drought periods. The modified SP showed the best agreement with the reference simulations, keeping the soil unsaturated during winter and maintaining moisture and fluxes closer to the reference levels in summer. This result was consistent across both crops and rooting profiles. At the crop level, the different boundary conditions had no significant effect on key crop variables (e.g., biomass and leaf area) and their respective response to the climate scenarios. Although our results are limited by observation availability and model parameterization uncertainty, they demonstrate the potential application of process-based models for decision-making in controlled-system facilities.

Snow plays a key role in land-surface processes by modulating the surface energy balance, soil thermal insulation and water availability. However, the influence of snow on water and energy fluxes in land surface models remains insufficiently understood. To improve simulations of the coupled water–energy cycle, we developed a snow data assimilation (snow-DA) within the Encore Community Land Model coupled to the Parallel Data Assimilation Framework (eCLM-PDAF; https://github.com/HPSCTerrSys/eCLM), enabling assimilation of both snow depth and snow water equivalent (SWE). In the Snow-DA experiment, we assimilate daily snow depth with a one-dimensional Ensemble Kalman Filter (EnKF), updating the liquid and ice SWE components across all snow layers; snow depth is then adjusted through its correlation with SWE. We evaluated the performance of the snow-DA framework by comparing snowpack variables as well as heat fluxes such as latent heat flux (LE), sensible heat flux (SH), ground heat flux (GH), and soil temperature, between data assimilation (DA) and open-loop (OL) simulations at eleven selected Integrated Carbon Observation System (ICOS) sites across Europe. The sites span different observation periods within 2017–2024. Each OL and DA experiment used 100 ensemble members, generated through perturbation of meteorological variables and key snow related parameters. A multiplicative inflation factor of 0.95 and observation error of 0.2m are applied across all sites. Results across ICOS sites showed that DA substantially improved snow variable estimates compared to OL simulations. On average, the root mean square error (RMSE) of SD decreased by 27.3%, and the correlation coefficient (R) increased by 0.06. DA also improved the timing of snow cover duration, yielding a more realistic seasonal snow cover evolution when compared with satellite-based observations. Although overall changes in land-surface heat fluxes were modest, the improved snowpack reduced RMSE during the melt season for LE by 9.5%, evaporative fraction (EF) by 1.6%, and soil temperature by 20.8%. Although the energy balance was evaluated, and LE and EF improved, snow DA degraded the performance of SH and GH at most sites, indicating possible coupling bias between modeled variables for energy partitioning, or representativeness errors between tower-based and modeled fluxes. Overall, this study enhances the representation of snow processes in a land surface model and encourages further research into the modeling of associated water and energy balance mechanisms. Future work will investigate regional responses of water flux components, including runoff, evapotranspiration, and soil water content, to further examine snow data assimilation impacts on water availability.

Droughts are Europe’s costliest natural disasters, with damages estimated at 621 million Euros per event. Changing precipitation patterns and rising atmospheric water demand are increasingly affecting terrestrial ecosystem functioning in Europe, with profound implications for sustainable water resources management and ecosystem carbon uptake. However, responses are driven not only by drought type and severity but also by diverse land surface properties, including soil texture and vegetation functional traits. A clear understanding of how water deficits propagate to inhibit ecosystem functioning is needed to assess drought risk for specific ecosystems under a warming climate. This study uses Community Land Model v5 (CLM5) simulations over Europe from 1960 to 2024 to identify drought events as spatiotemporal clusters and to systematically determine their propagation across hydrological compartments (e.g., from precipitation to root-zone soil moisture) and their impacts on gross primary production (GPP) and transpiration (T). We find that precipitation droughts often propagate into soil moisture droughts, especially during large-scale droughts, such as in the years 1995, 2003, and 2018. However, soil moisture droughts can also emerge even when precipitation deficits are not typically classified as drought events, for example, when vapor pressure droughts increase evaporation over a prolonged period. Further, we compare trends of drought characteristics and show increasing dynamics in the propagation of vapor pressure droughts and increasing severity of soil moisture droughts. These anomalies interact across multiple time scales to drive a wide, though predominantly negative, range of GPP and T responses: Short-term anomalies can already cause significant impacts on dry ecosystems and grasslands, while having only minor effects in humid ecosystems. These results are essential for understanding ecosystem-specific impacts during discrete drought events and for identifying ecosystems whose functioning is under increased risk as drought frequency and severity increase under climate change in Europe, essentially supporting EU Adaptation Strategy and the Water Framework Directive.

Accurate representation of soil and plant hydraulic processes is crucial for predicting forest water and carbon fluxes under variable climate conditions. At the Forest Research sites across Germany, we applied an Ensemble Smoother with Multiple Data Assimilation (ESMDA) coupled with the Community Land Surface Model (eCLM; https://github.com/HPSCTerrSys/eCLM) to optimize soil and vegetation parameters. Site-specific observations included soil moisture, evapotranspiration (ET), and dynamic canopy conductance, with the latter derived from sap-flow measurements. Across all sites, ensemble simulations were performed for 2012 – 2024 using 120 ensemble members in which key parameters controlling soil hydraulics, photosynthesis, stomatal behavior, and plant hydraulics were perturbed.We tested several data assimilation configurations. Assimilating soil moisture alone improved simulated soil water content, reducing RMSE by 5–50% across soil depths, but had limited impact on ET, gross primary production (GPP), or net ecosystem exchange (NEE). In contrast, assimilating both soil moisture and ET further constrained vegetation parameters, resulting in modest improvements in ET, GPP and NEE, while maintaining a large ensemble spread that captures a high percentage of observations. Optimized soil and plant hydraulic parameters also enhanced the representation of seasonal plant water stress, capturing summer stress dynamics more realistically in both wet and dry years, with stronger hydraulic limitation during dry years.These results indicate that correcting soil water availability alone is insufficient to improve plant water use and photosynthesis. Including ET and canopy conductance observations provides additional constraints, strengthening the interactions between soil moisture, transpiration, and carbon uptake. This demonstrates that multi-variable data assimilation is needed to effectively reduce uncertainty in both soil and plant hydraulics, and that direct physiological measurements, such as sap-flow, can further enhance model predictions of both water and carbon fluxes.

Plant hydraulic traits regulate water transport in the soil–plant–atmosphere continuum and mediate the coupling between soil moisture availability, stomatal regulation, and ecosystem carbon uptake. Mechanistic representations of plant hydraulics in land surface models, such as the Community Land Model (CLM), improve the accuracy of simulated vegetation fluxes, particularly under drying soil conditions¹, but they also introduce additional parameters that can be difficult to constrain and can strongly influence model outputs. Global ensemble perturbation experiments in CLM have shown that plant hydraulic parameters are among the most influential controls on evapotranspiration, although their relative importance varies across regions². Yet, how the sensitivity of these parameters varies across plant functional types (PFTs) and seasons remains largely unexplored.In this study, we investigated the sensitivity of simulated vegetation water potential and water and carbon fluxes to five key plant hydraulic parameters, including stomatal behavior (medlyn slope), plant and root conductance (kmax and krmax), cavitation resistance (psi50) and root distribution (β) using eCLM (https://github.com/HPSCTerrSys/eCLM). Ensemble simulations were performed for 13 ICOS sites across Europe, covering four climate zones and five PFTs, over the period 2009–2018. The selected parameters were varied within PFT-dependent ranges following previous perturbation experiments²,³, resulting in a total of 336 ensemble members. Variance-based parameter sensitivities (main effects, two-way interaction effects, and total effects) were quantified using the GEM-SA global sensitivity analysis framework based on Gaussian process emulation4. Emulators were trained on monthly averages for each station and each output variable individually.Across simulations, medlyn slope and kmax showed the strongest effects on simulated water and carbon fluxes (ET, Tr, GPP, NEE) with main effects explaining more than 60% of the variance, while two-way interaction effects contributed only marginally. However, parameter sensitivities varied substantially among PFTs, with distinct patterns in the relative importance of dominant parameters for Mediterranean evergreen broadleaf forests, temperate deciduous forests, and evergreen needleleaf forests. Sensitivities also varied seasonally, with the remaining parameters—particularly psi₅₀—becoming increasingly influential under dry summer conditions. Most notably, seasonal shifts in the direction of parameter effects on canopy transpiration were detected at drought-prone Mediterranean sites: higher medlyn slope increased transpiration during spring, but led to reduced transpiration during summer, reflecting earlier stomatal closure under increasing plant hydraulic stress.Our results show that model sensitivity to plant hydraulic parameters varies across PFTs and seasons, reflecting changes in model behavior across environments. These findings motivate further model development and refinement of plant hydraulic and stomatal process representation to ensure consistent performance across seasons, especially during drought.References 1Kennedy et al. (2019). 10.1029/2018MS0015002Kennedy et al. (2025). 10.1029/2024MS0047153Eloundou et al. (2024). 10.5194/egusphere-egu24-160864O’Hagan (2006). 10.1016/j.ress.2005.11.025

Groundwater serves as a crucial freshwater resource for people and ecosystems, playing a vital role in adapting to climate change. Yet, its availability and dynamics are affected by climate variations, changes in land use, and abstraction. Despite its importance, our understanding of how global change will influence groundwater in the future remains limited. Multi-model ensembles are powerful tools for impact assessments; compared to single-model studies, they provide a more comprehensive understanding of uncertainties and enhance the robustness of projections by capturing a range of possible outcomes. However, to date, no ensemble of groundwater models has been available to assess the impacts of global change. Here, we present the new Groundwater sector within ISIMIP, which combines multiple global, continental, and regional-scale groundwater models. We describe the rationale for the sector, the sectoral output variables that underpinned the modeling protocol, and showcase current model differences and possible future analysis. Currently, eight models are participating in this sector, ranging from gradient-based groundwater models to specialized karst recharge models, each producing up to 19 out of 23 modeling protocol-defined output variables. To showcase the benefits of a joint sector, we utilize available model outputs of the participating models to show the substantial differences in estimating water table depth (global arithmetic mean 6–127 m) and groundwater recharge (global arithmetic mean 78–228 mm yr⁻¹), which is consistent with recent studies on the uncertainty of groundwater models, but with distinct spatial patterns. We further outline synergies with 13 of the 17 existing ISIMIP sectors and specifically discuss those with the global water and water quality sectors. Finally, this paper outlines a vision for ensemble-based groundwater studies that can contribute to a better understanding of the impacts of climate change, land use change, environmental change, and socio-economic change on the world's largest accessible freshwater store – groundwater.