Validation gaps in Earth Science

Representative open questions

Showing 30 of 138 — one per source paper, highest-quality first.

• G6-1.5K-MCB: Marine Cloud Brightening scenario design for the Geoengineering Model Intercomparison Project (GeoMIP) in CESM2.1, E3SMv2.0, and UKESM1.1 (2026) · doi

  As CMIP transitions to the seventh iteration with updated ESMs and revised emission trajectories, the G6-1.5K-MCB scenario design choices must be evaluated and potentially optimized for compatibility with next-generation Earth system models to define the G7-1.5K-MCB scenario.

• A decision tree to guide precautionary and adaptive tiering in marine nursery identification and management (2026) · doi

  The decision tree framework for marine nursery identification lacks empirical validation across different ecoregions and fish life histories. Specific testing is needed to determine how precautionary and adaptive tiering thresholds should be adjusted for species with different ontogenetic habitat shifts, particularly for elasmobranchs and demersal fish with poorly documented juvenile recruitment dynamics.

• The 2022 hydraulic stimulation at Utah FORGE: investigating fracturing mechanisms and testing forecasting approaches (2026) · doi

  The study demonstrates that stress field, hydraulic parameters, and spatio-temporal seismicity distribution must be jointly compared to constrain triggering mechanisms, but the methodology for integrating these three datasets in real-time ATLP decision-making requires further development and validation across multiple EGS projects.

• Identifying regions that can constrain anthropogenic Hg emissions uncertainties through modelling (2026) · doi

  The paper identifies mercury measurement stations in specific regions (Antarctic, Arctic, Cape Point) but does not provide detailed validation of modeled anthropogenic Hg emissions against observational data from artisanal and small-scale gold mining (ASGM) regions. Direct comparison of model predictions with field measurements in ASGM-dominant areas is needed to constrain emission uncertainties in these high-uncertainty source sectors.

• Human productive capacities, institutions, and environmental sustainability in Sub-Saharan Africa (2026) · doi

  The KRLS machine learning estimator and MEDSEM mediation analysis were validated only through fixed-effects regression robustness checks; alternative nonlinear methodologies should be tested to confirm whether the identified nonlinearity in human productive capacity effects on climate action and ecological reserves is robust across different estimators.

• Assessing the outdoor thermal comfort impact of nature-based urban interventions in dense informal settlement upgrading processes: the case of Barrio 20 in Buenos Aires (2026) · doi

  The study employs a scenario-based simulation approach (baseline, current post-intervention, desired future scenario S3) to reveal opportunity costs of regulatory constraints, but does not validate these simulated thermal outcomes against empirical microclimatic monitoring data collected in Barrio 20 since July 2023, limiting verification of model accuracy for dense informal contexts.

• Validation of TROPOMI and WRF-Chem NO 2 across seasons using SWING+ and surface observations over Bucharest (2026) · doi

  Dedicated seasonal validation campaigns for TROPOMI NO2 tropospheric columns are critically lacking, particularly for winter and spring over polluted urban areas. The literature review reveals validation studies are concentrated in summer (aircraft) and fall (MAX-DOAS), limiting comparability across seasons and product versions (v2.1–v2.4).

• Revisiting the global budget of atmospheric glyoxal: updates on terrestrial and marine precursor emissions, chemistry, and impacts on atmospheric oxidation capacity (2026) · doi

  The relationship between glyoxal formation yields from isoprene oxidation and formaldehyde production in atmospheric models needs validation using aircraft observations (e.g., ATom, SENEX campaigns) under varying atmospheric conditions, particularly in the remote free troposphere where OH radical concentrations differ significantly from continental boundary layer regions.

• Global atmospheric methanol emissions inferred from IASI satellite measurements and aircraft data (2026) · doi

  The snow fraction calculation (Eq. A16) uses ERA5 snow depth with a maximum threshold (SDmax) dependent on LAI, but the paper does not validate how this simplified approach represents methanol deposition over snow-covered surfaces or discuss whether snow-specific methanol uptake coefficients differ from those for other surfaces.

• Dynamics of island mass effect – Part 2: Phytoplankton physiological responses (2026) · doi

  The artificial sharp increase in Ωsat across the outbound transect near the Society Islands was attributed to anomalies in normalized fluorescence line height (nFLH) or photosynthetically available radiation (iPAR) estimation from satellite products; improved validation and calibration of these satellite algorithms for coastal island environments is needed to accurately resolve iron stress responses.

• CMIP7 Data Request: atmosphere priorities and opportunities (2026) · doi

  The NOAA heat index calculation at 2 metres relies on conversion between Fahrenheit and the model's native temperature units, but no systematic evaluation has been conducted on how temperature unit conversion errors propagate through the heat index formula or affect inter-model comparisons of daily maximum heat index fields.

• Remote sensing image enhancement and water eutrophication prediction based on atmospheric-water multimodal information fusion (2026) · doi

  Although the Att_Conv2D_LSTM_Cross_GAN achieves 6.51% improvement over Att_Conv2D_LSTM_GAN, the paper does not investigate whether this performance gain is statistically significant or whether confidence intervals can be established for the SSIM, PSNR, and COSIN metrics across multiple independent runs or cross-validation folds.

• Transport of volcanic aerosol from the Raikoke eruption in 2019 through the Northern Hemisphere (2026) · doi

  The study validates CLaMS simulations against balloon-borne measurements at only two specific locations (Lhasa and Boulder); the model's ability to reproduce enhanced BSR455 signals and dilution processes driven by upwelling and lateral mixing needs validation across additional geographic regions and multiple eruption events beyond the 2019 Raikoke case.

• Fracture Complexity and Seismicity Influenced by Mineral Heterogeneity and Fluid Viscosity in Hydraulic Fracturing of Simulated Crystalline Rock (2026) · doi

  The maximum-likelihood b-value estimation method (Equation 22) uses a fixed binning width (ΔMbin = 0.1) based on previous work, but the sensitivity of magnitude–frequency distribution statistics to this binning parameter across different fluid viscosity regimes and heterogeneous mineral compositions has not been tested, limiting the robustness of viscosity-induced b-value comparisons.

• Regionally important developmental foraging grounds for green turtles and hawksbills in the Southwest Indian Ocean (2026) · doi

  Citizen science contributions to sea turtle monitoring should be restricted to supplementary uses (individual identifications, resightings, spatial distribution data) rather than abundance estimates, as untimed search effort and selective reporting of sighting days (not zero-sighting days) creates systematic overestimation of sightings per unit effort (SPUE) in neritic foraging habitats.

• Potential of micro glass/nano glass/fibreglass as pollutants species in environmental matrices (2026) · doi

  SEM-EDS analysis of micro-glass and nano-glass particles is affected by matrix effects and particle heterogeneity that compromise semi-quantitative elemental composition accuracy. Systematic investigation of matrix effect corrections and heterogeneity characterization protocols for engineered glass particles in complex environmental matrices is absent.

• A garden of forking paths: an ecosystem-based fisheries management case study of Florida red tides (2026) · doi

  The effectiveness of management actions implemented through the Fishery Ecosystem Initiative loop requires peer review and reassessment, but no standardized protocol for evaluating outcomes of harvest control rules or other HAB-related management responses in real-time has been established for application in future red tide events.

• Unraveling non-monotonic responses of the El Niño–Southern Oscillation to post-2100 global warming (2026) · doi

  The recharge-oscillator (RO) model analysis was limited to MIROC-ES2L simulations with extended output through 2300; validation of the RO diagnostic framework parameters (Bjerknes feedback R, ocean damping ε, discharge-recharge efficiency F1 and F2) across the full ensemble of 12 CMIP5/CMIP6 models for post-2100 scenarios remains incomplete, particularly for quantifying how these dynamical parameters shift under non-monotonic ENSO amplitude responses.

• Linking hosts, landscapes, and climate to advance zoonotic arbovirus forecasting (2026) · doi

  The study selected 12 specific bird species as suspected EEEV hosts based on prior Culicoides melanura blood meal analyses, but did not validate whether the absence of other potential avian hosts in the correlation analysis affects the model's ability to explain observed EEEV seroprevalence patterns across Florida.

• The interplay of habitat quality and temperature shape demographic patterns of mule deer (Odocoileus hemionus) in North America (2026) · doi

  The leave-one-out cross-validation (LOO) model selection approach was used to identify best-fitting covariate combinations, but the study did not present predictive performance validation on geographically independent mule deer populations or herd units outside the Wyoming study area to assess model generalizability across broader North American contexts.

• Marine chemistry special issue: Aerosol chemistry and impacts on the ocean (2010) · doi

  The confounding effects of multiple environmental factors (light, silicic acid, phosphate, iron availability) and aerosol elemental stoichiometry on biological responses to aerosol iron deposition need quantitative separation. Boyd et al. (2010) note these factors limit observed biological responses, but experimental designs isolating individual nutrient limitation effects during aerosol deposition events are underspecified.

• Environmental chemistry is the importance of chemistry to the environment (2014) · doi

  The allelopathic and cytoprotective effects of E. jambolana extract on Lactuca sativa growth under mercury chloride exposure have been documented at laboratory scale, but field-level validation under realistic soil-water-plant mercury contamination scenarios remains absent.

• BEYOND OIL: AZERBAIJAN’S CLIMATE ACTION IN PRACTICE AND ITS EMERGING LOW-CARBON PATH (2026) · doi

  The paper establishes that the 2021 renewable energy and energy efficiency laws create a 'governance and financing foundation,' but does not assess the effectiveness or compliance rate of the Energy Efficiency Fund (established March 2024) through measurable KPIs, disbursement rates, or outcome metrics tied to the 2030 renewables objective.

• Atmospheric moisture transport anomalies and vegetation response in arid coastal ecosystems: insights from the 2017 coastal El Niño in northern Peru (2026) · doi

  The persistence and ecological significance of transient moisture anomalies created by extreme events like the 2017 coastal El Niño for vegetation recovery in degraded arid systems remain uncertain; long-term monitoring studies combining atmospheric moisture transport observations with vegetation resilience metrics are needed to quantify ecological restoration potential.

• Earthquake Induced Landslide Identification to Support Landuse Planning in Rapid Growing Settlements Area of Imogiri and Dlingo Sub-District, Yogyakarta (2026) · doi

  The Ca-Markov settlement expansion model was calibrated on 2013-2022 transition patterns, but the study does not validate whether this longer-term calibration remains accurate for earthquake-induced landslide susceptibility when settlements encroach into the identified high-risk zones (1.93 km² in Wukirsari). Quantitative assessment of how settlement proximity to steep slopes (>800 m/km²) and weak Semilir Formation lithology affects actual landslide triggering during seismic events (238.222 gal, VII MMI) is missing.

• From Installation to Preservation: An Inspection-Based Case Study of Drainage Vulnerability During Active Construction (2026) · doi

  The inspection-based case study documents responses to drainage exposure and sediment entry at one active construction site, but does not establish whether the corrective response patterns (re-covering, detail correction, area-wide reinforcement) generalize across different site conditions, drainage configurations, or construction phases. Comparative multi-site inspection data is needed to validate the applicability of these response categories.

• Machine learning driven land surface temperature prediction and urban heat risk assessment in the Gambia (2026) · doi

  The 2040 LST and urban heat island projections are scenario-based estimates derived from projected LULC maps, but no sensitivity analysis quantifies how uncertainty in land-use change scenarios propagates through the XGBoost model to affect the reliability of heat stress predictions for policy planning.

• Atmospheric Simulation Chambers in the ACTRIS Research Infrastructure (2026) · doi

  Water vapor measurement techniques were intercompared in AquaVIT-1 (Fahey et al., 2014), but equivalent systematic intercomparison exercises for humidity measurement accuracy and performance across the ACTRIS atmospheric simulation chamber network—which require humidified aerosol experiments under controlled conditions—have not been documented.

• Coal seam porosity and water resources estimation based on big data analysis and deep learning models: A case study in the Surat Basin, Australia (2026) · doi

  Fracture length in the DFN is assumed as 'two times the bed thickness' and visualization fractures near wellbores assume 500m length and 50m height, but these assumptions are not validated against the mapped 69 faults from the mine database or the 919 image log-interpreted fractures. No sensitivity analysis is presented on how fracture length assumptions affect calculated water resources estimation or porosity predictions.

• Comprehensive study of anthropogenically unstable territory around the Rivne NPP, Ukraine (2026) · doi

  The study validates the GIS framework and DInSAR methodology for the Rivne NPP over a 2022-2025 period, but lacks comparison with ground truth data from the comprehensive geodetic stations network mentioned in the text. Quantitative validation analysis comparing InSAR-derived vertical displacements against GNSS observations from the geodynamic polygon stations should be conducted to establish measurement uncertainty bounds.