Earth and Environmental Sciences

Thin clouds substantially attenuate spectral signals in satellite imagery, leading to systematic underestimation of macroalgal bloom coverage. Existing thin-cloud correction methods have critical limitations for dynamically drifting blooms: they typically require paired cloudy-clear observations or stable surface conditions, which are rarely available, or inadequately account for cloud-induced attenuation, particularly in the near-infrared band that is essential for algal quantification. This study proposes a single-scene thin-cloud correction framework for dynamic macroalgal blooms that removes both additive scattering and multiplicative attenuation from visible to NIR bands. By exploiting local atmospheric consistency between algal patches and adjacent water, the method infers localized cloud effects from neighboring water pixels. Cloud-free and cloud-affected water pixels are first identified to estimate clear-sky and cloudy top-of-atmosphere reflectance. A linear conversion model derived from radiative transfer theory then relates these two reflectances through coefficients C T and C R , whose relationship is constrained by a MODTRAN-based lookup table (LUT). The coefficients are solved by combining the conversion model with a quadratic polynomial relationship between C T and C R established from the LUT. Validation using GaoFen‑1 and Landsat‑8 imagery against near-concurrent Sentinel‑2 observations over the Southern Yellow Sea demonstrates effective near-infrared restoration. Using two complementary validation strategies—macroalgal area estimation from real images and reflectance evaluation from simulated cloudy data—the proposed method achieves consistently higher accuracy than representative state‑of‑the‑art methods. Application to golden tide events in the East China Sea further confirms its robustness and transferability. The method provides a practical, physics-guided solution for monitoring dynamic macroalgal blooms under thin-clouds contamination.

Abstract Recent works suggest there are periods when the Sun encountered massive interstellar cold clouds which compressed the heliosphere to within Earth's orbit, exposing Earth to interstellar galactic cosmic rays and energetic particles of heliospheric origin. We model 10 Be production in Earth's atmosphere during possible interstellar cloud encounters and supernovae. We find that, if the heliosphere is compressed to 0.2 AU ( AU), the 10 Be production rate is elevated above the background. To be distinguished in marine sediments (iron‐manganese crusts) beyond terrestrial variability, this signal must be sustained for Myr ( Myr). A 0.7 AU ( AU) compression increases the 10 Be production rate above the background, which may be detectable if the event lasts Myr. We find that the 10 Be peak observed by Koll et al. (2025, https://doi.org/10.1038/s41467‐024‐55662‐4 ) 10 Ma is too prolonged to be produced by a supernova, but could be attributed to an interstellar cloud crossing.

Abstract Phenological shifts represent one of the most evident biological responses to climate change. This study used the two-band Enhanced Vegetation Index derived from Moderate Resolution Imaging Spectroradiometer satellite data to analyze changes in the growing seasons of arable land, broad-leaved forest, coniferous forest, and grassland across the central European region from 2000 to 2022. Phenological metrics included the start, end, and length of the growing season. Nonparametric Theil–Sen regression indicated an earlier start of the season (median of trends for all land cover classes ranging from -9.3 to -17.5 days per decade), a later end of the season (ranging from +8.1 to +11.4 days per decade), and consequently longer length of the season (ranging from +13.3 to +25.0 days per decade), with advancing start of the season identified as the primary driver of length of the season prolongation. The most pronounced changes across all phenological metrics were associated with arable land. Elevation zone-based analysis showed that the highest median of trends of start and length of the season was found for arable land and coniferous forest in elevations below 600 m a.s.l., with values decreasing progressively with increasing elevation. The analysis of environmental zones revealed a higher median of trends of the start and length of the season in cool-moist and warm-mesic zones, while end-of-season-related changes were more pronounced in cold-wet zones. This study demonstrates coherent long-term trends in vegetation phenology across different land cover classes, with stronger acceleration over arable land, highlighting potential implications for agricultural systems.

Marine Isotope Stage (MIS) 3, the period from 57 to 30 thousand years ago (ka), is critical to understand how glaciers and ice sheets advanced into the Last Glacial Maximum (LGM; 30–17 ka). Previous estimates of ice-volume changes and their equivalent global mean sea level (GMSL) during MIS 3 remain uncertain, particularly regarding the extent of Northern Hemisphere ice-sheet reduction and exact GMSL values during this period. Here we present a new GMSL record from late MIS 3 to the LGM (ca. 39–18 ka) inferred from paleo-water depth estimates of photosymbiotic large benthic foraminiferal assemblages in upper slope sediment cores from the Great Barrier Reef. The relative sea-level-based GMSL record reconstructed using a Bayesian statistical model, revealed long-term mean trends characterized by incrementally lowering sea levels divided into five periods. Late MIS 3 GMSL decreased from –101 m (average with a 95% confidence interval (CI) of –89 to –113 m) at ca. 39 ka to –119 m (average with a 95% CI of –114 to –124 m) at ca. 30 ka, then finally reached down to –137 m (average with a 95% CI of –131 to –142 m) during the LGM. A relatively steep sea-level fall into the LGM—approximately 17 m in magnitude—began at ca. 35 ka and took ca. 7000 years, implying an acceleration of large-scale ice sheet advance into the LGM.

Abstract During a 2023 field campaign, a NASA ER‐2 research aircraft intercepted an upward jet‐like electrical discharge at 20 km altitude over an intense tropical thunderstorm. This event was analyzed using airborne multispectral optical sensors, electric field mills, X‐band radar, passive microwave, and ground‐based radio networks. Optical observations revealed a streamer‐dominated discharge characterized by strong ultraviolet (∼340 nm) emissions and limited thermalization, explaining the lack of airframe damage. Radio observations indicated that the discharge was initiated by a narrow bipolar event near cloud top. Electric field measurements found that fields due to charge on the aircraft were close to dielectric breakdown threshold immediately prior to the event. Radar data showed the aircraft was navigating between intense convective cells when hit, while passive microwave indicated vertically aligned ice crystals due to electric fields. This is likely the first documented instance of an aircraft intercepting an upward‐propagating transient luminous event from cloud top.

Plant functional traits provide critical indicators of ecosystem functioning, biodiversity maintenance, and plant adaptive strategies under environmental change. However, hyperspectral retrieval of mangrove functional traits remains challenging because trait-related spectral signals are strongly influenced by species composition, canopy structural heterogeneity, and intertidal background effects, leading to unstable trait-spectrum relationships across species and scales. To address these limitations, this study developed a genetic algorithm-optimized extreme learning machine (GA-ELM) framework for robust retrieval of mangrove functional traits under heterogeneous mangrove spectral conditions. Three key traits, including chlorophyll content (Cab), equivalent water thickness (EWT), and leaf mass per area (LMA), were retrieved for four dominant mangrove species using leaf-level hyperspectral measurements and canopy-scale GF-5 imagery. GA-ELM generally outperformed PLSR, RF, and conventional ELM in species-specific comparisons. Leave-one-species-out validation yielded R 2 ranges of 0.61–0.75 for Cab, 0.60–0.72 for EWT, and 0.66–0.76 for LMA, indicating moderate predictive capability for previously unseen species, with performance varying among species and traits. At the canopy scale, GA-ELM achieved stable estimation accuracy for Cab, EWT, and LMA based on GF-5 imagery ( R 2 > 0.75), demonstrating that the framework can be independently calibrated for canopy-scale retrieval. Independent cross-regional validation yielded R 2 values of 0.60, 0.49, and 0.44 for Cab, EWT, and LMA, respectively, indicating moderate but reduced cross-regional transferability. Spectral sensitivity and band-selection analyses revealed species- and trait-dependent patterns that provided context for interpreting variation in predictive performance. These results remain conditional on one GF-5 acquisition per site and 30 m mixed-pixel conditions, and broader tidal-stage and cross-sensor validation is still required. These findings demonstrate the potential of GA-ELM for hyperspectral retrieval and spatial mapping of mangrove functional traits.

Abstract Using satellite‐derived dust optical depth (DOD), reanalysis meteorology, and a global topsoil flash drought (TSFD) inventory for 2003–2022, we find that TSFD events are associated with increased dust loading at the global scale. During TSFD events, 44.7% of land pixels exhibited positive DOD anomalies. The net global DOD anomaly increased significantly over the study period; however, this trend cannot be attributed solely to TSFD‐related effects, as changes in background environmental conditions may also have contributed. Compound TSFD–strong dust (TSFD_SD) events occurred most frequently under sparse vegetation and moderately elevated wind conditions. SHapley Additive exPlanation (SHAP) analysis showed that DOD variability during TSFD events was more strongly associated with soil moisture, background aridity, and vegetation state than with wind speed. These relationships are consistent with a mechanism in which rapid topsoil drying and vegetation stress increase surface erodibility and may enhance dust loading, although the observational analyses do not establish causality.

The past decade has seen rapid advancements in satellite remote sensing data availability and analysis. In coastal environments, this has enabled regional to global assessments of shoreline trends, a key practical application for management and planning. However, coastal zones are inherently dynamic, while instantaneous shorelines only capture the state of the system at a particular moment in time. This study introduces a novel approach for monitoring dynamic coastal environments using satellite imagery, adapting the principle from ground-based video monitoring systems. The SatVar approach presented here uses freely available, medium resolution satellite imagery and Google Earth Engine. The method takes advantage of the increasing availability of cost-free Sentinel-2 imagery and the flexibility of collections acquired over different time periods to automatically detect highly dynamic zones through variance composites, image compositing and segmentation to obtain medium-term trends of coastal dynamics. Natural boundaries are rarely abrupt, so a zone-based (instead of a line-based) approach provides more nuanced information about transitional zones. The identified highly dynamic zones correspond to morphological processes described in literature, and are also used for shoreline extraction. This combined information can enrich understanding of coastal dynamics and overcome some of the limitations of instantaneous shorelines. The approach is tested at three locations with different morphodynamic characteristics and tidal regimes: Sancti Petri (Spain, mesotidal), Narrabeen (Australia, microtidal) and Sefton (UK, macrotidal). Potential applications to monitor changes outside the upper shoreface such as coastal dune dynamics or back-barrier environments are also discussed.

Abstract We investigate ion velocity‐space features within the exhaust of magnetopause reconnection using a global hybrid‐Vlasov simulation. The Hermite transform and Gaussian Mixture Model (GMM) are applied to quantify the complexity of velocity‐space structures that arise during the mixing of magnetospheric and magnetosheath ion populations. Using the Hermite transform, we calculated the enstrophy metric to quantify available free energy. From the GMM multi‐beam decomposition we obtained pseudothermal energy. We find that pseudothermal energy appeared due directed beam motion accounts for nearly half of the apparent thermal energy within the exhaust. The enstrophy evolves similarly to the pseudothermal energy estimate selected via the optimal number of Gaussian components. These results revise ion energy partitioning during magnetopause reconnection and suggest an approach for defining thermal energy in non‐Maxwellian plasmas.

Agriculture suffers from more intense, prolonged, and frequent extreme temperature events such as heat waves, droughts, and frosts. Agrivoltaic systems offer an opportunity to mitigate such events by deploying photovoltaic (PV) panels above crops. In this study, measurements of plant, air, and soil temperatures, along with solar and infrared radiation, wind speed, and soil moisture demonstrate that plants benefit from cooler microclimates beneath PV panels during extreme heat events, and conversely from warmer microclimates during cold nights. More precisely, during the 2025 heat wave in France, plant temperatures beneath PV panels were observed to be up to 12 °C lower than the 45 °C recorded ones in a control zone without PV panels. Additionally, during radiation-driven spring nights, PV panels helped maintain positive plant surface temperatures. In contrast, plant surface temperatures dropped to -3 °C in the control zone, reaching the threshold at which irreversible crop damage typically begins. Quantifying the impact of PV panels on agricultural yield requires a multi-scale temporal analysis that captures their instantaneous effects on the microclimate, the occurrence of sub-daily extreme events, and their influence on crop growth throughout the season. To this end, the Farquhar photosynthesis model, using the experimental data as inputs, and the STICS crop growth model were both evaluated under stressed and non-stressed conditions. The results suggest that when extreme temperatures occur the PV-induced loss in solar radiation can be turned into a gain in agricultural yield.

Abstract Climate change poses an increasing threat to agricultural production and food security. While many European countries have developed National Adaptation Strategies (NASs) and National Adaptation Plans (NAPs), the alignment between proposed policies and their supporting mechanisms remains under-explored. This study maps adaptation frameworks across ten European countries—Finland, Austria, Switzerland, Belgium, the Netherlands, Denmark, Estonia, Spain, Latvia, and Lithuania—to clarify diverse national policies and identify how specific instruments are prioritized to address adaptation goals. Drawing upon 20 official documents available on the Climate-ADAPT platform and using ATLAS.ti for qualitative analysis, 131 climate adaptation policies in agriculture were extracted to form a thematic synthesis. Then, we determined which policy instruments (categorized into four groups: information-based, legal and regulatory, economic, and infrastructural and physical instruments) were associated with each policy. The findings indicate that while these frameworks provide strategic guidance, they place significant emphasis on information-based and awareness-raising measures. This suggests a distinct pattern in European climate governance: a preference for farmer engagement and voluntary action over more direct regulatory or economic interventions.

Abstract Earthquakes are notoriously irregular, yet occasionally neighboring faults rupture in close succession, repeatedly, as if synchronized. Over successive cycles, differences in recurrence intervals, overall irregularity and external stressors should all drive neighboring faults out of phase. We propose that synchronization emerges only when fault interaction is phase dependent and exceeds the accumulated misalignment. The competing effects are quantifiable from the geometry of neighboring faults, yielding a conditional existence prediction for synchronization that we test first against simulations and then observations. Among repeating earthquake families in the central San Andreas Fault, the overall prevalence of synchronized pairs decays rapidly with increasing separation, tracking their stress transfer. Among neighboring megathrust segments in the 1000‐year historical record for Japan, the same threshold identifies segments in Nankai and Hokkaido as uniquely poised for synchronization. This work offers a mechanistic basis for earthquake synchronization and, more broadly, shows how interaction shapes earthquake recurrence.

Abstract Motivated by an observed asymmetry in boundary layer convergence lines (BLCLs) between the Kubuqi and Ulan Buh desert borders of the Hetao Oasis in northern China, this work aims to document the differences in BLCL frequency and convective initiation and their potential drivers. Using radar, surface Automatic Weather Station, and reanalysis data in June–August from 2012 to 2022, we found that the Kubuqi border had more frequent, earlier‐forming, and more convective BLCLs, whereas the Ulan Buh border had fewer, shorter‐lived, and more tortuous ones. These contrasts were associated with sharper vegetation and thermal gradients, higher Bowen ratio, and a more favorable background wind on the Kubuqi side. Our findings demonstrate that BLCL behavior is highly sensitive to surface, flow, and terrain heterogeneity, even around a single oasis.

Climate extremes driven by climate change, such as heatwaves and severe droughts, have significantly impacted the ecosystem of the European Alps. Climate projections indicate that the frequency and intensity of such extreme events in this region will continue to rise. However, it remains unclear how heatwaves and extreme droughts affect the long-term evolution of local vegetation. Long-term time-series of satellite-based vegetation indices and climate records were analyzed to catch effects of climate extremes on spatiotemporal trends of plant phenology. MODIS data from the Alpine space was processed for a 23-year range (2001 to 2023). Correlation between anomalies in the length of the growing season (LOS) and extreme droughts and heatwaves were analyzed. Results indicated that the impact of these climate extremes on plant phenology varies significantly across different phenological stages and regions. Heatwaves led to an earlier onset of SOS primarily in the western high-altitude areas (-0.23 to -0.41 days yr⁻¹), whereas in the lower eastern mountainous regions, they resulted in a later onset of SOS. The consistency analysis between climate extremes and LOS showed that negative LOS anomalies were primarily caused by extreme droughts, which account for 65.4% of cases. Additionally, in drought-dominated areas, approximately 78% of the LOS shortening can be attributed to earlier EOS. Regions under extreme drought stress were mainly located in the southwestern and eastern low-altitude areas of the Alpine space. Furthermore, compound drought-heatwave events significantly shorten LOS across different areas (coincidence rate = 0.68). The study highlighted how remote sensing can advance understanding of mixed effects of extreme droughts and heatwaves on plant phenology considering also the spatial context.

Abstract High spatial and temporal resolution wind forcing is essential to capture the highly variable atmospheric circulation driving ocean currents in shallow waters and across complex coastal features in the Mississippi Sound. In this study, we examine the sensitivity of modeled hydrodynamics to sea‐land breeze (SLB) circulation. Two versions of the CONCORDE Meteorological Analysis (CMA) wind fields were applied as forcing in the COAWST‐based model: an hourly 1‐km resolution data set and a 24‐hr low‐pass filtered version of this data set (CMA24). Inter‐comparison of the resulting solutions isolates the impact of diurnal SLB circulation. Our analysis demonstrates that removing diurnal, SLB‐dominated wind variability modifies circulation patterns and impacts the heat exchange between the atmosphere and estuarine‐shelf waters within the Mississippi Sound. This study estimates its contribution to summer latent heat flux to be 18 W/m 2 . These results underscore the importance of high‐resolution atmospheric forcing for representing coastal ocean dynamics in modeling applications.

ABSTRACT Evidence on climate adaptation innovations remains fragmented, limiting evidence‐informed policy. This meta‐analysis of 418 studies (2010–2023), stratified by outcome and design with publication‐bias adjustment, synthesizes effectiveness, equity, and scalability. Hybrid approaches showed a 39% advantage over single‐domain innovations ( g = 1.24), although bias correction reduced this by 18%. Outcomes varied across domains. Equity improved when women's participation exceeded 40%, yet 73% of technological interventions showed elite capture. Portfolio approaches are recommended, supported by adaptive, inclusive governance.

Abstract Differential compressional stresses (<100 MPa) that do not result in permanent deformation have traditionally been considered insufficient to modify paleomagnetic signals; however, recent numerical predictions have challenged this assumption. Using quantum diamond microscopy, we experimentally determined the effects of low differential stresses on the remanent magnetization of columnar basalt samples from Seljadalur, Iceland. We show that although individual magnetic dipoles rotate by up to 40°, the bulk magnetization direction remains largely unchanged. These rotations reflect opposing single‐domain–like and single‐vortex–like behaviors that cancel directionally; however, the bulk magnetization intensity decreases. This behavior is consistent with micromagnetic model predictions for mixed–domain‐state samples. Given that differential stresses 1 GPa are common in undeformed rocks distal to earthquake rupture zones and impact craters, it is essential to account for stress effects when interpreting paleomagnetic data from geological materials.

Abstract Cyanobacterial harmful algal blooms (CyanoHABs) pose significant environmental and public health risks through toxin production, antibiotic resistance gene dissemination, and the disruption of biogeochemical cycles. Although extensive metagenomic investigations have explored microbial interactions between Cyanobacteria and coexisting members, the potential of incorporating functional indicators into early CyanoHAB monitoring systems remains underexplored. In this study, we employed metagenomic binning combined with machine learning approaches to analyze seasonal cyanobacterial succession and coexisting microbial communities across 19 sites in Erhai Lake across four seasons. We documented distinct niche shifts driven by the spatiotemporal heterogeneity in seasonal cyanobacterial succession alongside the formation of unique niches triggered by this asynchrony. Specialized microbial members (SN-MAGs) were identified as candidate bioindicators associated with early-stage cyanobacterial proliferation. Functional profiling showed r/K selection and metabolic adaptations in these SN-MAGs, including stress resistance and altered carbon metabolism, linked to cyanobacterial dynamics. Machine learning models leveraging these features were constructed and retrospectively evaluated. Cross-lake validation using an independent Lake Taihu dataset confirmed associations among these features and microbial community shifts preceding Microcystis proliferation. Our findings highlight the potential value of the dynamics of specific bacteria as a sensitive indicator of cyanobacterial dynamics. Moreover, we propose that these identified functional markers could potentially be developed into targeted molecular assays (e.g., qPCR assays), offering possible utility in predicting Microcystis population expansion.

As the basic component of urban space, the accurate extraction of buildings is the key basis for the application of geographic information. In complex urban scenarios, extracting buildings from high-resolution remote sensing images often faces issues such as blurred boundaries, severe background interference, and insufficient multi-scale feature fusion. To solve these problems, an Edge-Fused Multi-Scale Attention Network (EFMSANet) is proposed in this paper. The network achieves this through a synergistic integration of explicit edge guidance, multi-scale feature fusion, and attention-enhanced feature encoding. An edge detection branch based on dilated convolutions and residual mapping extracts boundary probability maps from the input image, which are then adaptively integrated into decoder features through a spatial-channel dual-attention fusion module that selectively enhances boundary regions while suppressing background interference. A Dual-Branch Multi-Scale Feature Fusion Module (DB-MSFF) captures both global context and local detail in parallel to accommodate buildings of varying scales. In building extraction experiments on Satellite Dataset II, the Aerial Imagery Dataset, and the Massachusetts Buildings Dataset, IoU reached 64.1662%, 88.1049%, and 70.2682%, and F1 scores were 78.1722%, 93.6763%, and 82.5382%, respectively. In the boundary extraction experiments, compared with existing methods, the proposed EFMSANet achieves IoU values of 50.9732%, 83.9798%, and 72.8380% on the three datasets, respectively, which outperforms the suboptimal methods by 1.76, 2.51, and 0.61% points and obtains the optimal results. Experimental results show that the proposed method can achieve high-precision and robust building segmentation in complex urban scenarios.

Abstract Biomass burning (BB) degrades indoor air quality both through direct emissions and via smoke sorption onto surfaces, resulting in persistent pollutant re-emission and transformation. However, the dynamics of this multiphase process remain unclear. This study investigates the re-emission, chemical aging, and secondary product formation of smoke residues on three representative indoor surfaces (glass, cotton, and gypsum). Smoke-contaminated surfaces re-emitted 120–180 distinct volatile organic compounds (VOCs). After 2 weeks of ventilation, the median emission rates of various VOCs from porous cotton and gypsum (3.3 and 4.2 μg m–2 h–1) far exceeded that from glass (0.06 μg m–2 h–1), reflecting their greater sorption capacity and longer pollutant persistence. The emission increased by factors of 7 and 4 on glass and cotton when relative humidity (RH) rose from 20% to 80%, attributed to enhanced initial sorption. Smoke-sorbed gypsum showed a strong ozone (O3) uptake coefficient ((1.6 ± 0.2) × 10–5 at 50 ppb O3, 20% RH), approximately 20 times that on glass, confirming its role as a key indoor O3 sink. Heterogeneous transformation of surface residues produced secondary VOCs, including aldehydes and carboxylic acids, further deteriorating indoor air quality. These findings demonstrate the critical and prolonged influence of surface-retained BB emissions on indoor air chemistry.

Abstract Not in Perspective articles.

Coastal zones provide recreational ecosystem services across land and sea, yet spatial data for ecosystem accounting is limited. We compared Strava activity data with survey-based public participation GIS (PPGIS) data across the Oslofjord, Norway, to assess spatial and sociodemographic biases in crowdsourced recreation accounting. Strava over-represented activities along the coastal belt and across marine areas, while under-representing island visitation and leisure-oriented activities. Spatial bias was partly associated with age, gender, and income. Although Strava underestimated absolute recreation activity levels relative to PPGIS, the datasets showed strong agreement in relative spatial variation among municipalities. Our findings demonstrate that Strava data can support mapping of recreation patterns, but may misrepresent the magnitude and distribution of coastal recreation activity. These biases may influence the identification and prioritization of recreational areas in planning and accounting. Combining Strava data and representative survey-based approaches may improve recreation assessments for coastal planning and accounting for physical and monetary recreation services.