We created the first comprehensive, high-resolution, and easily accessible bathymetry dataset for the three main branches of the Danube Delta. By combining four data sources, we obtained a detailed representation of the riverbed, with resolutions ranging from 2 to 100 m. This dataset will support future studies on water and nutrient exchanges between the Danube and the Black Sea and provide insights into the delta's buffer role within the understudied Danube–Black Sea continuum.

LakeBeD-US is a dataset of lake water quality data collected by multiple long-term monitoring programs around the United States. This dataset is designed to foster collaboration between lake scientists and computer scientists to improve predictions of water quality. By offering a way for computer models to be tested against real-world lake data, LakeBeD-US offers opportunities for both sciences to grow and to give new insights into the causes of water quality changes.

Vertical profile observations are key to understanding regional air pollution but remain scarce due to existing limits. This study presents a high-time-resolution (ca. 15 min) dataset of aerosol, nitrogen dioxide, and formaldehyde vertical profiles from 32 sites in China (2019–2023) using passive remote sensing. It documents vertical distribution, seasonal variations, and diurnal patterns, revealing long-term trends. Data are available at Zenodo under https://doi.org/10.5281/zenodo.15211604.

The high-resolution autonomous underwater vehicle (AUV) dataset for the South China Sea (SCS) provides 13 491 temperature and salinity profiles and covers 463 d of experiments. To our knowledge, the resolution and length of this dataset are enough to detect the asymmetry, vertical tilt, and temporal evolution of mesoscale eddies (MEs) and the corresponding submesoscale processes. The dataset is expected to improve the accuracy of current and biogeochemistry numerical models. More projects conducting AUV experiments will be promoted in the future.

Freshwater from Arctic land ice loss strongly affects the Arctic and North Atlantic oceans. Datasets describing this freshwater discharge have low resolution and do not cover the entire Arctic. We statistically enhanced coarse-resolution climate model data – from approximately 6 km to 250 m – and routed meltwater towards the coastlines to provide high-resolution data covering all Arctic regions. This approach has far fewer computational requirements than running climate models at high resolution.

Forest age is critical for carbon cycle modeling and effective forest management. Existing datasets, however, have low spatial resolutions or limited temporal coverage. This study introduces China's annual forest age dataset (CAFA), spanning 1986–2022 at a 30 m resolution. By tracking forest disturbances, we annually update ages. Validation shows small errors for disturbed forests and larger errors for undisturbed forests. CAFA can enhance carbon cycle modeling and forest management in China.

Based on a deep-learning method, we established a global classification dataset of daytime and nighttime marine low-cloud mesoscale morphology. This aims to promote a comprehensive understanding of cloud dynamics and cloud–climate feedback. Closed mesoscale cellular convection (MCC) clouds occur more frequently at night, while suppressed cumulus clouds exhibit remarkable decreases. Solid stratus and MCC cloud types show clear seasonal variations.

Grounding line discharge is a measure of the amount of ice entering the ocean from an ice mass. This paper describes a dataset of grounding line discharge for the Antarctic Ice Sheet and each of its glaciers. The dataset shows that Antarctic Ice Sheet grounding line discharge has increased since 1996.

Air pollution adversely affects health, ecosystems, and infrastructure. In this work, we have developed a methodology and created a dataset for describing air pollution sources for the heating of individual houses for the Western Balkan region. The dataset has high spatial resolution and can help explain the air pollution situation in the region.

Understanding how leaves absorb carbon from the atmosphere is essential for predicting changes in global forests. Young leaves play a key role in this process, but their efficiency has been difficult to measure at large scales. Using satellite data, we developed a new method to track the seasonal patterns of young leaves’ photosynthetic capacity from 2001 to 2018. Our dataset helps scientists better understand forest growth and how ecosystems respond to climate change.

We develop P-CAME, a long-term gridded emission inventory for China spanning from 1978 to 2021. P-CAME enhances the accuracy of emissions mapping, identifies potential pollution hotspots, and aligns with observed Hg⁰ concentration trends. With its improved spatial resolution and reliable long-term trends, P-CAME offers valuable support for global emissions modeling, legacy impact studies, and evaluations of the Minamata Convention.

In this study, we estimate emissions of nitrogen oxides (NO_x) in 2022 at high resolution at the global scale, using satellite observations. We provide maps of the emissions and identify several types of sources. Our results are similar to the EDGAR emission inventory. However, differences are found in countries with lower observation densities and lower emissions.

IAR-HP (Italian Alpine Region HydroPower) is a detailed inventory of large hydropower systems in Italy's Alpine Region, aimed at improving their inclusion in hydrological modeling by providing relevant information with a consistent level of detail. It includes structural, geographic, and operational data for over 300 hydropower plants and their related reservoirs and water intakes. Validated through modeling, IAR-HP accurately reproduces observed hydropower, capturing 96.2 % of actual production.

Accurately quantifying the distribution of soil profile  SOC (soil organic carbon) stocks is crucial for carbon sequestration and mitigation. The detailed spatial subsoil SOC data are the scientific basis for environmental protection, as well as for the development of Earth system models. Based on multiple environmental variables and soil profile data, this study used machine learning approaches to evaluate  the SOC stocks and their spatial distribution at a depth interval of 20–100 cm in various ecosystems.

We developed a 3 h, 1 km surface soil moisture dataset for the contiguous United States from 2015 to 2023 using the spatio-temporal fusion method. This dataset effectively combines the distinct advantages of two long-term surface soil moisture datasets, which is also the first hourly-level 1 km soil moisture dataset at the continental US scale. The new dataset could provide new insight into the fast changes in soil moisture along with drought and wet spell occurrences.

Many waterbodies undergo nutrient decline, called oligotrophication, globally, but a comprehensive dataset to understand ecosystem responses is lacking. The OLIGOTREND database comprises multi-decadal chlorophyll a and nutrient time series from rivers, lakes, and estuaries with 4.3 million observations from 1894 unique measurement locations. The database provides empirical evidence for oligotrophication responses with a spatial and temporal coverage that exceeds previous efforts.

After Russia's 2022 invasion of Ukraine, Europe's energy dynamics shifted significantly. Our study introduces updated datasets tracking changes in natural gas supply, usage, and transmission within the EU27&UK. We discovered that Europe adapted to losing Russian gas by increasing LNG (liquefied natural gas) imports and shifting to renewables. Our insights could shape future energy policies and climate research.

Seismic paleochannel interpretation is vital for georesource exploration and paleoclimate research yet remains time-consuming. While deep learning offers automation potential, it is limited by the lack of labeled data. We present a workflow to simulate geologically reasonable 3D seismic volumes with diverse paleochannels, generating a large-scale labeled dataset. Field applications demonstrate its effectiveness. The dataset and codes are publicly available to support future research.

We present a global geospatial database of cropland and pastures for the year 2015. We built these data by fusing satellite-based land cover data with agricultural census data using machine learning. This database is an update to an earlier version representing the year 2000. It can be used to study issues such as land use, food security, climate change, and biodiversity loss. We provide a reproducible code base that can be used to easily update the product for future years.

Particulate Observing Scanning Polarization (POSP) on board the second GaoFen-5 (GF-5(02)) satellite is the first space-borne ultraviolet–visible–near-infrared–shortwave-infrared (UV–VIS–NIR–SWIR) multi-spectral cross-track scanning polarimeter. Due to its wide spectral range and polarimetric capabilities, POSP measurements provide rich information for aerosol and surface characterization. We present the detailed aerosol/surface products generated from POSP's first 18 months of operation, including spectral aerosol optical depth, aerosol-size-/absorption-related properties, surface black-sky and white-sky albedos, etc.

Denitrification and anaerobic ammonium oxidation (anammox) are two important nitrogen removal pathways that convert reactive nitrogen into dinitrogen gas. Here, we construct a global database on actual nitrogen loss rates, covering over 30 years of observations, measured in coastal and marine sediments. This work provides a global overview of the biogeography and potential controlling factors of denitrification and anammox and highlights the potential applications of this database.

Optimal atmospheric dynamic conditions are essential for convective storms. This study generates a dataset of high-resolution divergence and vorticity profiles using the measurements of a radar wind profiler mesonet in Beijing. The negative divergence and positive vorticity are present ahead of rainfall events. This suggests that this dataset can help improve our understanding of the pre-storm environment and has the potential to be applied in weather forecasting. 

The Histoire physique de la mer, written by Luigi Ferdinando Marsili in 1725, was one of the first treatises to analyse the science of the sea. However, it is difficult to understand Marsili's original data. This paper reports the results of a major effort that has been undertaken to re-evaluate Marsili's observations, converting historical measurements into modern units – water weight to water density – with bathymetric profiles mapping the locations where these measurements were made and sea level variations alongside consideration of the associated error.

Building on the foundation of Palynodata, the Global Acritarch Database (GAD) added 29 new fields, 4531 new entries, 2 238 366 new metadata points, and 415 new references, resulting in a database comprising 115 860 entries, 43 fields, 3 050 852 metadata points, and 7791 references. GAD represents records from 1146 different sampling sites spanning geological history from the Precambrian to the Phanerozoic, and the fossil records include 1456 genera and 9865 species (excluding sp.).

The air–sea CO₂ flux in coastal waters plays a key role in the global carbon budget but remains poorly understood. In 2021, the Tara schooner collected 14 000 km of CO₂ fugacity (fCO₂) data along the South American coast. This dataset improves our understanding of fCO₂ in the under-sampled Brazilian coastal region and provides a unique insight into the complex biogeochemistry of the Amazon River–ocean continuum.

To address the limitations of short time spans in satellite data and spatiotemporal discontinuity in site records, we reconstructed global monthly burned area maps at a 0.5° resolution for 1901–2020 using machine learning models. The global burned area is predicted at 3.46 × 10⁶–4.58 × 10⁶ km² per year, showing a decline from 1901 to 1978, an increase from 1978 to 2008 and a sharper decrease from 2008 to 2020. This dataset provides a benchmark for studies on fire ecology and the carbon cycle.

In situ observations of snow water equivalent (SWE) are critical for climate applications and resource management. NorSWE is a dataset of in situ SWE observations covering North America, Norway, Finland, Switzerland, Russia, and Nepal over the period 1979–2021. It includes more than 11.5 million observations from more than 10 000 different locations compiled from nine different sources. Snow depth and derived bulk snow density are included when available.

We present a new set of multi-temporal lidar metrics of ecosystem structure derived from four national ALS (airborne laser scanning) surveys of the Netherlands (AHN1–AHN4), capturing vegetation height, cover, and structural variability over the last 2 decades (1998–2022). Around 70 TB point clouds have been processed to ready-to-use raster layers at 10 m resolution (~ 59 GB), enabling a wide use and uptake of ecosystem structure information in biodiversity and habitat monitoring and ecosystem and carbon dynamic modelling.

In this study, we use machine learning models to produce the first global maps of Pb concentrations and isotope compositions in the ocean. In line with observations, we find that (i) the surface Indian Ocean has the highest levels of pollution, (ii) pollution from previous decades is sinking in the North Atlantic and Pacific oceans, and (iii) waters carrying Pb pollution are spreading from the Southern Ocean throughout the Southern Hemisphere at intermediate depths.