Monitoring sea ice is important for understanding climate change and supporting safe travel in the Arctic, but creating detailed maps usually requires time-consuming manual work. In this study, we trained a computer model to learn from large amounts of unlabelled satellite images before using only a small number of labelled examples. This approach produced accurate ice maps while greatly reducing the need for manual labelling, making large-scale monitoring more practical.

We investigate the potential of drone-based thermal cameras in the Swiss Alps and found they can misread surface temperatures when the camera warms or cools. Lab and field checks showed some correction methods exaggerate extremes. Calibrating on site with ground sensors gave the best results as well as the implementation of a site-specific validation. This improves temperature maps needed to track melting snow and ice, and unstable ground in rapidly changing mountains.

Earth Observation archives now encompass multispectral imagery, yet producing analysis-ready Sentinel-2 time series remains a critical bottleneck. We present an open-source Python package that builds data cubes from Spatio-Temporal Asset Catalog catalogues and integrates cloud masking, co-registration, and super-resolution in one workflow. Validated on high-performance computing, it enables rapid, reproducible cube construction and updates, improving temporal consistency in a braided river.

Land and sea surface temperatures, collectively called skin temperature, are available from satellite measurements since few decades now and provide a holistic view of the Earth's climate system. Here we review 5 different datasets of Tskin to show that climate change signature is clear from the different datasets with notable exception, such as a significant cooling in India, parts of Africa and the Pacific.

The shrinking of glacier frees large areas worldwide in alpine environment which are progressively colonized by the vegetation. Monitoring this process often relies on in situ botanical observations which are challenging to obtain. Here, we quantify accurately the uncertainty of the vegetation cover measured by optical images from Unmanned Aerial Vehicles. We find that, despite biases, this increasingly used technology extend and complete the data collected during such field campaign.

Water hyacinth (WHY) are one of the most costly aquatic invasive species. Regular monitoring of WHY would the management of infested waterboddies. We present an algorithm to derive daily maps of WHY from satellites, featuring an interpolation scheme, which led to 55% more daily WHY maps of Hartbeespoortdam Reservoir. The algorithm is light-weight and can be run on a cloud-computing platform. This allows applications in (sub-)tropical regions with limited access to computing power and/or funding.

This study summarizes challenges of operating consumer-grade UAVs over high-latitude Arctic pack ice. High ice-floe drift and the lack of commercial grid-mission solutions north of 85° N hinder the generation of void-free orthomosaics and DEMs. We present an autonomous high-latitude grid-flight approach with pre- and post-flight drift correction, applied during ArcWatch-2 (Aug/Sep 2024), enabling ice thickness, surface type and pond depth retrievals.

New satellite observations are used to improve our understanding of agricultural land on a global scale. Innovative artificial intelligence methods are employed to combine the available information. This improvement is evaluated using independent observations and varies depending on the characteristics of instruments and models.

We mapped the riverbed through a 28 km canyon in southern France using airborne laser measurements, where field surveys are difficult and risky. By reprocessing the full recorded signal and combining it with targeted sonar checks, we recovered deeper and more complete underwater terrain than standard approaches. The resulting riverbed elevation model preserves key channel shapes needed for improved flood simulations, habitat assessment, and river management.

High-resolution satellite imagery was used to map the expansion in number and area of glacial lakes across Kyrgyzstan from 2016 to 2024. Small, high-elevation, glacier-connected lakes grew most, whereas larger, lower-elevation lakes remained stable. Including these often omitted small dynamic lakes in the 2024 inventory advances comprehensive monitoring of glacier–lake systems and high-mountain hydrology, highlighting how glacier retreat and topography shape lake distribution.

San Francisco International Airport is slowly sinking. Using satellites that measure ground movement with millimeter precision, we mapped subsidence across the airport and found that areas built on thick, soft bay sediments sink fastest. This sinking adds directly to rising sea levels, threatening runways with flooding and cracking decades sooner than climate projections alone suggest. Satellite monitoring can help airport managers identify vulnerable areas before costly damage occurs.

Passive microwave (PMW) satellite data have been widely used to monitor Antarctic sea ice velocity (SIV) on a pan-Antarctic scale. We present a comprehensive evaluation of a PMW-derived Antarctic SIV by leveraging high-resolution SIV fields derived from synthetic aperture radar imagery. We found a widespread ~1 km d^-1 underestimation of ice drift speed across the Southern Ocean, caused by optimal interpolation.

Repeated airborne laser surveys were used to study how forest structure can be tracked over time in Switzerland. By comparing different years and methods, canopy height estimates were found to be consistent, while vegetation density results varied depending on how the data were collected. This highlights the need for consistent methods and data collection for reliable long-term analysis.

This study examines the rapid expansion of unregulated gold mining in the Ecuadorian Amazon rainforest. Using different types of satellite images, the spread of mining over time is analyzed and reproducible workflows are presented. Results show significant growth, especially in recent years, often starting along rivers and expanding into surrounding forests, highlighting the value of satellite monitoring for environmental protection and informed decision-making.

Soil moisture (SM) plays a key role in weather, agriculture, and water management. While satellites can measure SM from space, obtaining accurate, frequent measurements throughout the day remains challenging. Here, we explore how deep learning models can improve sub-daily SM estimates. Our approach focuses on capturing spatial patterns and adapting to local conditions. Using data from the ASCAT satellite instrument, we show that this model can produce reliable SM estimates multiple times a day. 

Mangrove forests are vital for climate protection and coastal communities, but their extent is often uncertain. This study uses a satellite-based artificial intelligence approach (AlphaEarth Foundations) to map mangroves across Southeast Asia and Papua New Guinea from 2017 to 2024. Achieving over 80 % accuracy, the results provide consistent maps and improved area estimates with uncertainty, supporting better monitoring and climate reporting.