Land cover monitoring is crucial to understand land transformations at a global, regional and local level, and the development of innovative methodologies is necessary in order to define appropriate policies and land management practices. Deep learning techniques have recently been demonstrated as a useful method for land cover mapping through the classification of remote sensing imagery. This research aims to test and compare the predictive models created using the convolutional neural networks (CNNs) VGG16, DenseNet121 and ResNet50 on multitemporal and single-date Sentinel-2 satellite data. The most promising model was the VGG16 both with single-date and multi-temporal images, which reach an overall accuracy of 71% and which was used to produce an automatically generated EAGLE-compliant land cover map of Rome for 2019. The methodology is part of the land mapping activities of ISPRA and exploits its main products as input and support data. In this sense, it is a first attempt to develop a high-update-frequency land cover classification tool for dynamic areas to be integrated in the framework of the ISPRA monitoring activities for the Italian territory.

In recent years, deep learning (DL) algorithms have been widely integrated for remote sensing image classification, but fewer studies have applied it for land consumption (LC). LC is the main factor in land transformation dynamics and it is the first cause of natural habitat loss; therefore, monitoring this phenomenon is extremely important for establishing effective policies and sustainable planning. This paper aims to test a DL algorithm on high-resolution aerial images to verify its applicability to land consumption monitoring. For this purpose, we applied a convolutional neural networks (CNNs) architecture called ResNet50 on a reference dataset of six high-spatial-resolution aerial images for the automatic production of thematic maps with the aim of improving accuracy and reducing costs and time compared with traditional techniques. The comparison with the National Land Consumption Map (LCM) of ISPRA suggests that although deep learning techniques are not widely exploited to map consumed land and to monitor land consumption, it might be a valuable support for monitoring and reporting data on highly dynamic peri-urban areas, especially in view of the rapid evolution of these techniques.

The Land Use, Land-Use Change, and Forestry (LULUCF) and agriculture sectors are increasingly central to global climate policy. They play a crucial role in climate mitigation strategies, as land acts as a carbon sink that needs to be enhanced and as a source of greenhouse gas (GHG) emissions that must be reduced. In the European context, the LULUCF Regulation (EU 2018/841), revised in 2023, aims for 310 Mt CO2eq net removals by 2030 and requires spatially explicit land-use representations to monitor land dynamics and assess policy impacts.Within the Horizon project AVENGERS (Attributing and Verifying European and National Greenhouse Gas and Aerosol Emissions and Reconciliation with Statistical Bottom-up Estimates), a methodology was developed to generate an IPCC-compliant land-use map by integrating multiple Copernicus Land Monitoring Service (CLMS) products. In national GHG inventories, the operational use of spatial explicit data is often limited due to restricted temporal coverage, inconsistencies with national statistics, and challenges in interpreting mixed classes and land-use/land cover definitions. This methodology provides a transparent approach to reconcile inventory data with high-resolution spatial datasets.The approach combines the CLC Plus Backbone geometry with CORINE Land Cover (CLC) and ancillary CLMS datasets, including the High-Resolution Layer Crop Types and Priority Areas monitoring products (e.g., Coastal Zones, Riparian Zones, and Protected areas). Multiple layers were integrated using overlay techniques and priority rules, resulting in an harmonized map at 10-m spatial resolution. CLC attributes were aggregated to IPCC land use categories, allowing direct comparison between mapped areas and inventory surfaces.Preliminary validation involved cross-checks with national land-use activity data to ensure reliability of mapped areas across LULUCF categories. The resulting maps enable the spatialization of inventory-based LULUCF and agriculture emissions, producing gridded emission datasets based on improved spatially explicit land-use information. These datasets are suitable for use as input (priors) in atmospheric inversion modelling, a top-down emissions estimation method supporting policy evaluation.The methodology is designed to be replicable across all European countries covered by CLMS data and to be updated approximately every 2–3 years, in line with the regular update cycle of CLMS products. The methodological framework is modular and flexible, based on a spatial data storage and management scheme developed by ISPRA, which allows the integration of additional datasets and adaptation to different territorial contexts. The approach was applied and tested in three national case studies for the year 2018—Italy, Sweden, and the Netherlands—with specific adaptations introduced to account for distinct territorial characteristics. This first implementation represents a promising step and provides a solid foundation for further refinements and future developments, supporting the production of high-resolution land-use maps helpful for national inventory agencies and inversion modelling experts.