In many applications, Earth observation (EO) land use and land-cover products must represent conditions at a specific reference date in the past. This requirement arises in contexts such as retrospective land-use assessments, baseline mapping, and regulatory frameworks that rely on historical cut-off dates. A prominent example is the EU Regulation on deforestation-free products, which aims to ensure that key commodity supply chains are deforestation-free after 31 December 2020. As EO data and automated processing are increasingly used for large-scale mapping to support policy and operational decision-making, the need for rigorous, independent validation has become more critical. However, retrospective validation is challenging because field observations collected in the present may not reflect conditions at the reference date.

This work presents a reproducible validation framework developed under the Copernicus4GEOGLAM country-level support service. The approach combines probability-based sampling with two complementary reference data streams: structured field surveys and interpretation of very high resolution (VHR) satellite imagery. Field data serve both as high-quality ground truth and as a basis for developing interpretation keys that guide consistent visual interpretation of archived VHR imagery for the reference year. This is particularly important in areas where field data collection is not feasible due to cost, accessibility, or security constraints.

Retrospective interpretation is supported by clear decision rules, the use of multi-date imagery where necessary, and structured quality-control procedures to ensure consistency across interpreters. The framework has been successfully applied to validate a 2020 national land-use map in Côte d’Ivoire, with additional applications underway for a cocoa probability map in Cameroon and a coffee map in Honduras.

By clearly separating map production from evaluation, this independent validation framework enhances credibility, increases user confidence, and supports more informed use of EO-derived products across diverse operational and policy contexts.

JackDaw is a metadata-driven GeoLLM and GeoRAG environment designed to transform heterogeneous Earth observation, climate, agriculture and biodiversity data into actionable Earth Intelligence. Its objective is to move from passive access to geospatial datasets towards decision-ready knowledge products that are traceable, reproducible and usable by farmers, land managers, public authorities, researchers and policy stakeholders. The concept is aligned with the GEO vision of Earth Intelligence as the transition from Earth observation data provision to actionable insight generation.

The core principle of JackDaw is that the language model does not act as a source of truth. Instead, it works as a controlled reasoning and orchestration layer connected to authoritative data catalogues, validated analytical services and governed computational tools. User questions are translated into structured workflows defining area of interest, time period, thresholds, thematic constraints and uncertainty limits. JackDaw then uses machine-actionable metadata to discover relevant datasets and services through standards such as OGC API – Records, STAC, DCAT/GeoDCAT and ISO 191xx-based metadata profiles.

Technically, JackDaw combines GeoRAG for spatial vector information, raster and multidimensional data-cube access for Earth observation and climate time series, and textual RAG for scientific, policy and advisory evidence. This enables integrated workflows for crop stress detection, management zoning, drought and heat-risk interpretation, climate adaptation planning, habitat pressure assessment, land-use change monitoring and biodiversity restoration support.

Typical outputs include maps, ranked risk areas, alerts, recommendations, uncertainty flags, explanatory briefs, provenance records and reusable workflow descriptions. JackDaw can be deployed on local infrastructure, national or regional nodes, HPC environments or edge/cloud platforms, supporting data sovereignty and transparent governance. The showcase presents JackDaw as a reproducible operational pattern for trustworthy, standards-based and user-oriented Earth Intelligence services.

The Data Integration and Analysis System (DIAS) is one of the Post‑2025 GEO Work Programme, categorized under “Open Data, Open Knowledge, and Infrastructure,” which provides foundational support to other GWP activities. DIAS serves as a platform that develops a data lake enabling the creation of interactive, decision‑support applications for cross‑cutting issues by integrating domain data with climate change and weather prediction datasets—thereby expected to contribute to the development of Earth Intelligence.

In this session, Dr. Masaki Yasukawa will introduce a project that utilizes DIAS: the Drought Warning System in Ghana, which was demonstrated in 2025.

In this project, the team co‑designed and co‑created a drought early warning system together with local disaster‑management organizations and researchers. The system leverages several datasets available through DIAS, including satellite observations, rainfall forecasts, and topographic data.

Furthermore, the project places strong emphasis on capacity development, including deepening understanding of drought mechanisms, learning the structure of the model and its input/output datasets, and gaining practical skills to effectively use the system.

This showcase aims to demonstrate how GEOGLAM CropWatch follows the co-design principles and operationalizes Earth Intelligence for agriculture and food security in developing nations. Our primary objectives include:

(1) to present CropWatch Cloud's latest co-designed updates, including user-driven crop mapping, near-real-time drought/flood monitoring, and crop yield prediction with national partners. We will share evidence from Nigeria and Thailand, where local institutions have directly shaped platform features to fit local farming systems.

(2) to articulate a replicable co-design framework that moves beyond one-way technology transfer. We will show how joint problem definition, iterative feedback loops, and capacity building have enabled Nigerian analysts to integrate EO-based crop monitoring into routine agricultural monitoring workflows.

(3) to forge new South–South partnerships. We seek to engage additional developing countries, particularly those in Africa, Southeast Asia and Latin America in co-developing next-generation Earth intelligence services. By sharing open methodologies and cloud-based tools, we aim to scale impact across diverse agro-ecological zones.

The main objective of the project is to develop a territorial intelligence tool based on Earth Observation and Geographic Information Systems to model plastic waste transport from river basins into the Ecuadorian Ocean through the interinstitutional RuPO initiative. This tool is intended to strengthen environmental management by generating technical evidence that supports decision-making and public policy design aimed at reducing marine pollution across coastal ecosystems. To achieve this, the project integrates official geospatial information, scientific datasets, and contributions from public institutions, academia, non-governmental organizations, and local communities, allowing the identification of the main sources of plastic pollution throughout the territory. In addition, the initiative analyzes the spatial dynamics of waste movement through river systems, coastal zones, and beaches, considering the influence of natural processes and human activities. The results will help identify priority areas for early intervention and prevention before waste reaches the marine environment. The project also promotes cooperation among national and international stakeholders in order to establish a coordinated response to this growing environmental challenge. Finally, the initiative seeks to create an innovative and replicable methodological model that can be implemented in other Latin American countries, contributing to stronger regional strategies for marine ecosystem conservation and for reducing the impact of plastic pollution in the world’s oceans.