Składanie ofert zakończone. Udzielenie zamówienia lub sprostowanie tego ogłoszenia: e-mail, gdy rejestr to zobaczy.Zaloguj się
Germany - Research and development services and related consultancy services - DE_391 Destination Earth Machine-Learning Demonstrators
Zamawiający: European Centre for Medium Range Weather Forecasts (ECMWF)
Kraj: Niemcy
Szacunkowa wartość: 1 000 000 €
Tryb: Przetarg nieograniczony
Branża (CPV): Badania i rozwój (73000000)
Aperlena: https://aperlena.com/pl/t/germany-research-and-development-services-and-related-consultancy-services-de-391-destination-earth-machine-learning-52648b8303
Ogłoszenie urzędowe: https://ted.europa.eu/en/notice/-/detail/497744-2026
Udzielony
Opublikowano
Opis
ECMWF, as one of the Entrusted Entities for the Destination Earth (DestinE) initiative, entered into contracts for the provision of services aimed at implementing DestinE ML-based demonstrators. Contracts are expected to implement ML Demonstrators on different topics, addressing needs in different domains. To this end, contracts were planned to be placed in different Lots: Lot 1: Supporting water resilience in Europe; Lot 2: Demonstrating DestinE in the domain of food security; and Lot 3: Combining heterogeneous climate projections for climate adaptation. As result of this ITT, ECMWF awarded two contracts under lot 1 and two contracts under lot 3 to the two highest ranking tenderers of each due to exceptional value, and no contracts under Lot 3 due to unsuitable tenders.
The objective of the first contract awarded under Lot 1: DestinE ML Demonstrator for Water Resilience combines Large Language Models (LLMs) and Graph Neural Networks (GNNs) with physics-based hydrological and water quality models. This hybrid system could enable users to interact with complex simulations using natural language, configure “what-if” scenarios, and receive tailored, interpretable outputs-instantly and intuitively. The "what-if" scenarios-such as changes in land use, nutrient loads, pollutant emissions, or infrastructure-will be translated into adjustments of model schematizations, model parameters, nutrient loads, and pollutant emissions. These adjustments will be used to simulate the effects on water availability, water quality (e.g., nitrogen, phosphorus, PFAS), water temperature, and/or sediment dynamics. The simulations will utilize multi-model climate change scenarios available from DestinE, enabling adaptive water resilience planning. The DT climate storyline simulations of present and future climate will be used to develop storylines to show the need and options for better water resilience. The objective of the second contract awarded under Lot 1: The contract covers the development and demonstration of an ML application that provides probabilistic forecasts of freshwater temperature extremes in the Scheldt river (Belgium) under climate change, with a focus on supporting climate-resilient energy systems and industrial cooling operations. It exploits Climate DT simulations and implements a “storylines-framework” to present plausible future heatwave scenarios and their impact on freshwater temperatures.; Tenders under this Lot were asked to suggest how to target the implementation of a Demonstrator that supports a clearly defined stakeholder group in their tasks linked to agricultural market information systems, food security mechanisms, or resilience planning. However, Lot 2 produced no satisfactory results.; The first contract under Lot 3 focuses on combining information from the high-resolution climate simulations in the DestinE Climate DT with global CMIP6 and regional EURO-CORDEX climate projections, using artificial intelligence (AI). The Climate DT provides globally consistent, high-resolution data that explicitly resolves key climate processes for three models, reducing model uncertainty and bias, while CMIP and EURO-CORDEX ensembles suffer from biases and coarse resolution but can contribute long, multi-model time series that quantify structural uncertainty for climate services. Using the FRAIMHiRes AI downscaling method, trained on Climate DT simulations, the contract will try to transfer the potential added value of high-resolution DT data to (potentially longer) coarser-resolution DT ensembles by creating a “ClimateDT FineScaler”. The resulting downscaled ensemble will be evaluated for its statistical robustness, added value over original projections and CMIP/CORDEX simulations, and suitability for climate impact and adaptation studies. The second contract under Lot 3 develops and demonstrates a machine-learning application for high-resolution regional dynamical downscaling of global and regional climate scenarios (using the Anemoi framework), integrated within -and trained on data from- the Destination Earth Climate DT. By combining Climate DT simulations with CMIP6, EURO-CORDEX and national climate scenarios, the contract aims to increase the size of European climate ensembles, improve uncertainty quantification, and deliver high-resolution climate information for adaptation planning across Europe.
Oficjalne źródło
Źródło: TED - Tenders Electronic Daily (Publications Office of the EU)
Zawsze weryfikuj szczegóły w oficjalnym ogłoszeniu.
Ostatnia weryfikacja w TED - Tenders Electronic Daily (Publications Office of the EU) 1 miesiąc temu
udzielono: Norwegian Meteorological Institute - MET Norway7 927 375 €
Historia zmian
Każdy przypadek ponownej publikacji tego ogłoszenia przez oficjalne źródło ze zmieniona treścią. Monitorujemy wszystkie źródła, które indeksujemy, w tym regionalne kopie tego samego ogłoszenia.
Przeczytaj przed złożeniem oferty
Wartość szacunkowa:
Ustalono na 1 000 000 €
Data udzielenia:
Udzielono 19 stycznia 2026
Przedmiot i kryteria:
Opis został rozszerzony. Przeczytaj ponownie przed złożeniem oferty.
Wartość szacunkowa:
Ustalono na 1 000 000 €
Obserwuj ten przetarg, a napiszemy do Ciebie, gdy tylko się zmieni.
Załóż konto