DPS 7

Danube DeltaMonitoring system and knowledge of sediment flow spilled in the Black Sea

Organization: NIHWM/UDJG (Romania)

In DPS7 the University of Galati intends to prove that using specific equipment for sediment monitoring is possible to improve the knowledge of sediment budget, especially in extreme events, and to obtain a better knowledge of sediment budget transported through the Sulina branch into the Black Sea.

”Dunarea de Jos” University of Galati will be involved in Dalia project in Pilot 7 Danube Delta, Sulina branch, through research about:

• water, sediments, biodiversity;
• developing of prediction models;
• forecasting models based on water quality and sediments;
• developing of a deep-learning.

Within this pilot we will develop new, innovative methods of modelling as machine learning, deep-learning and neural networks in order to establish an analytical framework, suitable to assist environmental management strategies within the Danube River. Thus, in order to achieve this goal, the following activities are considered:

-           The developing of prediction models related to both water quality matrix, suspended sediments concentrations and the presence and accumulation of different type of sediments;

-           The developing of forecasting models based on water quality and sediments time series data by using ARIMA;

-           The developing of a deep-learning based framework, capable of identifying the water quality status, as well as sediments presence and accumulation degree, within the Danube River study sector.

These results will be able to be replicated in all the project Pilots.

The use of new, innovative methods such as machine learning, deep learning and neural networks in order to establish an analytical framework, suitable to assist environmental management strategies within the Danube River is developed based on 3 pillars, as follows:

• The 1^st pillar consists of the use of machine learning algorithms for developing virtual sensors for monitoring the quality of water and sediments, based on prediction models.

• The 2^nd pillar targets to develop forecasting models, based on water quality and sediments time series data, such as Danube lighthouses for decision management.

• The 3^rd pillar establishes a deep-learning-based framework for the identification of water quality and sediment status.

More informations

...