|Paper title||Operational modeling and forecasting of bio-physical properties in alpine lakes|
|Form of presentation||Poster|
Water quality remote sensing is increasingly used in an operational context, and several studies in particular for perialpine lakes showed how hydrodynamic modeling can greatly improve the utility of remotely sensed products. Conversely, remotely sensed products can help to improve the performance of hydrodynamic models as a source of dynamic input data, by means of data assimilation, or for validation. With such an interdisciplinary integration of Earth observation techniques, we can take advantage of the forecasting capabilities of data-driven hydrodynamic lake modeling and the synoptic coverage, as well as a regular sampling of high-resolution satellite imagery, i.e., from Sentinel-2.
A first, operational framework that partially established the integrated usage of Earth observation data for Lake Geneva resulted from the ESA project CORESIM (www.meteolakes.ch). As part of the ESA Regional Initiative for the Alpine Region, the project AlpLakes aims to extend this framework functionally and spatially. The two main objectives of AlpLakes are to integrate Sentinel-2 transparency products in hydrodynamic models in order to improve their performance, and to update the models with a particle tracking module for validation with Total Suspended Matter (TSM) estimates from Sentinel-2 data. Ultimately, the project aims at understanding the short- and long-term evolution of the dynamics of freshwater systems with a particular focus on altitudinal and latitudinal gradients. For this purpose, we selected eleven lakes north and south of the Alps as test sites, covering a wide range of morphological and hydrological features, trophic status, and climatic conditions.
We use Sentinel-2 products to derive information of the light penetration and turbidity at a high temporal and spatial resolution. Our workflow is based on remote sensing image processing, field data acquisition, model setup and calibration via data assimilation, and real-time operational model publication on an open-access web-based platform. Sentinel-2 Secchi depth products obtained with state-of-the-art algorithms (e.g., QAA) will be validated with monitoring data. Dedicated field campaigns will be conducted to improve performance by means of generalized inherent optical properties for lakes in the Alpine region. Such products are crucial to constrain and improve the hydro-thermodynamic models, as transparency information is used in the heat fluxes models to parameterize the distribution of the incoming solar radiation in the water column and hence to correctly reproduce the lake thermal structure.
Similarly, existing algorithms for TSM retrieval will be tested and optimized for the use case of Sentinel-2 and the Alpine region. The resulting TSM maps are used to validate the simulated flow field and understand the transport dynamics in the lakes. To this aim, we use a Lagrangian particle tracking module coupled with the three-dimensional hydrodynamic model. Spatial patterns identified in Sentinel-2 images will serve as a proxy for the particle tracking seeding area and particles concentration. This allows tracking the evolution of detected spatial structures in Sentinel-2 image driven by turbulence and mixing processes in the lake. The accuracy of this method will be assessed by comparing the predicted evolution of particles paths with the succeeding Sentinel-2 TSM products.
For dissemination of and user interaction with the combined Sentinel-2 products and hydrodynamic simulations, we will provide hindcasting, real-time, and forecasting functionalities in a web-based platform on the basis of Datalakes (https://www.datalakes-eawag.ch/). This will allow open access to all results, provide a common tool for scientists, decision makers and the broader public, and improve the management of lakes in the Alpine lakes as well as the public perception of environmental processes in their immediate living space.