Day 4

Detailed paper information

Back to list

Paper title MAPAQUALI - Customizable modular platform for continuous remote sensing monitoring of aquatic systems
  1. Claudio Barbosa Brazilian National Institute for Space Research - INPE Speaker
  2. Evlyn Novo Brazilian National Institute for Space Research - INPE -Brazil
  3. Rogerio Flores Junior Brazilian National Institute for Space Research - INPE -Brazil
  4. Felipe Carlos Brazilian National Institute for Space Research - INPE -Brazil
  5. Daniel Maciel Brazilian National Institute for Space Research - INPE -Brazil
  6. Vitor Gomes Institute for Advanced Studies (IEAv)
  7. Gilberto Ribeiro Queiroz Brazilian National Institute for Space Research - INPE -Brazil
  8. Karine Reis Ferreira Brazilian National Institute for Space Research - INPE -Brazil
  9. Vitor Martins Michigan State University
  10. Felipe Nincao Begliomini Brazilian National Institute for Space Research - INPE -Brazil
  11. Thainara Lima Brazilian National Institute for Space Research - INPE -Brazil
  12. Rejane Paulino Brazilian National Institute for Space Research - INPE -Brazil
  13. Felipe Lobo Federal University of Pelotas (UFPel)
  14. Marie Paule Bonnet DR2 Institut de Recherche pour le Développemen
  15. Raianny Leite do Nascimento Wanderley Brazilian National Institute for Space Research - INPE -Brazil
Form of presentation Poster
  • A7. Hydrology and Water Cycle
    • A7.06 EO for monitoring water quality and ecological status in inland waters
Abstract text MAPAQUALI - Customizable modular platform for continuous remote sensing monitoring of aquatic systems

The water resource is vital, not only for maintaining life on Earth, but also for supporting economic development and social well-being as the sustainable growth of all the nations depends upon water availability. Approximately 12% of the planet's surface fresh water available for use circulates through the Brazilian territory. Due to this water availability, Brazil has an extensive number of large artificial and natural aquatic ecosystems. This water availability places Brazil in a privileged position, but it also poses a great challenge for sustainable use and monitoring of these natural resources. For instance, the nutrient inflows to lakes and hydroelectric reservoirs from irrigated agriculture and sewage from nearby cities significantly contribute to the eutrophication process and the systematic occurrence of cyanobacterial blooms. These blooms can be harmful and produce toxins that lead to a series of public health problems. Even when not harmful, they impair fisheries and the recreational use of those water bodies. These environmental impacts on aquatic ecosystems need to be determined and monitored, mainly in reservoirs, as energy sources, besides being renewable, must be clean. This study summarizes the integrated effort of specialists in hydrological optics, aquatic remote sensing, and computer science to build a customizable modular platform, named MAPAQUALI. The platform allows a continuous monitoring of aquatic ecosystems based on satellite remote sensing, and the integration of bio-optical models derived from in-situ measurements. The platform will generate and make available, for aquatic ecosystems for which it is customized, a spatiotemporal information about water quality parameters: Chlorophyll-a, Cyanobacteria, Total Suspended Solids, Secchi disk depth, diffuse attenuation coefficient (Kd), and bloom events alerts (especially cyanobacteria).

The MAPAQUALI platform comprises the following modules: Data Pre-processing; Bio-optical Algorithms; Query and View WEB.

The Data Pre-processing Module (DPM) generates and catalogs Analysis Ready Data (ARD) [10.1109/IGARSS.2019.8899846] collections, which are input data for the Bio-optical Algorithms Module (BAM) for water quality products generation. The DPM has data acquisition, processing and cataloging functionalities. The DPM structure is flexible for adding new processing tasks or even new functionalities. The following processing tasks are available in the current implementation of MAPAQUALI: query and image acquisition from data providers (Google Cloud Platform or Brazil Data Cube Platform [10.3390/rs12244033]); atmospheric correction procedure with 6SV model; water bodies’ identification and extraction; cloud and shadow masking; sunglint and adjacency corrections. The BAM comprises parameterized/calibrated/validated algorithms using Brazilian inland water in situ bio-optical datasets (LabISA – INPE), and OLI, MSI, and OLCI simulated spectral bands. Algorithms were parameterized for OLCI sensor only for aquatic systems having the suitable size, such as large lakes in the Amazon floodplain. In addition, to ensure the best possible accuracy, we developed a semi-analytical [10.1016/j.isprsjprs.2020.10.009; 10.3390/rs12172828], hybrid [10.3390/rs12010040], machine learning [10.1016/j.isprsjprs.2021.10.009], and empirical [10.3390/rs13152874] algorithms, using in situ data representative of the full range variability of the apparent and inherent optical proprieties. These algorithms are achieving accurate results. For example, the hybrid algorithms for Chl-a have an error of 20% (MAPE = 20%). Machine learning algorithms, for estimating water transparency, presented errors of approximately 25%. Moreover, Kd algorithm for oligotrophic reservoir resulted in errors of 20%. The Query and View WEB is a web portal providing resources for searching for the aquatic systems integrated into the platform and returns the products available for each of them.

Tools for viewing and analyzing time series are available in this module. Registered users to the platform can freely download products and images from the ARD archive. Additionally, users can consume any available data through our web geoservices enabled database, such as Web Map Service (WMS) or Web Feature Service (WFS).

For the integrated application of DPM and BAM processing tasks, we are using the process orchestration infrastructure available on the Brazil Data Cube Platform [10.3390/rs12244033]. In this way, MAPAQUALI platform can perform all operations periodically, which allows continuous monitoring of the aquatic systems under consideration. At the end of each execution, the newly generated data products are cataloged, and made available for consulting in monitoring activities.

In our ongoing efforts, we are customizing water quality bio-optical algorithms to four aquatic ecosystems: two multi-user reservoirs, a set of lower amazon floodplain lakes, and one nearshore coastal water. As modular and customizable, others aquatic ecosystems can be easily inserted into the platform.