EBLab Numerical was founded as a means to provide direct access to expertise in numerical methods suitable for marine and aquatic systems. The creation of EBLab Numerical was inspired by discussions within The Entangled Bank Laboratory about structural blockages to promoting acceptance of quantitative methods in ecological research and training programs.
EBLab Numerical develops expertise in applied data and image analysis, hydrodynamics and ecosystem modelling, individual-based modelling in ecology, and in, optimisation and data assimilation. We mainly develop tailored numerical solutions for Environmental Impact Assessment projects in coastal and offshore areas.
Through partnerships, we promote the development of quantitative system-based EIA principles and practices with the long-term objective of improving the predictive capacities of ecological and environmental studies.
To discuss your project, please contact:
guarini.jm.micrent@gmail.com
Numerical Resources
Coastal hydrodynamic models. Since we deal regularly with the complex dynamics of coastal environments, we have acquired a 3D hydrodynamic modelling system. We often implement nested models to improve the representation of boundary conditions for our Ecosystem-Based Environmental Impact Assessment projects. Currently we have the capability to represent the entire English Channel with nested high-resolution simulations of embayments, on both the coasts of France and the UK. We also have a functioning model of Mediterranean Sea hydrodynamics with nested high resolution simulations at locations in the Eastern Basin.
Custom codes. All custom methods are integrated within our SPatially Oriented-Tools in Ecosystem Modelling framework (SPOT'EM). Any custom codes developed are supplied in an open format to the project client, along with the complete model framework description.
Some of our projects
1. Image analysis and processing for benthic facies mapping.
The analysis of aquatic and marine systems depends on adequate knowledge of the sediment facies at a project or study site. In the past, this required extensive sampling and sediment grain size analysis in dedicated facilities. Advances in the quality of underwater imagery now make it possible to use numerical methods instead to map quickly grain size distributions for large sections of benthic environments (10s of kilometers). EBLab Numerical developed a new approach based on existing USGS methods using a wavelet analysis to simplify the process. The method was applied successfully to map the distribution of riverbed surface sediments with underwater photography. We also developed a user interface to further automate the analysis.
2. Assessing marine water quality at local scales with multi-spectral sensors.
Ocean color indices estimated from satellite imagery are used to evaluate global primary productivity in surface waters. Similar approaches could be used to monitor productivity in coastal zones at smaller scales by mounting custom-designed light sensors on UAVs. Light reflectance measured by multispectral sensors could then be exploited to build new proxies to provide near real-time snapshots of primary productivity at scales relevant to local authorities. However, the standard remote sensing methods available to treat these data should be re-examined, because at smaller, local scales, mathematical models can provide better estimates than statistical correlations (even if this requires a site-specific calibration of model parameters). EBLab Numerical contributed an original approach that models the fundamental relationships between the reflectance measurements at different wavelengths. The goal is to have a mechanistic description suitable for the small spatial scales (on the order of 10s of centimeters) that dominate experimental and monitoring programs. With this approach we can produce more accurate estimators of chlorophyll a and turbidity from sensor imagery, than purely statistical methods.[1]
[1] McEliece, R.; Hinz, S.; Guarini, J.-M.; Coston-Guarini, J. Evaluation of Nearshore and Offshore Water Quality Assessment Using UAV Multispectral Imagery. Remote Sens. 2020, 12, 2258. https://doi.org/10.3390/rs12142258
3. Evaluating environmental quality with sustainable, eco-forecasting methods: Monitoring ecosystem health with bivalve "valvometry".
EBLab Numerical is leading an initiative to use valvometry for environmental impact assessment (EIA) projects. EBLab Numerical brings expertise in modelling complex dynamic systems to improve sensor calibrations that when combined with other datastreams (e.g. acoustics, hazardous algal blooms, water quality parameters ...) can generate alerts about changes in the environment. The method is being adapted to monitor shellfish aquaculture systems with the aim to provide a longer forecast window that should allow managers enough time to take any necessary preventive measures. [2,3]
[2] Guarini J.-M., Coston-Guarini, J., Comeau, L.A. 2020 [preprint] Calibrating Hall-Effect valvometers accounting for electromagnetic properties of the sensor and dynamic geometry of the bivalves shell. bioRxiv 2020.12.20.423648; doi: https://doi.org/10.1101/2020.12.20.423648.
[3] Guarini J.-M., Coston-Guarini, J., Comeau, L.A. 2020 [preprint] Interactions between discrete events and continuous dynamics in the regulation of scallops valve opening: insights from a biophysical model. bioRxiv 2020.12.25.424408; doi: https://doi.org/10.1101/2020.12.25.424408.
[4] Guarini, J.-M.; Hinz, S.; Coston-Guarini, J. (2021). Designing the Next Generation of Condition Tracking and Early Warning Systems for Shellfish Aquaculture. J. Mar. Sci. Eng. 2021, 9(10), 1084; doi: 10.3390/jmse9101084.
4. Calculating the propagation error in sediment spill estimates for dredging applications.
EBLab Numerical collaborated recently with LimOce Environmental Consulting on the problem of estimating the uncertainty of the total amount estimate of sediment spilled over board during dredging operation. Dredgers have specific requirements in terms of estimate accuracy, but this accuracy depends on both the sampling data dispersion and the way that errors propagate in the calculation of the total amount from different flux estimates. In this very interesting study, we have developed and programmed in Python a series of functions dedicated to estimate the propagation of uncertainty from basic measurement dispersion using bootstrap-based algorithms. It contributed to identifying problems in the calculation sequence, hence improving the monitoring program, while providing the required estimator values.
It was also an occasion to contribute orienting Environmental Impact Assessment toward quantitative methods that consolidate the discipline and outcomes.