The overarching goal is to improve the understanding of the population dynamics of these polychaetes to guide sustainable resource management in the context of fishing and environmental changes. To reach this goal, the successful candidate will develop a method based on image analysis to estimate population dynamics without direct contact with the sand. Current methods (digging or pumping) are labor-intensive, cover small areas, and cannot achieve frequent sampling. The project proposes to obtain images from ground or aerial vehicules (drones) and to automate image analysis using classical and/or machine learning methods to detect and analyze lugworm faecal casts. Ongoing collaborations with two other labs, UMR LamCube and UMR CRIStAL, showed the feasibility of this approach based on preliminary results.
The position is fully funded through a research contract between UMR LOG and the French Observatory for Biodiversity (OFB).
Context : Arenicolid lugworms are polychaetes mainly represented in Europe by two species
(Arenicola marina and A. defodiens). They are ecologically and economically important,
constituting a significant part of the marine biomass on European sandy beaches, and play a
crucial role in the benthic ecosystem, serving as bioindicators. These psammivorous organisms
extract organic matter from the sand and produce faecal casts at the surface of the sediment.
Found ubiquitously across European tidal beaches (from Sweden to Portugal) and even in North
Africa, their populations are threatened by intensive fishing, primarily for bait. While other
countries have introduced regulations to manage marine worm fishing pressure, France has
lagged behind. A recent prefectural decree restricted the use of pumps and imposed limits on
the number of worms collected to combat illegal bait sales. This has sparked controversy among
anglers, highlighting the need for better understanding of these species’ population dynamics.
Additionally, hemoglobin extracted from their blood is used as a human blood substitute for
organ preservation before transplants, and an increase of the population exploitation is expected
to this end.
Current Research and Problem Statement : Several studies have been conducted by the UMR
8187 LOG (laboratory of Oceaonology and Geosciences), including PhD work and regional
initiatives, focusing on the life cycle, metabolism, and population dynamics of these lugworms
in North of France. These studies confirmed the presence of two species and assessed the risks
of unregulated fishing, which could harm population sustainability (De Cubber et al., 2018). A
dynamic energy budget (DEB) model was developed to understand the life cycle of A. marina,
revealing a complex life history involving larval dispersal and migration (De Cubber et al.,
2019, 2020, Brocquart et al., 2022). The most recent study (De Cubber et al., 2023) takes a
mechanistic approach, combining DEB models with individual-based models (IBM) to examine
the population dynamics of A. marina and A. defodiens along the northeastern Atlantic coast.
The results highlight intra- and interspecific competition under favorable environmental
conditions, while unfavorable conditions lead to different impacts on the populations, with
sharp declines for A. defodiens and atypical processes for A. marina. This approach offers
promising insights for predicting the evolution of lugworm populations under environmental or
human-induced pressures. This IBM tool has the potential to be utilized by resource managersin the future, but it still requires full validation due to a lack of extensive data on the structure
and dynamics of lugworm populations.
Objectives and Methods : The overarching goal is to improve the understanding of the
population dynamics of these polychaetes to guide sustainable resource management in the
context of fishing and environmental changes. To reach this goal, the successful candidate will
develop a method based on image analysis to estimate population dynamics without direct
contact with the sand. Current methods (digging or pumping) are labor-intensive, cover small
areas, and cannot achieve frequent sampling. The project proposes to obtain images from
ground or aerial vehicules (drones) and to automate image analysis using classical and/or
machine learning methods to detect and analyze lugworm faecal casts. Ongoing collaborations
with two other labs, UMR LamCube and UMR CRIStAL, showed the feasibility of this
approach based on preliminary results.
Key Responsibilities:
Conduct research, experiments and analyze data.
This work is structured in two main phases:
1. Image Acquisition and Analysis: The first phase involves capturing images of lugworm
faecal casts, which can be done either by a single operator or using a UAV (unmanned aerial
vehicle). Once images are collected, they will be processed to:
o Detect and count faecal casts to determine their abundance.
o Measure each cast’s size, specifically its diameter, which relates directly to the size of the
worm, providing data to characterize the population structure.
Field sampling will follow a spatio-temporal design adapted to account for the approach's
limitations and needed precision. Machine learning techniques will be employed for automatic
image analysis, specifically focusing on accurate segmentation of faecal casts to facilitate
robust shape and size measurements. Additionally, further improvements are necessary to refine
the correlation between cast diameter and worm size.
2. Estimating Population Dynamics: In this phase, we will apply geostatistical methods and
mixture models (e.g., Bhattacharya’s method) to interpret the spatio-temporal dynamics of the
lugworm population. These analyses will allow us to correlate population structure and
dynamics with environmental variables, including beach morphology, sediment composition,
patchiness, organic matter content, and temperature.
This two-step approach provides a comprehensive framework for understanding lugworm
population dynamics and supports future improvements in predictive modeling for resource
management."
Publish results. At least two papers in high quality journals or international conferences should
be published in the course of the position
Collaborate with other researchers within the lab and from two other labs. UMR LOG is a
pluridiciplinary lab in which collaborations with several groups in ecology and sedimentology
will be strongly encouraged.Ongoing collaborations with computer scientists and civil
engineers from two other labs, UMR LamCube and UMR CRIStAL, should be strengthened.Possibly supervise master students and contribute to grant writing.
Required Qualifications
• Educational background: Ph.D in ecology and population dynamics with a strong interest
in marine applications and numerical approaches
• Specific skills: coding skills (R, python, …), data and image analysis (ImageJ, …),
geostatistics, machine learning, field research
• Publications in the research field.
Duration and Start Date: 18 months, Start expected in first trimester 2025
Funding and Salary
• The position is fully funded through a research contract between UMR LOG and the
French Observatory for Biodiversity (OFB)
• Salary range between 2000 and 2500 € per month including health insurance and 5 to
9 week-vacations
Location
The position is based at UMR LOG (Laboratory of Oceanology and Geosciences), at the marine
station of Wimereux (university of Lille), North of France, by the Eastern English Channel.
UMR LOG is a pluridisciplinary laboratory promoting interdisciplinarity research (see
https://log.cnrs.fr/ for details).
Mentorship and Collaboration Opportunities
The position will be supervised primarily by two researchers, Prof. Sébastien Lefebvre (marine
ecology and numerical ecology), and Dr. Sylvie Gaudron (Marine invertebrates life-history
traits and DEB modeling). Collaborations with other researchers from the LOG teams Interest,
Geosed and Geolit are highly encouraged.
The post-doctoral fellow will also be supervised by researchers from two other laboratories
located in Lille, UMR LamCUBE (e.g., Nicolas Bur) and the Color Imaging team of UMR
CRIStAL (e.g., Olivier Losson). One-day travels to Lille and video meetings, especially at early
stages of the research work, will be planned to ensure supervision by these two teams on image
acquisition and analysis aspects.
Application Process
Detail the required documents for the application:
o CV and cover letter
o Research statement.
(c) GdR IASIS - CNRS - 2024.