Simulation of Human Activities Dynamics (DAHU)
applied to the marine environment

Matthieu Le Tixerant, Mathias Rouan, François Cuq

Laboratoire Géomer (UMR 6554 CNRS), Université de Bretagne Occidentale
Institut Universitaire Européen de la Mer, Plouzane (FR)

Coastal seas are the seat of multiple, sometimes rival, human activities interacting with the environment. A better understanding of these interactions are essential to sustainable development.
In this context, a focus on the ways in which human society uses and exploits the environment is indispensable. Such a focus can reply to important questions posed by scientists who need to evaluate the role of human activity on the ecosystems evolution, and by administrators researching information offering a global and interdisciplinary territory vision. Appropriate compilation and communication tools are thus needed to process various information on spatial and temporal dimensions of the dynamics of human activities.
Specific problems arise from the marine environment. The legibility of phenomena related to human activities, in such open space, is very complex. This paper focuses on the specific uses of the marine environment, and particularly of the coastal seas.

The methodology suggested: a simulation toolkit of Human Activities Dynamics (DAHU) applied to the marine environment.
There are few structured information sources that both describe human uses in a detailed way and give a global view of their spatial and temporal dynamics. This paper presents a methodology and technology implemented for the DAHU simulation toolkit. In order to produce relevant information, the system must be able to process a large amount of technical data of various nature and origin, related to the progress of human activities.·

Modelling principles
The specificity of the methodology consists in considering space as the major constraint of simulation: for each modeled activity, a specific space reference frame acts as principal calibration data.
An interdisciplinary approach is essential. Human marine activities are subjected to many variables due to the physical and biological environment, to weather conditions and to legal and socio-economical constraints (without which it is impossible to understand the system).
In order to support a determinist approach, professional activities are selected since their archetypal characteristics facilitate their modelling.
The adopted methodology, although applicable elsewhere, is based on the case study of the Iroise Sea on the French coast.

Space-time modelling
A typology of the different activities is first carried out. Then, to understand their progress, a space and time disintegration of initial data, is carried out. For each activity, a simulated environment is built as a potential practice territory and calendar: spaces and periods of activities progress, are identified by the superposition of various "thematic spatio-temporal filters" that integrate environmental, climatic, legal and socio-economic constraints. Spatialization is achieved by projecting activities constraints in a Geographical Information Data Based exploited by a GIS (Geographical Information System). In the same way, on a one year cycle, the superposition of various temporal filters allows to build a calendar of potential practice activity, by identifying periods when practice is possible.
This step is essential because, in most cases, existing data are not connected with spaces and periods concerned with activities. This is particularly true in marine environment where there is no given space of reference (administrative limits) from which the activities description is possible.

DAHU simulator functioning
The simulator follows a traditional operation diagram on three stages : preparation of input calibration data, modelling and analysis of results.
Stage 1: Preparation of input calibration data (pre-processor)
The input data, first formatted in the space-time modelling phase, are pre-processed in a format exploitable by the model, in order to connect space, temporal and statistical data in a Space Reference Data Base (SRDB).
Stage 2 : Modelling (processor)
This simulator is built on an architecture directly inspired by a Multi-Agents System (MAS). To each activity corresponds an autonomous agent able to interact with the beforehand modelled environment. The results of these interactions are spatial and temporal constraints, which force activities progress. The system is specified in UML and implemented in an oriented object language (C++).
Stage 3: Results analysis (post-processor)
Exploitation of results via a GIS allows restitution of relevant cartographic representations. Theses representations provide a useful synthetic form of information for integrated management. The fact that activities models are coupled to a GIS, then it allows to integrate input data and to analyse output data.

Current system functions
Currently, the system is able to provide a daily description of simultaneous human activities progress on coastal sea. With various space-time scales, it is possible to quantify and show where and when the activities are supposed to be present. The potential conflicts are detected by highlighting conflicting zones and periods of practice.
Applications of such system on commercial fishing activities of the Iroise Sea are presented in the paper.

Conclusion / prospects
One of the main objectives of this work is to relate quality environmental information with the information concerning human activities in order to evaluate the potential impact of anthropical uses on the environment. This simulation system may be applied to address problems of daily management and to evaluate the consequences of management decisions on the coastal environmental system.

References