GeneticLand: Discovering future landscapes under climate change scenarios using genetic algorithms

Project is funded by FCT/MCES, the Portuguese Science Foundation. Grant no. POCTI/MGS/37970/2001. Project started in Sep/2002 and finished in Jan/2005.

Research team

We are collaborating in this project with researchers at Universidade Nova de Lisboa and Universidade de Évora. The project leader is Prof. Júlia Seixas at FCT/UNL. The members of the research team at the University of Algarve are,

Project summary

The simulation of future behaviours of physical systems under climate change scenarios raises the possibility of improving the state of knowledge of the physical systems' underlying processes. Some examples include the relationships between hydrological processes at the land surface and processes within the atmosphere. Although some progress has been made, the assessment of related impacts (e.g. soil erosion) have been constrained by three issues: (1) the limited understanding of processes relating climate forcing and hydrological responses, (2) the lack of methods to couple different multi-scale (space and time) processes and patterns, and (3) the lack of techniques to generate future system states according to forecasted variables.

For simplicity and feasibility, the major driving force governing the carrying capacity of semi-arid landscapes is assumed to be hydrological soil erosion. Within this framework, the project has two novel research goals. The first one is the development of a method to integrate both large scale and small scale physical processes governing soil erosion. The second one is to design a genetic algorithm capable of generating landscapes according to global objectives (e.g., minimization of soil loss, maximization of net primary productivity), subject to local constraints (e.g., spatial coherence) and guiding requirements (e.g. policy land uses).



Last updated by Fernando Lobo, 14/Mar/2006.