Finding efficient management policies for forest plantations through simulation Models and Simulation Project 2013 http://disi.unal.edu.co/~gjhernandezp/sim/
Goal: Find a “good” management policy for a forest plantation Initial decisions Interventions Tree model Plantation – stand model Optimization by evolution simulation - genetic algorithm 2
Tree Model Photosynthesis http://www.phschool.com/science/biology_place/biocoach/photosynth/overview.html
Photosynthesis Biomass Water Temperature Sunlight Minerals Photosynthesis Biomass Crown- leaves and branches (LA Leaf Area) Roots - weight Trunk- (D diameter at breast height in cm and H Height )
Spatial model of the tree crown growth
Plantation model-celular automaton
Competition Index Competition in forestry, refers to the action caused by a tree over another and the second prevents normal development and growth. For individual trees, competition is usually estimated by a competition index. http://www.worldagroforestrycentre.org/sea/?q=node/124
Plantation parameters Area Time interval Tree species Local variation of soil fertility Tree genetic strength
Forest Management decisions Initial decisions: distance and planting pattern
Forest Management actions Pruning Thinning When and which ones?
Experimental setting Area: 100m x 100m, distance between 1 0m and 9m with 1 m grid size Time interval: one year Tree species: shade intolerant
Neighborhood radio R: 6m Local variation of soil fertility: 0.9 + (U[0,1]*0.1) Tree genetic strength: 0.9 + (U[0,1]*0.1) Daily insolation: N (1,0.2) Daily rain: N (0.8,0.4) Daily temperature : N (25,10) Policy: (pattern, distance, density cell size, density threshold, thinning criteria) Thinning criteria: random, min D, max D, closest to average D.
Results Netlogo simulation of a plantation evolution under a policy Find a “good” policy with a Genetic Algorithms