Task 3 - Forested landscape management planning

Duration: 24 months

Start date: March 19th 2020

Person * months: 41

Researchers involved: Adrian Pascual Arranz; Brigite Botequim; José Borges; Liliana Ferreira; Marco Marto; Marlene Marques; Marta Mesquita; Susete Marques; Vladimir Bushenkov and 2 Masters to be hired

Consulting: Marc McDill and Sandor Tóth

The objective of this task is to develop an approach to integrate wildfire risk and spread into landscape-level management planning and to optimize the spatial-temporal of management options, namely fuel treatments thus minimizing expect loss from wildfires. It reads the data and information from Tasks 1 and 2, it requires computing resources and it is structured into three sub-tasks.
 

3.1 Spatial optimization of fuel treatments
 

In this sub-task we will develop an approach to optimize the fuel treatments spatial distribution in any given planning period. The sub-task will consider the decision space generated in Task 1 and the corresponding potential landscape configurations under each climate scenario. It will consider further the forest management planning wildfires simulator developed in Task 2 and the corresponding ignition probability, flammability and wildfire spread probability associated to each stand. The spatial optimization model will target the minimization of the expected loss from wildfires and will build from enumeration of every possible fuel treatment spatial distribution over the forest stands. The solution will build from the development and parameterization of heuristic techniques – e.g., simulated annealing, tabu search, genetic algorithms – to address this specific spatial optimization problem.

3.2 Spatial and temporal optimization of management options

 

In this sub-task we will extend the approach developed in 3.1 to address dynamic problems, encompassing several planning periods and multiple management options (e.g. harvests, fuel treatments). It will consider the same inputs from Task 1 (decision space) and Task 2 (wildfire simulator). The spatial and temporal optimization model will target the maximization of the return from the harvests, plus the expected value of the landscape at the end of the planning horizon. The solution will consist of the optimal configuration of treatments, e.g. the one that best balances the opportunity cost of harvesting at a different time with the gains in reducing the expected loss from fire in each climate scenario. The solution will build like in 3.1 from the development and parameterization of heuristic techniques – e.g., simulated annealing, tabu search, genetic algorithms – to address this specific spatial-temporal optimization problem.
 

3.3 Combining spatial optimization of fuel treatments with multiple criteria approaches
 

In this sub-task, we aim at integrating the spatial optimization of fuel treatments with multicriteria methods. We will extend the approaches developed in 3.1 and 3.2 to address the provision of a wider range of ecosystem services. It will consider the same inputs from Task 1 (decision space) and Task 2 (wildfire simulator). The multiple criteria method will integrate the optimization models developed in 3.1 and 3.2 and will generate its Pareto frontier to provide information about trade-offs between wildfire protection goals and the provision of other ecosystem services. The solution will build from the algorithms developed in 3.1 and 3.2 as well as from decomposition approaches to Pareto frontier methods that may address the complexity of the combinatorial optimization problem.
 

© 2019 - PCIF/MOS/0217/2017