%0 Research article
%T Integrating wind disturbances into forest planning: a stochastic programming approach
%A Eyvindson, Kyle
%A Kangas, Annika
%A Nahorna, Olha
%A Hunault-Fontbonne, Juliette
%A Potterf, Maria
%D 2024
%J Silva Fennica
%V 58
%N 4
%R doi:10.14214/sf.23044
%U https://silvafennica.fi/article/23044
%X Forest disturbances challenge our ability to carefully plan for sustainable use of forest resources. As forest disturbances are stochastic, we cannot plan for the disturbance at any specific time or location. However, we can prepare for the possibility of a disturbance by integrating its potential intensity range and frequency when developing forest management plans. This study uses stochastic programming to integrate wind intensity (wind speed) and wind event frequency (number of occurrences) into the forest planning process on a small coastal Finnish forest landscape. We used a mechanistic model to quantify the critical wind speed for tree felling, with a Monte Carlo approach to include wind damage and salvage logging into forest management alternatives. We apply a stochastic programming model to explore two objectives: maximizing the expected forest net present value or maximizing the even-flow of income. To assess the effects of improper wind risk assumptions in planning, we compare the results when optimizing for correct versus incorrect wind intensity and frequency assumptions. When maximizing for net present value, the impacts of misidentifying wind intensity and frequency are minor, likely due to harvests planned immediately as trees reach maturity. For the case when maximizing even-flow of income, incorrectly identifying wind intensity and frequency severely impacts the ability to meet the required harvest targets and reduces the expected net present value. The specific utility of risk mitigation therefore depends on the planning problem. Overall, we show that incorporating wind disturbances into forest planning can inform forest owners about how they can manage wind risk based on their specific risk preferences.