Lead Inventor:
Maria Uriarte, Ph.D.
Foresty Management Involves Timber Yields, Biodiversity, Harvesting:
Forestry management has the challenging task of integrating ecological concerns with economic objectives. Sustainable harvesting recommendations must take into account multiple criteria such as timber yields, structural complexity, biodiversity, and other desirable forest attributes. Current harvesting models are ineffective at aligning multiple ecological variables such as structural biodiversity with economic demands. Furthermore, the spatial pattern of a harvest dramatically affects the shift of competitive balance amongst the remaining unharvested trees. Most current algorithms for managing forests are not able to capture complex spatio-temporal interactions. Thus, incorporating such fine-scale spatial interactions into harvesting regimes is crucial in optimizing desired forest characteristics.
Managed Forests Optimize Harvests with Forest Simulator Algorithm:
This invention designs harvesting plans that optimize multiple forest attributes through machine learning methods coupled with a forest simulator. The outcomes of hundreds of potential harvesting schedules are projected over many years. Then a specialized algorithm which models non-linear interactions is used to identify those schedules with favorable outcomes. Thus, this algorithm can identify the best harvesting regimes while coordinating multiple ecological constraints and maximizing timber yield.
Applications:
• Software to determine harvesting regimes that maximize timber value while addressing ecological concerns
• Software to optimize carbon sequestration and storage
Advantages:
• Can integrate the non-linear spatial-temporal interactions between trees and their subsequent effects on forest development
• Can explore the effects of a very large range of inherently conflicting variables (e.g. ecological concerns with economic goals)
• Algorithm can be trained with a small number of examples
• Produces a time savings relative to using a forest simulator of four orders of magnitude
• Considers natural forest development dynamics as the benchmark against which harvest regimes are evaluated
• Can be adapted to a wide range of goals and can be applied to any forest simulator
Opportunities:
• Sponsored research funding of continued application development
• Licensing
Patent Status: Patent Pending (WO 2009/015069)