Assessing the influence of deer browsing on forest ecosystem services with new stand-level optimization and agent-based models

鹿的觅食行为可能会严重影响森林的发展，造成严重的环境后果。虽然现有研究集中分析了鹿食食对林分组成和生长的生物化学物理影响，但缺乏对鹿食食对生态系统服务提供影响的综合评价。在此，我们提出三个相互关联的研究领域来填补这一空白。1)在不确定性条件下，提出创新的林分级多准则优化方法，以获得最优满足利益相关者偏好的理想林分作为基准。提出的改进这些模型的创新包括对遥远未来发生的ES的处理以及对ES供应的鲁棒性的评估。第一个研究领域将考虑木材供应、碳固存、地下水补给、地下水质量（例如氮污染）、经济回报和生物多样性作为决策标准（以枯木为指标）。2)以最优组成的基准林分为基础，提出了模拟实验，以阐明鹿食对林分的影响。我们建议在森林景观模型中实施基准森林，并将其暴露给鹿。为了模拟鹿的冲击，我们将概念化一个基于主体的模型。在这样的模型中，幼树和小鹿是自主代理，其中多代理研究和模拟框架将支持在这些代理之间建立相互作用。当前项目“BioWild”的观察和现有文献将提供信息，以建立幼树对浏览的反应和鹿对森林特征的反应的规则。3)假设鹿的密集觅食会对林分的最优构成和管理形成约束，我们建议推导鹿对树木影响的影子价格。这些将包含基准森林优化中包含的所有ES的成本。我们期望这一研究概念能够全面评估大量鹿在树上觅食对环境的影响。

Deer browsing may strongly affect the development of forests with severe environmental consequences. While the existing research has intensely analyzed the bio-chemo-physical impact of deer browsing on the composition and growth of forest stands, a comprehensive assessment of the influence of deer browsing on the provisioning of ecosystem services (ES) is still lacking. Here we propose three interrelated research fields to fill this gap. 1) We suggest innovative stand-level multiple criteria optimization under uncertainty to derive ideal forest stands to be used as benchmarks that will optimally meet stakeholders’ preferences. The proposed innovations for enhancing such models include the treatment of ES occurring in the far future and the assessment of the robustness of ES provisioning. This first research field will consider timber provisioning, carbon sequestration, groundwater recharge, groundwater quality (for the example of nitrogen contamination), economic return and biodiversity as decision criteria (with dead wood as an indicator). 2) Based on the optimally composed benchmark forest stands we suggest simulation experiments to elucidate the impact of deer browsing. We propose implementing the benchmark forests in a forest landscape model and exposing them to deer browsing. For simulating the impact of deer, we will conceptualize an agent-based model. In such a model, young trees and deer are autonomous agents, where a framework for multi-agent research and simulation will support establishing interactions among such agents. Observations of the current project “BioWild” and existing literature will provide information to set up rules for the young trees’ response to browsing and the deer’s response to forest characteristics. 3) Assuming that intense deer browsing will form constraints for the optimal composition and management of a forest stand we suggest deriving shadow prices for the deer’s influence on the trees. These will encapsulate the costs for all ES included in the optimization of our benchmark forests. We expect that this research concept will allow a comprehensive assessment of the environmental impact of heavy deer browsing on trees.