Optimization of Riparian Buffer Zones with Respect to Nature and Economics : Multi-objective optimization in forestry

Sammanfattning

Riparian buffer zones are forested areas left unharvested next to streams andwatercourses in order to protect water quality, biodiversity, shading, and otherecological functions. In Swedish forestry, these zones are often planned usingfixed-width rules, which are simple to apply but do not account for the spatialvariation in hydrology, tree height, species composition, or timber value. Thiscan result in areas being protected where the ecological benefit is relativelylow, while more sensitive locations may receive insufficient protection. At thesame time, every retained tree represents an economic opportunity cost for theforest owner.This thesis investigates how riparian buffer zones can be optimized so thatecological values are preserved while economic losses from reduced timberextraction are minimized. The problem is suitable for an applied mathematicsand industrial engineering thesis because it combines spatial data processing,ecological modeling, economic valuation, and mathematical optimization in areal forestry planning context. The difficulty lies in representing ecologicalvalue using available proxy data and in producing buffer-zone layouts thatare not only numerically efficient, but also spatially coherent and practicallyinterpretable.The method was developed for watercourses in Asa Research Forestusing raster and vector data provided through Skogforsk in collaborationwith Svenska Lantbruksuniversitetet (SLU). Ecological value was modeledfrom depth-to-water index, tree height, tree species, shade contribution, anddistance to water. Economic value was approximated from tree height andspecies-dependent timber prices. A binary graph-cut optimization modelwas then implemented in Python, where each raster pixel was classified aseither protected or harvested. The model included spatial regularizationand monotonicity constraints to avoid fragmented solutions and to ensurecontinuous buffer zones extending outward from the watercourse. Severaltrade-off scenarios were evaluated by varying the ecological weight parameter.The results show a clear trade-off between preserved ecological valueand harvested revenue. Compared with fixed-width buffer alternatives, theoptimized variable-width solutions can preserve the same ecological valuewhile retaining higher estimated timber revenue in several cases. The thesistherefore suggests that spatially adaptive buffer planning can provide a moretransparent and efficient decision-support basis for forestry operations nearwatercourses.