Avhandlingar

Disputation i Skogsindustriella produktionssystem: Magnus Persson

Avhandlingens titel:

Utvärdering av gallringsmetoder och bedömningsmetoder för förbättrat skogsbruk i barrdominerade produktionsskogar i södra Sverige

Forskarutbildningsämne:

Skogsindustriella produktionssystem

Fakultet:

Fakulteten för teknik

Datum:

Fredag 4 november 2022 kl 10:00

Plats för disputation:

Hus N, N1017, Linnéuniversitetet, Växjö

Opponent:

Professor Lauri Mehtätalo, Natural Resources Institute Finland (LUKE), Finland

Betygsnämnd:

Professor Lars Eklundh, Lunds universitet
Professor Annika Kangas, Natural Resources Institute Finland (LUKE), Finland
Professor Jari Hynynen, Natural Resources Institute Finland (LUKE), Finland

Suppleant:
Professor Johan Fransson, Institutionen för skog och träteknik, Linnéuniversitetet

Ordförande:

Professor Johanna Witzell, Institutionen för skog och träteknik, Linnéuniversitetet

Handledare:

Professor Johan Bergh, Institutionen för skog och träteknik, Linnéuniversitetet

Biträdande handledare:

Docent Emma Holmström, Sveriges lantbruksuniversitet, Docent Johan Sonesson, Skogforsk, Dr. Magnus Petersson, Södra skogsägarna

Examinator:

Professor Stergios Adamopoulos, Institutionen för skog och träteknik, Linnéuniversitetet

Spikning:

Fredag 14 oktober 2022 kl 14:00 på Universitetsbiblioteket, Växjö

Abstract

Much of our knowledge about coniferous production stands and associated silvicultural guidelines are based on field experiments which naturally have been established in rather small, homogenous management units. The assumingly large gradient in within-stand variation in practical forestry complicates an efficient use of the guidelines. The aim of the thesis is to evaluate within-stand variation at first thinning: its extent, its various effects on silviculture and its importance for future stand development. Also, to explore the utility of the optical satellite system Sentinel-2 for monitoring thinning and classifying tree species.

The survey showed an unprecedented within-stand variation before thinning for various stand attributes, but the observed thinning operations did not reduce the within-stand variation in any of the attributes. The stands were also thinned heavily (Paper I). The thinning detection model (Paper II) based on Sentinel-2 satellite imagery, separated unthinned, lightly thinned and heavily thinned sample plots with a moderate overall accuracy of 62% (kappa of 0.34). Also, annual volume growth after thinning was estimated with a root mean squared error (RMSE) of 1.8 m3 ha-1 y-1 (relative RMSE: 24%). The thinning regime were optimized (Paper III) based on the within-stand variation before thinning using the Heureka system, but no benefits in terms of stand economy or mean annual increment was found as compared to conventional stand-level thinning. Despite the non-significant results, high-resolution maps are probably needed anyway to support forest workers in thinning operations as the observed stands were heavily thinned. The dominant tree species on pixel level was classified over a large forest holding (Paper IV) using Sentinel-2 with a high overall accuracy of 88.2 %. This illustrate the usefulness of Sentinel-2 and that a practical implementation is feasible.

I conclude that there is overall a lot to gain from basing decisions on the within-stand variation and in the implementation of silviculture. The Sentinel-2 system proved its relevance for practical forestry, since monitoring of thinning response, thinning detection and tree species maps will reduce uncertainties for the management of planted forest.