Olha Nahorna defended her PhD thesis «Assessing the importance of improved forest data for decision-making processes» September 26th, 2025.

Value of information assessment, adapted from Ståhl (1994) and Eid (2000)

The topic of the trial lecture was «Climate change impacts on forest planning and implications for information needs».

Thesis abstract

The availability and quality of forest inventory data are critical for ensuring effective planning decisions. The errors in the inventory data could propagate through simulation and optimization models, leading to suboptimal planning decisions and potential economic and ecological losses. While improved data acquisition and processing approaches are continually being developed, the key question remains: how useful are these improvements for the decision-making? This thesis addresses this question through the value of information (VoI) assessment, with a focus on extending existing VoI frameworks to account for multiple planning objectives, decision-makers’ preferences, and uncertainty in the inventory data.

Typical VoI assessments focus on a single economic criterion, such as net present value. This thesis expands that framework by exploring multi-objective VoI using different formulations of the optimization models.  To explore how VoI can change with the decision-makers’ preference information, various approaches to integrate preferences were assessed. This included simple preference parameters, explicit numerical targets, or relative distance to the ideal values. In two studies, the uncertainty in both the studied and reference data were accounted for through the use of stochastic programming. The concept of value of improved information (VoII) was introduced to capture that reference data does not need to provide perfect information.

The thesis is organized around three studies, each addressing a relevant challenge related to data acquisition or processing approaches in Norwegian forest planning. The first study evaluated the VoI of using cell-level inventory data, as opposed to stand-level aggregated data. The results demonstrated that using cell-level data as basic unit for simulation and optimization, subject to segment- or stand-boundary constraints, consistently improved the quality of the decisions, in comparison to the use of conventionally aggregated stand-level data. The second study applied a VoII analysis to compare airborne laser scanning-based inventory approaches, demonstrating that more accurate approaches offer greater flexibility for decision-makers, particularly when planning priorities may shift over time. Less accurate inventory approaches may still be cost effective in cases where only a single objective is considered. The third study used VoII to assess different site index determination approaches and found that, in certain cases, lower-cost data sources can support decision-making outcomes comparable to those achieved with more expensive alternatives.

Overall, the findings highlight that the VoI is context-dependent, shaped by planning objectives, uncertainty, and decision-makers’ preferences. High-quality data may lead to significantly improved decisions in some cases, while offering little to no added value in others. VoI assessment proves to be an effective tool for comparing and evaluating different data acquisition or processing strategies. This set of studies recommends decision-makers to integrate VoI assessment, especially when expanded to multi-objective and uncertainty-aware framework, into the process of evaluation of new data acquisition and processing approaches.

Supervisors:
Kyle Eyvindson, Norwegian University of Life Sciences (NMBU)
Terje Gobakken, Norwegian University of Life Sciences (NMBU)

Evaluation committee:
Göran Ståhl, Swedish University of Agricultural Sciences
Olalla Díaz-Yáñes, ETH Zurich
Erik Næsset, Norwegian University of Life Sciences (NMBU)