Parcs

ID	TITEL	DATENTYP	AUTOR	JAHR
NPHT	Bibliographie NPHT	Project	Nationalparkrat Hohe Tauern	2013
Bemerkungen/Beschreibung
Abstract
Loss of biodiversity threatens the provision of ecosystem services and erodes the foundation
of civilization. As biodiversity monitoring is lacking, extent of global biodiversity crises and
thresholds for ecological collapses are largely unknown. Main reasons are severe knowledge
gaps in indicator choice and aggregation next to limited availability of resilient long-term data.
The aim of the doctoral project is to advance forest biodiversity assessments in Central Europe
by applying machine learning. Following research questions are targeted: (1) Is actual
European biodiversity monitoring reliable? (2) How can understanding of indicator-indicandum
relationships be extended? (3) How can forest biodiversity be assessed reliable in Central
Europe? Machine learning (''R randomForest'') provides new insights into indicatorindicandum
relationships and intercorrelation within indicator sets. The approach
demonstrates the forest stand characteristics indicated by forest structural biodiversity
indicators and highlights the importance of indicator choice. Negative binary generalized
models and generalized linear models prove that national biodiversity monitoring systems fail
to report actual biodiversity loss in Europe. Hence, a novel Biodiversity Composite index (BCI)
based on forest inventory and forest typing data is tested. BCI delivers high-resolution spatial
maps of ecosystem-, species-, genetic-, and biodiversity. Advantages of the BCI approach are
easy transferability, cost-efficiency, forest type rankings and a monitoring system in line with
the Convention on Biological Diversity. In the case study Tyrol, Central Europe, coniferous
forest types display higher potential to maintain biodiversity than deciduous and mixed forests.
BCI supports decision-making in forest policy (e.g., cost-benefit analysis), biodiversity
conservation (e.g., restoration priorities) and Sustainable Forest Management. Monitoring with
BCI can help to halt forest biodiversity loss on the national scale.
Eintrag Nr. 53825

Biodiversity assessments with machine learning based on forest inventory data
