Using forest history and spatial patterns to identify potential high conservation value forests in Romania

Abstract

Naturally dynamic forests have a high proportion of biotopes with old large trees, diverse vertical and horizontal structure at multiple scales, and much dead wood. As such, they provide habitat to species and ecosystem processes that forests managed for wood production cannot provide to the same degree. Whether termed old-growth, ancient, virgin, intact, primeval or continuity forests, a major challenge and need is to map such potential high conservation value forest for subsequent inclusion in functional habitat networks for biodiversity conservation in forest landscapes. Given that the delivery time of natural forest properties is much longer than of industry wood, we explore the usefulness of using historical maps to identify forests that have been continuously present for 220 years (potential old-growth) versus 140 years (potential aging forest) in a case study in the Romanian Carpathian Mountains (see Online Resource 1). While the total forest cover increased by 35 % over the past two centuries, the area of potential aging and potential old-growth forest declined by 56 and 34 %, respectively. Spatial modelling of edge effects and patch size for virtual species with different requirements indicated an even greater decrease in the area of functional habitat networks of old-growth and ageing forest. Our analyses show that compared to simple mapping of potential high conservation forests, the area of functional habitat patches is severely overestimated, and caution is needed when estimating the area of potential high conservation value forests that form functional habitat networks, i.e. a green infrastructure. In addition, the landscape and regional scale connectivity of patches needs to be considered. We argue that the use of historical maps combined with assessment of spatial patterns is an effective tool for identifying and analyzing potential high conservation value forests in a landscape context.