Online or in Stockholm
Together with the Swedish EPA we are inviting to a seminar where we present a national (Swedish) model for predicting the relative likelihood for forest with high conservation value.
The seminar will be held in Swedish. Abstract in English follow below
Abstract: To conserve forest biodiversity, it is imperative to maintain and restore sufficient amounts of functional habitat networks. Therefore, the location the remaining forests with natural structures and processes over landscapes and large regions is key information. We integrated machine learning (Random Forest) and open landscape data to scan all forest in Sweden with a 1ha spatial resolution with respect to the relative likelihood of harboring High Conservation Value Forests (HCVF), based on data from each 1 ha patch and its surrounding forests. Using independent spatial stand- and plot-level validation data, we confirmed that our predictions (ROC AUC in the range of 0.89 - 0.90) correctly represent different levels of forest naturalness, continuously from high to low likelihood values and intensively managed to natural forests. Given ambitious national and international conservation objectives in forest landscapes with ongoing forestry, our model and the resulting wall-to-wall mapping fill an urgent gap for assessing the achievement of evidence-based conservation targets, spatial planning, and forest landscape restoration.
Jakub Bubnicki, Per Angelstam, Grzegorz Mikusiński, Johan Svensson, Bengt Gunnar Jonsson