Specification of threshold values for cultivation of tree species facing climate change using marginal occurrences (MARGINS)

Funded by the Bavarian State Forest Authority, the MARGINS project will forecast reactions of temperate tree species to increasing temperature and drought based on a combination of species distribution modeling and study of populations at the warm and dry edges of the distributions. Background Ecological niche models identify thresholds for physiological existence (i.e., fundamental niche) or observability of species (i.e., realized niche). The model output are probabilities of occurence, whereas the highest probabilities are found in the niche center, with probabilities converging to zero towards the margins. Validation with real observations reveals that beyond certain thresholds no occurrences of the species in question can be ascertained.

The specification of these threshold values for the commercially important tree species of Bavaria is the central scope of the MARGINS project. Aims Niche models applied to geographic areas result in species distribution models (SDM), where the probabilities of occurrence are mapped onto geographic zones. Occurrences just before the distributional margins represent populations under extreme climatic influence. Hence, symptoms induced by climatic change should become visible in these populations "at the rear edges" first. Studying these extreme populations is the core part of the MARGINS project, and SDM are used to identify interesting populations, and later to transfer the obtained results to future conditions in Bavaria.

Aiming at the combination of the strengths of both statistical niche modeling and effect-oriented case studies in a space-for-time approach, the MARGINS project aims to:

1. define the geographic zones where Bavaria's six most important tree species (Norway spruce, Scots pine, Silver fir, common beech, sessile oak and pedunculate oak) grow at or very near their distributional margins,
2. localize populations in these zones,
3. characterize the environmental conditions defining the boundaries of the distribution area,
4. validate, correct, and specify state-of-the-art niche models in the most critical warmer part,
5. derive thresholds for commercial use of the species,
6. identify reactions when reaching or transgressing critical thresholds,
7. and to apply the improved thresholds to diverse regional climate scenarios in Bavaria.

The HSWT workpackage within MARGINS studies vegetation composition and plant traits of tree populations. We collect plot data in ain the field and assemble a reference database using existing data from partners throughout Southern Europe.

The warm-dry range edge of the tree species was mapped using the newly developed "climatic marginality" model, which quantifies the ecological distance to the warm-dry niche edge (Mellert et al. 2015). Together with local forest experts, 45 study stands were identified in Romania, Bulgaria, Slovenia, Italy, France, Spain and Baden-Württemberg. On site, soil, ground vegetation, stand and physiologically relevant characteristics of the target tree species (annual ring structure, leaf morphology) were sampled. The vegetation data from beech forests were compared with a Bavarian reference data set on the basis of ordination and pointer species analysis.

In 2013 and 2014 45 MARGINS-plots have been sampled at the southern distribution edge of beech, spruce, fir, Scots pine and sessile oak. We found a total of 432 vascular plant species, among them 59 tree species. Species composition was highly similar to the one found in Bavarian forests. Site index (dominant height at 100 yrs.) indicated average growth conditions. Radial growth in Spain was limited by high summer temperatures and low precipitation of the previous year, with a diminishing influence of the north-atlantic oscillation. In 12 beech stands throughout the Mediterranean it was demonstrated, that radial growth chiefly depended on summer temperatures, with a diminishing influence of precipitation in the western, yet still significant effects in the central and eastern regions. In many plots a diminishing growth trend was detectable.

The concept of “site marginality” was elaborated in a paper submitted to the journal Forestry. Models of the niche edge are designed to define climatic limits of forest tree species more sharply, but they are limited by a lack of high resolution soil information. Based on the newly built occurrence database, the outer tolerance limit was defined as the most extreme climate (in terms of warmth and aridity), at which the specis occurs under favourable conditions of soil and relief. Based on the systematic sample of ICP Forests Level-I plots, the so-called “impact” was estimated from an ensemble of 63 climate change scenarios. Marginality proved to be a robust estimate of the sensitivity of a given situation (tree species, current climate) towards climate change. As a summary parameter, impact explicitly encompasses the diversity of scenarios and model uncertainty. Such a representation of climate risk improves decision support in tree species selection under climate change.