The study extent includes mainland Norway.

The TFLs have originally been re-sampled in six different ways (Table 1): i) Most of the re-sampling has been carried out by in-situ re-mapping (79 entries), that is to re-visit locations during field-work. ii) Nine entries have been re-sampled through in-situ comparison of old and young forest. The age of the young forest at higher elevation, is then contrasted with the elevation of older forest at lower altitudes. iii) Eight entries were registered through comparing empirical TFLs with climatic TFLs, i.e. documenting elevational difference between lowest land use disturbed TFLs with uppermost climatic TFLs. iv) One entry was re-sampled by comparing the present day forest line with the historical forest line identified on an old photograph, whereas v) one forest line entry was re-sampled based on oral information on the elevation of previous forest line. vi) Two entries were re-sampled by comparing previous TFLs with an updated map.

The TFL locations which have been georeferenced, were spatially located through three steps using GIS (ArcGIS v. 10.3). First, the locality names provided by the authors were used to locate sites based on a query in the standard Norwegian topographic maps (NMA 2017). If locality names had changed, they were identified with the help of contemporary analog maps (historical maps). Second, we located the provided aspect on an aspect map derived from a 20 m digital elevation model over Norway. Third, we identified the most likely position at the reported elevation, by interpretation of both old and new aerial photos.

The georeferenced locations have two main potential sources of uncertainty: i) the uncertainty given by the original authors regarding the measurements of elevation and ii) the uncertainty regarding the exact spatial position along the combination of aspect and elevation.

The uncertainties have been reported in two ways: i) The altitudinal uncertainty (expected precision) has been divided into three categorical classes: high precision (± 5 m a.s.l.) reflects standard GPS quality, intermediate precision (± 10 m a.s.l.) reflects in-situ measurements using barometer, and comparisons of empirical and climatic forest lines, whereas low precision (± 25 m a.s.l.) reflects comparisons of old and young forest lines, comparison with old photo, map comparison, or where authors have reported uncertain measurements of elevation. ii) The uncertainty regarding coordinate precision has been reported as a vector length combining aspect and elevation at the sites. This uncertainty ranges between 160 and 5200 m, with an average of 1113 m.

All re-sampled records have been checked with aerial photos, and 76 of the locations (76 %) have been visited during field-work from 2013 to 2017. The oldest re-sampled records are more or less impossible to validate, since they are older than all available aerial photos and precise maps, and since most trees from that period are dead and lost through decomposition.

Twenty potential records have not been registered. Ten records were average numbers, and did not represent single locations, whereas the other ten records only showed the elevational change and not the elevation at the locations.

The TFLs have been re-sampled in the main mountain regions of Norway, but some regions have far more entries than others (Fig. 1). South-central Norway has most of the entries, whereas mid Norway and southeastern Norway lack entries. The coastal mountains have few and scattered entries. The dataset has been gathered within the uppermost parts of the boreal (forest line) and the lower parts of the alpine (treeline) bioclimatic regions (Bakkestuen et al. 2008).

Figure 1.  

Distribution of the re-mapped tree- and forest lines in Norway.

57°58'50'' and 71°11'8'' Latitude; 4°49'8'' and 31°3'26'' Longitude.

This dataset includes occurrence data from 1 species, mountain birch (Betula pubescens subsp. czerepanovii). There are two common synonyms, namely Betula pubescens subsp. tortuosa (Ledeb.) Nyman, and Betula pubescens var. pumila (L.) Govaerts. Misidentification of mountain birch (Betula pubescens subsp. czerepanovii), at the highest elevated locations, is highly unlikely. This is because it is the only known tree-forming birch species present at the boreal-alpine ecotone in Norway. In GBIF, the higher nomenclature was added to all records, from Kingdom to Subspecies.