Correction to: Ambio https://doi.org/10.1007/s13280-019-01198-7

In the original published article, some of the symbols ( , , , , △, ) in figure 1 were modified incorrectly ( , , , , , ) during the typesetting and publication process. The correct version of Fig. 1 is provided in this correction.

Fig. 1
figure 1

Maps showing the locations of lemming monitoring sites in (A) Circumpolar region and (B) Fennoscandia (delimited by the large grey square in panel A). Different shadings of green indicate the delimitations of bioclimatic subzones. The high Arctic, low Arctic and Subarctic are drawn according to CAFF (2013) and the Oroarctic, which is customarily pooled with the arctic tundra in global biome overviews and represents potential habitat for lemmings south of the Arctic proper, according to Virtanen et al. (2016). Numbers refer to the sites as listed in Table 1 and symbols indicate small rodent community composition. Time series of annual small rodent abundance estimates are presented for selected sites (graphs for the remaining sites are given in Fig S1). On the time series graphs, triangles represent lemmings (upwards and black for Lemmus, and downwards and white for Dicrostonyx), circles represent data not identified to species such as winter nests, qualitative indices or incidental observations, whereas grey squares represent voles (all species pooled). The colour of the y-axis indicates the data type: black refers to individuals caught per 100 trap-nights (6, 8, 41 and all Fennoscandia except 39), light blue to density in individuals per ha (13, 16, 17), green to winter nest density in nests per ha (11), orange to incidental observations (numbers seen per observer-day [1], or observer-hour [15]) and purple to qualitative indices (39, 44, 48, 49). Smoothed trend lines for the total abundance of lemmings are shown in light blue. 35H and 35 L refer to the highland and lowland series of Joatka (Table S3)

Full size image

The original article has been corrected.