Invasive earthworms unlock arctic plant nitrogen limitation

Abstract

Arctic plant growth is predominantly nitrogen (N) limited. This limitation is generally attributed to slow soil microbial processes due to low temperatures. Here, we show that arctic plant-soil N cycling is also substantially constrained by the lack of larger detritivores (earthworms) able to mineralize and physically translocate litter and soil organic matter. These new functions provided by earthworms increased shrub and grass N concentration in our common garden experiment. Earthworm activity also increased either the height or number of floral shoots, while enhancing fine root production and vegetation greenness in heath and meadow communities to a level that exceeded the inherent differences between these two common arctic plant communities. Moreover, these worming effects on plant N and greening exceeded reported effects of warming, herbivory and nutrient addition, suggesting that human spreading of earthworms may lead to substantial changes in the structure and function of arctic ecosystems.

Fig. 1. Documented presence of geoengineering earthworms in the Arctic.

Sites (red circles) in which Lumbricus sp. and Aporrectodea sp. have been found (utilizing the mineral soil as a habitat) in the arctic biome (purple shading) divided into three major geographical zones (sub-, low-, and high-arctic). Sites are compiled from previously published studies^,,,– and findings presented in the Global Biodiversity Information Facility database. Red rectangle indicates a reported finding of L. rubellus in Greenland where the specific site was not reported. The figure does not intend to provide a complete overview of known populations of geoengineering earthworms, but to illustrate that they can survive in the Arctic (mainly the sub-arctic zone). A common denominator is that earthworms occur adjacent to human introduction points. Thus, the map illustrates that human mediated introductions occur at circumpolar scale. The underlying map showing geographic areas of the Arctic is derived from the Arctic Biodiversity Assessment (http://grida.no/resources/6264).

Fig. 2. Earthworm effects on nitrogen uptake of Festuca ovina.

Shown are earthworm effects on N content (%) and δ¹⁵N of Festuca ovina, a plant present in both of the studied vegetation types, i.e., heath (H) and meadow (M). Individual replicates of samples from unlabeled (white diamond shape) and δ¹⁵N labeled (red circle, where darker values indicate overlapping data) mesocosms are shown with average values (black square, ±std. err). Main effects and interaction effects are presented as text for each panel (effect of the labeled litter is indicated using the ¹⁵N symbol) along with symbols indicating significance level (P < 0.05 are shown using *, and significance levels <0.001 are indicated as ***). Note the cut off in the y-axis to show the δ¹⁵N signatures of the unlabeled and the labeled litter. Source data are provided as a Source Data file.

Fig. 3. Earthworm effects on nitrogen uptake of common heath and meadow species.

Nitrogen content and δ¹⁵N signatures of dwarf shrubs found only in heath (V. vitis-idea and V. myrtillus) and forbs found only in meadow (S. alpina and B. viviparia) in response to geoengineering earthworms. Note that these plant species were only found in one of the two vegetation types and thus did not allow comparison of effects between vegetation types. Individual replicates of samples from unlabeled (white diamond shape) and δ¹⁵N labeled (red circle, where darker values indicate overlapping data) mesocosms are shown with average values (black square, ±std. err). Main effects and interaction effects are presented as text for each panel (effect of the labeled litter is indicated using the ¹⁵N symbol) along with symbols indicating significance level (P < 0.05 are shown using *, and significance levels <0.001 are indicated as ***). Note the cut off in the y-axis to show the δ¹⁵N signatures of the unlabeled and the labeled litter. Source data are provided as a Source Data file.

Fig. 4. Aboveground plant growth responses to earthworm additions.

Boxplot showing a the maximum height and b the highest number of floral shoots (lower panel) of graminoids (D. flexuosa and F. ovina) per mesocosm. Diamond box indicate the 25%- to 75%-percentiles (whiskers show 99% percentile) and black rectangle the median value. Mesocosms lacking the studied plants are given a value of 0. Significant effects are indicated with *(P < 0.05). Source data are provided as a Source Data file.

Fig. 5. Belowground plant growth response to earthworm additions.

Boxplot showing the fine root length at the onset of the experiment before the earthworm addition, and the fine root growth after earthworm addition in heath and meadow vegetation (mesocosm with labeled and unlabeled litter are pooled together). Periods when growth was measured are shown at the top. Diamond box indicate the 25%- and 75%-percentile (whiskers indicate the 99% percentile) and black rectangle show the median value. Also shown are all individual measurements (red diamonds). Treatment effects are written at the base of the panels and the level of significance indicated with *(P < 0.05) and **(P < 0.01). Source data are provided as a Source Data file.

Fig. 6. Effects of earthworms compared to other environmental drivers in a changing Arctic.

Shown are a conceptual comparison of effects (percentage change relative to average of control) caused by earthworms (worming) with effects seen in previous local experiments in sub-arctic northern Sweden simulating arctic environmental change including warming, herbivory, and fertilization. a Effects (mean ± std. err) on plant community N expressed as a function of vegetation type (BF = sub-arctic birch forest, H = heath, and M = meadow). Nitrogen data for warming are from open top chambers and from an alpine altitudinal gradient, the herbivory experiment was conducted using fences excluding voles and reindeer (shown effects are illustrating when grazers are present in relation to the ungrazed control) and reindeer feces were used in the fertilization experiment, where additions corresponded to about double and four times natural abundance. b Effects (mean ± std. err) on NDVI (i.e., greenness) expressed as a function of vegetation type. Experiments are similar as above with the addition of data from an open top chamber experiment simulating a mean annual temperature increase of 1 °C conducted at both an altitude of 500 and 900 m.a.s.l. Source data are provided in Supplementary Table 3, Supplementary Methods 1 and as a Source Data file.