Invasive earthworms unlock arctic plant nitrogen limitation

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.


Supplementary Figures
Supplementary Figure 1. Experimental setup of the mesocosm experiment. Platforms (black rectangles) where mesocsoms with heath (grey rectangles) and meadow (green rectangles) vegetation were installed. Mesocosm ID-numbers are shown within each mesocosm, while numbers in the bottom right corner of each platform represent blocks used for our applied block-designed statistics. Platforms that received worm addition are surrounded by red lines and mesocosms that received labelled litter addition are surrounded by blue lines; other mesocosms received unlabelled litter. Location of loggers recording air temperatures and soil moisture (moist.) are also shown. Figure 2. Minirhizotron images illustrating fine root growth and deposition of earthworm casting. Images from the same area taken in a) June, b) August and c) September. Scale of the images are shown using black arrows. Note how earthworm castings (marked with white dashed lines) have formed between August and September and how considerable fine root (white arrow) growth has occurred during the same time period. Figure 3. Nitrogen forms in cast, litter and humus. Data shown include water leachable nitrogen forms from the earthworm invasion gradient at Jiebren described in detail by Wackett et al 2018 1 . Analyzed matrixes include fresh earthworm casts (<7days old), aged earthworm cast (seasonal age), litter and humus. (a) The concentration of ammonium (NH4) and proportion of total dissolved nitrogen (± std. err) that are in organic forms in earthworm cast in relation to litter and humus not subjected to earthworms. (b) Dissolved nitrate (NO3) in the same leachate as above (± std. err).

Supplementary Tables
Supplementary Table 1

Supplementary Methods 1.
Species abundance (% cover) as determined by point intercept method was used as weight when calculating weighted plant community N content. The specific data used for the calculation are shown. Community N from other local studies was compiled from previous published data using plant functional group cover, besides the study of fertilization effects where biomass was used as weight. Community N and its response to increased temperature was derived from an experiment using open-top-chambers generating an increase in summer air temperatures of 1.5 °C chambers 2 and an altitudinal gradient representing a 3 °C increase in summer air temperatures 3 . Data from the effects of herbivory was compiled from an experiment where fences excluded grazing by voles or both voles and reindeer 2 . In the fertilization experiment, reindeer feces were added at dosages corresponding to about double and four times natural abundance 4 .
NDVI was derived from the same experiments and an additional experiment including warming experiments (open top chambers) from two elevations 5 . Here, previously unpublished NDVI from these studies are found in Table S3. Note that differences in NDVI between control sites, located at an altitude of 500 and 900 m.a.s.l. were used to reflect difference in greeness between these two altitudes and thus, this measure includes long-term effects.
The community-N (Ncom.) was calculated as: where P is the relative coverage (%) for plant functional groups in subscript (gram = graminoids; forb = Forbs; everg = evergreen shrubs; and decid = deciduous shrubs) and N is the measured N content (%) of the same functional groups (in subscript).