Effects of global warming are most pronounced in winter. A reduction in snow cover due to warmer atmospheric temperature in formerly cold ecosystems, however, could counteract an increase in soil temperature by reduction of insulation. Thus, soil freeze-thaw cycles (FTC) might increase in frequency and magnitude with warming, potentially leading to a disturbance of the soil biota and release of nutrients.

Here, we assessed how soil freeze-thaw magnitude and frequency affect short-term release of nutrients in temperate deciduous forest soils by conducting a three factorial gradient experiment with ex-situ soil samples in climate chambers. The fully-crossed experiment included soils from forests dominated by Fagus sylvatica (European beech) that originate from different winter climate (mean coldest month temperature range ΔT > 4 K), a range of FTC magnitudes from no (T = 4.0 °C) to strong (T = -11.3 °C) soil frost, and a range of FTC frequencies (f = 0–7). We hypothesized that higher FTC magnitude and frequency, respectively, will increase the release of nutrients. Furthermore, soils from cold climates with historically stable winter soil temperatures due to deep snow cover will be more responsive to FTC than soils from warmer, more fluctuating winter soil climates.

FTC magnitude and, to a lesser extent, also FTC frequency resulted in increased nitrate, ammonium, and phosphate release almost exclusively in soils from cold, snow-rich sites. The hierarchical regression analyses of our three-factorial gradient experiment revealed that the effects of climatic origin (mean minimum winter temperature) followed a sigmoidal curve for all studied nutrients and was modulated either by FTC magnitude (phosphate) or by FTC magnitude and frequency (nitrate, ammonium) in complex two- and, for all studied nutrients, in threefold interactions of the environmental drivers. Compared to initial concentrations, soluble nutrients were predicted to increase to 250 % for nitrate (up to 16 µg NO₃-N kg^-1DM), to 110 % for ammonium (up to 60 µg NH₄-N kg^-1DM), and to 400 % for phosphate (2.2 µg PO₄-P kg^-1DM) at the coldest site for strongest magnitude and highest frequency. Soils from warmer sites showed little nutrient release and were largely unaffected by the FTC treatments except for above-average nitrate release at the warmest sites in response to extremely cold FTC magnitude.

We suggest that currently warmer forest soils have historically already passed the point of high responsiveness to winter climate change, displaying some form of adaptation either in the soil biotic composition or in labile nutrient sources. Our data suggests that previously cold sites, which will lose their protective snow cover during climate change, are most vulnerable to increasing FTC frequency and magnitude, resulting in strong shifts in nitrogen and phosphorus release. In nutrient poor European beech forests of the studied Pleistocene lowlands, nutrients released over winter may be leached out, inducing reduced plant growth rates in the following growing season.

Here, we provide the raw data of a three-factorial regression experiment and the R-Code used in the hierarchical regression analysis of this raw data in Kreyling et al.: Soils from cold and snowy temperate deciduous forests release more nitrogen and phosphorus after soil freeze-thaw cycles than soils from warmer, snow-poor conditions.

for details see publication Kreyling et al.

The R-script reads in the provided raw data and contains the full hierarchical regression analysis as presented in the published paper including additional R-packages and creation of the figures. The raw data is the same dataset as used in the publication.