Methodology:

The following experiment description is based on Moser (2025):

General tracer experiment information:

We conducted in-situ melt experiments that have been designed to mirror short ROS-induced melt events and to trace percolation via visual features corresponding to structural imprint, thermal changes and SWI alteration of the snowpack.

The melt tracer is based on ultrapure (Milli-Q) water mixed with deuterium oxide (D2O) and blue food dye, following the procedure laid out in Moser (2025), to enable visually and chemically tracing percolation. The percolation tracer of known composition is applied to a defined surface area (40 x 20 cm) of the already assessed snow pit (data available: Moser et al., 2025). The prepared tracer characteristics (volume, temperature, composition) are known before it is evenly distributed onto the snow surface using a sprinkler.

We documented the specifics for all tracer percolation experiments in the respective snowpit, including:

Date: DD.MM.YYYY;

Experiment number: consistent across temperature monitoring, percolation depth and SWI results to enable linking of records;

Liquid volume: in millilitres, measured in the NERC Arctic Station lab at the start of the field day;

Liquid temperature: in deg C, measured using a digital thermometer immediately before application in the field;

SWI Tracer ID: system-internal identifier assigned from the fieldwork vial until data processing to enable continuous data tracking;

Tracer d2H and Tracer d18O: the d2H and d18O signature of the tracer liquid, both reported in permil (ppth) against V-SMOW;

Pouring time: is the number of seconds it took to apply the entire tracer liquid to the experiment area. Note: if the pouring time was, e.g., 15 seconds, this corresponds to the first 0.25 decimal minutes of the experiment in the temperature monitoring record.

Thermometer positions: each thermometer is numbered (consistent with the temperature monitoring record in a separate CSV file) and its respective position in the snow pit reported as height above the ground in centimetres.

Note that experiment no. 2 and no. 3 for the snowpit dug on the 25th of March were conducted on 26th March.

Temperature monitoring:

Digital thermometers were placed in regular intervals below the surface, e.g. every 2 cm in the top 10 cm, and used to record snow temperature at regular intervals throughout the percolation observation period. Experiments were concluded when temperatures in the affected snowpack, as reflected by the top thermometer, started to drop, thereby indicating the predominance of refreezing.

In the related CSV file, the first column identifies the experiment number. The second column gives the time since the start of the experiment in decimal minutes. Where values are provided for the 0-minute time step, this shows the temperatures at the start of the experiment. In the subsequent columns, temperature readings are provided for each numbered thermometer at each time-step. The number of thermometers used varies between snowpits, but is consistent within experiments.

Percolation result depths:

Upon completion of the percolation experiment, we used the blue colour of the tracer to measure the average and maximum percolation depth on the snow pit front. These measurements were repeated on two new snow pit fronts that were cut 5 cm and 10 cm into the experiment area to evaluate and ensure representativity. Maximum percolation depth is the greatest depth reached by blue dye during an experiment, and average percolation depth reflects surface melt layer thickness on the respective front.

The Percolation Results table contains:

Experiment number: given to allow linkage with corresponding "Tracer_Experiments_Info" CSV in the same folder.

For each of the snow pit fronts (0 cm = Front_0, 5 cm = Front_5, 10 cm into the experiment area = Front_10), we provide averag...(33)

Data collection:

General:

Depth measurements: Foldable ruler.

Timekeeping: Stopwatch feature in Clock app (Samsung Galaxy A52s, Android).

Temperature:

Digital Thermometers: Salter 544A HBBKCR and Salter 518 WHCR, 0.1 deg C precision.

SWI:

Polypropylene (BIOFIL) centrifuge tubes.

Picarro L2130-i.

All measurements made in the field were immediately recorded in a field notebook and digitised upon return to the NERC Arctic Station the same day. The Picarro measurements are primarily processed using the LIMS software. All further data processing of the above records was conducted using Microsoft Office Excel 365.

Data quality:

The following measures were taken for quality control:

General snow pit setup: We always assessed a snow pit wall that was permanently in the shadow and downwind to prevent artefacts through incoming solar radiation or snow re-deposition.

Percolation depth: Measurements of average and maximum percolation depth were repeated on at least 3 fronts for each experiment to achieve representative results.

Depth readings: percolation estimates and SWI samples taken below the surface after experiments are corrected for surface compaction and reported against the snow surface prior to the experiment.

SWI Sampling: Of each prepared tracer liquid, a small sample was kept and measured for its SWI signatures to ensure comparability in composition.

Empty cells: Where information could not be obtained or has been excluded due to erroneous data points, the cells in the CSV tables are filled with "na".