Aeolian samples were collected by the McMurdo Dry Valleys Long Term Ecological Research program (MCM LTER) using the Big Spring Number Eight (BSNE) isokinetic wind samplers at four sites within Taylor Valley. Further details on the collection process can be found in Diaz et al. (2017). In total, 55 aeolian sediment samples were stored in Whirl Pak or 50 ml Falcon tubes and housed at the University of Colorado, Boulder. For this specific experiment, 12 aeolian samples were selected that encompass a range of environments and conditions within the valley. Among these, two samples were collected at Taylor Glacier during the austral summer of 2014; five samples were collected at East Lake Bonney, with four collected during the austral summer of 2016 and one during the austral summer of 2014; one sample was collected at Lake Fryxell during the austral summer of 2013; and four samples were collected at Explorers Cove during the austral summer of 2016.

In this experiment, two distinct washing procedures were used for all equipment. Procedure 1 was used for cleaning equipment intended for the analysis of ions and nutrients. The process involved 5 to 10 rinses of 18.2 Mohm ultra-pure Milli-Q (MQ) water for low density polyethylene (LDPE) bottles, tubes, one filter tower, and crucibles. Procedure 2 was used for cleaning equipment for the analysis of trace metals and took place inside of a clean room. This procedure began with an initial rinse with MQ water for all bottles, tubes, and one filter tower. Falcon tubes were subsequently filled with 10% HNO₃ for 24 hours (in contact with the sample for <30 minutes), while the filter tower underwent a similar treatment, being immersed in 10% HNO₃ for two weeks after the initial rinse.

Following the initial rinse, bottles used for sample storage and the freeze-thaw cycles underwent a longer procedure, modified from GEOTRACES (Cutter et al., 2017). In summary, the bottles were filled halfway with 1% Citronox detergent and left upright for four days. After four days, the bottles were inverted for an additional four days. After eight days, the bottles underwent five rinses with MQ water to remove all residual detergent. Subsequently, the bottles were filled halfway with 10% HNO₃ and positioned upright in secondary containers for seven days. Following this period, the bottles were inverted for another seven days. The 10% HNO₃ was then discarded into waste containers, and the bottles were rinsed five times with pH 2 water. Finally, the bottles were carefully placed in double Ziploc bags. Throughout Procedure 2, precautions were taken to prevent any contact with the inside of the bottles. If contact occurred, the cleaning process was restarted from the beginning.

Sample Processing

All 12 sediment samples underwent initial homogenization in their original containers using an inversion technique, involving ten inversions. Next, the samples were dried in MQ water-cleaned ceramic crucibles in a drying oven at 125 °C for 24 hours.

Freeze-Thaw

All samples, each containing at least 1 g of sediment, underwent a standardized processing method. 1 g of the sample was carefully placed into a 50 ml Falcon tube, which had been either acid-washed or rinsed with MQ water. A 1:50 sediment-to-water leach ratio was chosen to mimic leaching experiments done by Toner et al. (2013). To achieve a 1:50 sediment-to-water ratio, 50 ml of chilled (3 °C to 4 °C) MQ water was added to the Falcon tube using a MQ-rinsed or acid-washed squirt bottle. After this, all Falcon tubes containing samples were positioned on a shaker table for 10 minutes to resuspend the sediments, which represented initial leaching upon deposition on ice or snow.

Each bottle was labeled with the designated number of freeze-thaw cycles it would undergo. One cycle involved freezing for three hours and thawing for nine hours, resulting in a total cycle time of 12 hours. After each sample or control had completed the proposed number of freeze-thaw cycles, the bottles were prepared for analysis.

Centrifuge Method (INC: Ion/Nutrient Centrifuge & TC: Trace Metal Centrifuge)

Following the shaker table, the Falcon tubes were subjected to centrifugation (20 minutes at 3,000 RPM) for either the Ion/Nutrient Centrifuge (INC) or the Trace Metal Centrifuge (TC) methods. The resulting leachate was carefully poured into a bottle that had been either acid-washed or rinsed with MQ water. This was named the initial leach (IL), also known as cycle 0, and was stored for analysis at -20 °C for the remainder of the experiment. The sediment pellet from the centrifugation process was poured into LDPE bottles, also either acid-washed or MQ rinsed, and labeled for the freeze-thaw procedure. 50 ml of fresh MQ water was added to this freeze-thaw bottle.

Filter Method (INF: Ion/Nutrient Filter & TF: Trace Metal Filter)

Following the shaker table, the contents in the Falcon tubes for the Ion/Nutrient Filter (INF) and Trace Metal Filter (TF) methods underwent separation using a filter tower equipped with a GF/F 0.7 μm filter. Between filtering steps, the flasks and filter towers were rinsed with either MQ water or pH 2 water, depending on the sample being analyzed. This filter method was included in the experimental design to investigate whether using a GF/F filter to separate sediments from melted glacier ice and snow (a common technique) is sufficient for freeze-thaw experimentation. Apart from initial separation of sediment from leaching fluid by filtering as opposed to centrifugation, the filter method is identical to the INC and TC method above.

Controls

A total of 32 controls were used: 16 controls were for ion and nutrient analysis and 16 controls were used for trace metal analysis. The latter set used acid-washed bottles, differing from the former set, which used MQ-rinsed bottles. Despite this difference, the procedures for both sets of controls were identical. To investigate the effects of freezing, thawing, filtering for GF/F filters and nucleopore filters, and MQ water, a series of controls (Controls 1-12) were used. The first set of controls, denoted as controls 1-2 and 7-8, aimed to quantify solute leaching from the filters themselves. Controls 1 and 7 were subjected to a single freeze-thaw and Controls 2 and 8 were subjected to 60 freeze-thaw cycles, the maximum in the study. The second set of controls, denoted as controls 3-4 and 9-10, quantified any leaching from the LDPE bottles. Controls 3 and 9 were subjected to a single freeze-thaw and Controls 4 and 10 were subjected to 60 freeze-thaw cycles. The third set, controls 5-6 and 11-12, quantified contamination from the filtering process, using a GF/F or nucleopore filter, respectively. Controls 5 and 11 contained MQ water that had passed through a GF/F filter on a pre-cleaned filter tower. Control 6 and 12 replicated the process with a nucleopore filter. Controls 13 – 32 were labeled reactivity blanks and were used to create a reference point for reactivity of samples at 4°C in water over time. They served as a baseline for ongoing chemical weathering. The first subset of controls (controls 13, 15, 17, 19, 21, 23, 25, 27, 29, and 31) assessed the reactivity of dust samples exposed freely to water for varying durations at 4 °C. The ten controls were processed similarly to regular samples. They underwent varying freeze-thaw time cycles (one cycle is 12 hours) and were stored at +4 °C for different durations. Specifically, Controls 13 and 15 measured the reactivity after 1 freeze-thaw cycle (equivalent to 12 hours). Controls 17 and 19 mirrored Control 13 and 15 but measured the reactivity after 5 freeze-thaw cycles (equivalent to 60 hours). Controls 21-23, 25-27, and 29-31 mimicked 10 cycles at 120 hours, 30 cycles at 360 hours, and 60 cycles at 720 hours, respectively. The latter subset of controls (controls 14, 16, 18, 20, 22, 24, 26, 28, 30, and 32) mirrored the experimental conditions of the first set, but the sediments were bound to GF/F filters.

Preparing for Analysis

After completing the designated freeze-thaw cycles and control steps, the bottles were removed from the refrigerator. Rather than collecting a subset of the leachate for sequential analysis and subsequently refreshing the sample, an individual bottle was used for each freeze-thaw duration (i.e., for each sample, there was a separate bottle for 1 FT, 5 FT, etc.). This approach was selected to better represent the conditions encountered in the field, where sediment often remains in the same meltwater for extended periods in disconnected cryoconite holes (Fountain et al., 2004). Refreshing the samples was considered a potential source of error, as introducing the sample to new water could lead to changes in reactivity. Once a bottle was removed for analysis following the final thaw, the sample underwent a mixing period for five minutes on a shaker table before filtration. A filter tower, equipped with a 0.4 nucleopore filter, was used to separate the leachate from the sediment. The resulting leachate was collected into designated bottles and labeled as the final leach (FL). For trace metal analysis, the FL was acidified to 2% HNO₃ to solubilize metals and minimize adsorption effects. Subsequently, all FLs were stored in the freezer at -20 °C until the analysis phase.

Nutrients

A total of 168 ILs and FLs underwent analysis: 28 controls, 100 INC bottles, and 40 INF bottles. Each bottle contained approximately 50 ml of leachate available for analysis, divided into two aliquots of ~25 ml each. One aliquot was analyzed for Nitrite + Nitrate (NO₃^- + NO₂^-), Orthophosphate (PO₄^3-), Total Ammoniacal N (NH₃ + NH₄⁺), and Silicate (Si) using a Skalar San ++ continuous flow nutrient analyzer at the University of Colorado, Boulder. A set of standards were prepared for NO₃^- + NO₂^-, PO₄^3-, NH₃ + NH₄⁺, and Si. Any samples exceeding the standard ranges by 10% underwent an automatic 1:5 dilution. A seven-point calibration curve was established using the San++ Flow Access software for Windows to determine sample concentrations. To assess accuracy, the results from the USGS Inter-laboratory Calibration Standard 156 and 160 were analyzed as unknowns, and the concentrations were compared to the published values (USGS, 2022; USGS, 2023). The accuracy and precision of all measured analytes ranged from 2.00% to 39.18% for accuracy and 2.84% to 19.72% for precision.

Anions
The second aliquot was used for the measurement of major anions (F^-, Cl^-, Br^-, SO₄^2-) using a Dionex ICS Integrion ion chromatographer with an AS11 column at the University of Colorado, Boulder. All IL sample leachates were diluted by 1:5 and the FL were analyzed undiluted. The leachates underwent analysis with seven incrementally increasing standards, in addition to a USGS Inter-laboratory Calibration standard. Accuracy and Precision for all measured analytes ranged from 7.54% to 10.46% for accuracy and from 1.28% to 4.70% for precision. Bicarbonate (HCO₃^-) was estimated using charge balance (Welch et al., 2010: Take the difference between the total anion equivalents, including alkalinity, and total cations and divide by the sum of anion and cation concentrations in equivalents).

Cations and Trace Metals

A total of 168 ILs and FLs, distributed among 28 controls, 100 TC bottles, and 40 TF bottles were sent to The Ohio State University Trace Element Research Laboratory (TERL) for analysis. Cations (Na, K, Ca, Mg) and a suite of elements (Al, Ba, Co, Cu, Fe, Mn, Mo, Pb, Si, Sr, U, V, Zn, As, Cd, Cr, and Ni) were analyzed by inductively coupled plasma optical emission spectroscopy (ICP-OES) on a PerkinElmer Optima 8300 in axial mode. Precision for all measured trace metals ranged from 1.28% to 4.70%. Finally, the TERL regularly runs external calibration standards, and all analytes are typically within 10% accuracy.