This study explores the long-duration (0.8-2.3Ma), time-averaged micrometeorite flux (mass and size distribution) reaching Earth, as recorded by the Transantarctic Mountain (TAM) micrometeorite collection. We investigate a single sediment trap (TAM65), performing an exhaustive recovery and characterization effort and identifying 1643 micrometeorites (between 100-2000μm). Approximately 7% of particles are unmelted or scoriaceous, of which 75% are fine-grained. Among cosmic spherules, 95.6% are silicate-dominated S-types, and further subdivided into porphyritic (16.9%), barred olivine (19.9%), cryptocrystalline (51.6%) and vitreous (7.5%). Our (rank)-size distribution is fit against a power law with a slope of -3.9 (R²=0.98) over the size range 200-700μm. However, the distribution is also bimodal, with peaks centered at ~145μm and ~250μm. Remarkably similar peak positions are observed in the Larkman Nunatak data. These observations suggest that the micrometeorite flux is composed of multiple dust sources with distinct size distributions. In terms of mass, the TAM65 trap contains 1.77g of extraterrestrial dust in 15kg of sediment (<5mm). Upscaling to a global annual estimate gives 1,555 (±753) t/yr – consistent with previous micrometeorite abundance estimates and almost identical to the previous South Pole Water Well flux estimate (~1,600 t/yr) and potentially suggesting minimal variation in the background cosmic dust flux over the Quaternary. The greatest uncertainty in our mass flux calculation is the accumulation window. A minimum age (0.8Ma) is robustly inferred from the presence of Australasian microtektites, while the upper age (~2.3Ma) is loosely constrained based on ¹⁰Be exposure dating of glacial surfaces at Roberts Butte (6km from our sample site).

We supply two datasets, both as excel spreadsheets, the first dataset are micrometeorite statistics from the TAM65 collection, the second dataset are wind data from a nearby weather station (Paola's (AWS) data from Talos Dome (40 km West of Miller Butte on the plateau side).

The TAM 65 data:
For this study we collected 15kg of loose sediment and investigated a subsample with a weight of 2540g. This was split into several size fractions (<63μm [12g], 63-100μm [33g], 100-200μm [149g], 200-400μm [90g], 400-800μm [179g], 800-2000μm [941g] and >2000μm [1138g]) using a series of uncontaminated cascade sieves. Each size fraction was then passed through a fine-grained magnetic mineral separator (three times per fraction), producing magnetic and non-magnetic portions. For each size fraction (and mag/non-mag subdivision) either the entire separation or a subsampled aliquot was then searched in an exhaustive picking effort. We ensured that each sample was searched completely by both researchers (MDS and LF), while also collecting particles with a range of optical properties (dark, irregular-shaped, rounded, smooth, fluffy, glassy, metallic, etc.) these efforts aimed to minimize the number of unpicked micrometeorites – thereby minimizing human biases affecting the collection effort.  All potential micrometeorites were imaged optically (external surfaces) and under BSE (back-scattered electron imaging for both their external and internal sectioned perspective).

Extraterrestrial identification: Micrometeorite identification was based on well-established textural criteria (e.g. the presence of igneous rims, magnetite rims and dehydration cracks, rounded to spherical particle morphologies) and geochemical properties (chondritic bulk compositions, Ni-bearing metal and high CaO, Cr₂O₃ olivines) as defined in Genge et al., (2008). Micrometeorites were classified following the system of Genge et al., (2008) and applying the updated suggestion established in van Ginneken et al., (2017) [e.g. the additional classification of μPO-type particles, which consist of numerous subhedral olivine crystals, typically <10μm in size and abundant small vesicles]. These data were used to determine population statistics, including relative abundances of different micrometeorite subtypes.

In total we picked 1643 micrometeorites, ranging in size between 100-1500μm. However, because partial searches were performed on the smaller (<400μm) size fractions the total expected micrometeorite population accounting for the unpicked micrometeorites lies at 3310 particles (>100μm) among the 2540g of sediment.

Size distribution: Following positive extraterrestrial identification, micrometeorite particle sizes were measured using the ImageJ software (Schneider et al., 2012). We analyzed their BSE external images to accurately characterize particle sizes. These data were used to calculate a whole-collection rank-size distribution (fitted against both a power law and an exponential function) and size distributions for each textural subtype of cosmic spherule.

Particle densities: We also calculated the densities for 57 cosmic spherules, with near-perfect spherical morphologies and within the size range 200-400µm; this included two G-type spherules and 55 S-type spherules. Particle masses were weighed using a Mettler Toledo mass balance (model – XP6/Z), with a precision of ±0.001mg. Particle diameters were measured under SEM, using BSE imaging of their exterior surfaces, as described above, these measurements have a precision of ±10µm. Particle volumes were then calculated using the volume of an oblate spheroid equation (v = 4/3π•b²•c [where b and c are the semi-major and semi-minor axes respectively]) and subsequently used to determine mass values for each particle.

Mass flux calculation: In addition, we calculated a mass flux estimate of cosmic dust falling to Earth per year, for comparison against existing estimates. This was achieved by estimating the total extraterrestrial mass preserved in the TAM65 trap (converting the micrometeorite abundance data into mass data, using both empirical [self-derived] and literature density constraints) and then upscaling this mass flux to a global flux (based on the area-size of the TAM65 trap and the global surface area) and finally dividing our upscaled mass value by the estimated age of the micrometeorite collection (using both the min and max age limits). Finally, we calculate the relative abundances of well-defined textural features previously identified in the literature including the presence of Fe-Ni metal and metal-sulfide beads (Genge and Grady, 1998), cumulate layering textures (Genge et al., 2016) and tailed or dumbbell-shaped morphologies (among melted cosmic spherules [Suttle et al., 2018]).

The Wind data:  Provides wind direction and strength data (over 58,000 indvidual records) from the Antarctic automated Weather Station (AWS) Paola at Talos Dome (~40 km due ESE of Miller Butte) with data recorded since 2003. This was used to gauge the effect of strong winds in (temporarily) preventing micrometeorite accumulation.

TAM65 data:
Sheet A - raw data from picked micrometeorites (the real or "empirical" population)
Sheet B - both the picked and expected micrometeorites (synthetic population)
Sheet C -  Population statstics for the TAM65 collection 
Sheet D - Calculating the mass contribution from small micrometeorites (50-100um)
Sheet E - Calculated masses for each particle in the 100-2000um size fractions.
Sheet F - Global time-averaged mass flux calculation
Sheet G - Peak fitting analysis of the size distribution data from TAM65 and Larkman Nunatak

Wind data: 
Sheet A - Input data
Sheet B - Wind rose diagram
Sheet C - Average wind speeds
Sheet D - Direction distribution
Sheet E - Speed distribution