New Zealand Southern Alps Blanketed by Red Australian Dust During 2019/2020 Severe Bushfire and Dust Event
Abstract
Episodic deposition of light absorbing impurities on glaciers reduces albedo and exacerbates snow melt. In 2019/2020 a devastating Australian bushfire and desert dust event combined with favorable meteorological conditions transported an unprecedented mass of impurities across the Tasman Sea turning the Southern Alps of Aotearoa New Zealand red. Here we use time lapse cameras, airmass back trajectories, snow impurity geochemistry, and remote sensing to quantify the timing, provenance, and mass deposition of the event. Deposited in late November 2019, the impurities were dominated by mineral dust with a distinct southeastern Australian geochemical fingerprint. The event deposited ∼4,500 ± 500 tons of red dust to Southern Alps permanent snow and ice with a mean dust mass concentration of 6.5 ± 0.7 g m−2. A southeast Australian desert dust storm generated by the same type of meteorological conditions as the 2020 New Year bushfires was the main driver of the glacier discoloration.
Key Points
-
Dust deposition to Southern Alps occurred in late November 2019 from Australian bushfire and desert dust storm event
-
Dust geochemical fingerprint and airmass back trajectories pinpoint a southeastern Australian provenance
-
Approximately 4,500 tons of red dust are estimated to have been deposited to the Southern Alps
Plain Language Summary
In 2019/2020, snow on the Southern Alps of Aotearoa New Zealand changed color from white to red overnight. While New Zealand air quality dropped following the severe 2020 New Year Australian bushfire event, the glacier discoloration was caused by a desert dust storm and weather conditions that transported an extraordinary amount of red dust from southeastern Australia across the Tasman Sea in late November 2019. Understanding the source and amount of dust deposited to the Southern Alps is required to assess the impact on glacial melt and snow biology.
1 Introduction
Extraordinary bushfires occurred in southeastern Australia in late 2019 and early January 2020 (Adams et al., 2020; Boer et al., 2020) promoted by the co-occurrence of climate variability events (positive Indian Ocean Dipole, negative Southern Annular Mode and positive El Niño-Southern Oscillation) and exceptionally dry conditions (Abram et al., 2021; Wang & Cai, 2020). An unprecedented mass of biomass burning aerosol was injected into the tropopause via strong pyrocumulonimbus convection and caused the strongest bushfire-derived stratospheric aerosol perturbation observed on record (Peterson et al., 2021). The resulting Southern Hemispheric aerosol optical thickness exceeded even that of the Pinatubo volcanic eruption (Hirsch & Koren, 2021). The smoke, reaching an altitude of 30 km (Ohneiser et al., 2020), dispersed over the Southern Hemisphere strongly influencing the climate and environment downwind of the plume. Reported impacts include the thinning of the stratospheric ozone layer over Antarctica (Damany-Pearce et al., 2022; Ohneiser et al., 2022; Stone et al., 2021), the influence on climate via cooling over oceanic cloud-free areas (Hirsch & Koren, 2021; Stocker et al., 2021), the radiative forcing of the Southern Hemispheric aerosol layer (e.g., Sellitto et al., 2022; Senf et al., 2023), and fertilizing Southern Ocean phytoplankton blooms with limiting micronutrients (Tang et al., 2021; Weis et al., 2022).
Directly downwind of the plume, air quality across New Zealand was reduced and exceeded the New Zealand PM10 (particulate matter less than 10 μm in diameter) National Environmental Standards for Air Quality (Davy & Trompetter, 2022). The same type of meteorological conditions that pushed smoke from eastern Australia across the Tasman Sea toward New Zealand also generated desert dust storms further west. Over the late 2019 to early 2020 period, New Zealand experienced a series of deposition events. In particular, around 6 December 2020, highly elevated PM10 observed in the Auckland region was attributed to coarse mineral dust from an Australian dust storm (Davy & Trompetter, 2022), and on 22 November 2019 satellites measured a coarse mineral dust event over the Tasman Sea (Li et al., 2021). A bushfire smoke episode followed from December 2019 to January 2020 where the plume affected the Tasman Sea, passing over New Zealand to the South Pacific Ocean (Li et al., 2021; Tang et al., 2021).
A large volume of light absorbing impurities was deposited on the Southern Alps significantly changing the color of the glacier surface from white to red (Pu et al., 2021). At the time, global media reported the snow discoloration being due to ash from bushfires (e.g. Roy, 2020), deposited around New Year 2019/2020; as we will show, both of these assumptions were incorrect.
While this was a particularly devastating event, Trans-Tasman dust transport is well documented with large episodic outbreaks of Australian dust over the last one hundred years reported to periodically affect New Zealand (Hesse, 1994; Hesse & McTainsh, 2003; Marx & McGowan, 2005; Marx et al., 2005a; Marx et al., 2005b; McGowan et al., 2000; McGowan et al., 2005). In the current climate, the southeast dust pathway over Tasman Sea is most active between December and March, with a source region in the southeastern sector of the Lake Eyre Basin and the southern Murray Darling Basin (Hesse & McTainsh, 2003; Prospero et al., 2002).
Despite the global significance of the 2019/2020 bushfire/dust storm event, the input of material to New Zealand glaciers has not been quantified. Here we use dust geochemical and isotopic fingerprinting, supported by airmass back trajectories, to trace the provenance and timing of dust deposition on the Southern Alps. We then use satellite remote sensing alongside ground data to estimate the total mass deposited on areas of permanent snow and ice. Our study describing the distribution, mass, and geochemical composition of dust on New Zealand glaciers from the 2019/2020 Australian bushfires/dust storm is an important and necessary baseline for further studies of the effect of this event on the physics and biology of glacial systems.
2 Materials and Methods
A detailed description of the methods is available in Text S1 in Supporting Information S1. Briefly, surface snow samples were collected from the Tasman/Haupapa, Fox/Te Moeka o Tuawe, Franz Josef/Ka Roimata o Hine Hukatere and Brewster glaciers in the Southern Alps, New Zealand following the Australian 2019/2020 event (Figure 1) (Adams et al., 2020; Boer et al., 2020; Davy & Trompetter, 2022; Gabric et al., 2010; Pu et al., 2021). Samples containing red dust were collected in 2020 and 2021, and background dust samples were collected in February 2023 (i.e., once the 2019/2020 dust layer was buried) to compare the dust geochemical composition of the 2019/2020 event to a non-Australian bushfire year. Potential source area (PSA) samples were also collected from local geological deposits on the Fox and Tasman glaciers in February 2023 to characterize the geochemical fingerprint of local source material. Sample information is reported Table S1 in Supporting Information S1. In addition, we analyzed an historic dust sample of the 1939 Australian Black Friday bushfire event collected from a house roof at Golden Downs, Tasman District, New Zealand.

Deposition of red dust to the Southern Alps following the 2019/2020 Australian bushfire/dust storm. (a) South Island, New Zealand with key locations mentioned in text. Red box shows area covered in panel (b) Key sampling sites in the Franz Josef, Fox and Tasman Glaciers. Red box shows area covered in panel (c) Snow sampling sites (white circles) in the Tasman Glacier accumulation area. Image is an orthophoto obtained via drone survey on 26/02/2020. (d) Sampling snow in the Fox Glacier accumulation area (Photo: J. Hunt 11/02/2020).
Utilizing the information from a timelapse camera installed at Brewster Glacier (Text S2 and Figure S1 in Supporting Information S1) combined with personal observations and published particle size data (Li et al., 2021), and observation of a local pilot [Greg Isbister, Helicopter Line Mt Cook, pers. Comm. 2023], we identified two key events. The first occurred in late November 2019, and the second over the New Year period. Airmass back trajectories were initiated at 2,500 m asl on the Tasman Glacier neve from 0000 UTC from 18 to 30 November 2019 and from 28 December 2019–5 January 2020 using the HYSPILT model (Stein et al., 2015).
To quantify the spatial extent and total dust mass deposition to the permanent snow and ice region of the Southern Alps, the dust mass was determined gravimetrically in snow samples in fine (0.2–5 μm) and coarse fractions (>5 μm) (Table S1 in Supporting Information S1). Non-linear regression, using a logarithmic equation of the form used by Di Mauro et al. (2024), was used to relate a spectral index from Sentinel-2 imagery to geolocated dust mass concentrations from the Tasman Glacier accumulation area, performance-tested using masses from the Franz Josef and Fox Glacier accumulation areas, and used to scale those point measurements across all areas of permanent snow and ice in the Southern Alps mapped by Baumann et al. (2021) (Figures S2–S3 in Supporting Information S1). Uncertainty was estimated by re-calculating all pixels using 95% confidence intervals and differencing from the final result.
Trace element concentrations and strontium (Sr) and neodymium (Nd) isotope ratios were determined on acid digested dust samples to fingerprint the dust source. The samples analyzed in this study were size-selected for comparison to provenance measurements made of dust from PSA. Trace elemental concentrations were analyzed by high resolution inductively coupled plasma mass spectrometry (HR-ICP-MS, Thermo Fisher Scientific Element 2; Table S2 in Supporting Information S1) and figures of merit, including the reproducibility of repeat measurements of certified reference materials are reported Table S3 in Supporting Information S1. Based on the measured Nd concentrations, dust samples from each glacier were combined where necessary to ensure that a minimum load size of 10 ng of Nd was available for isotopic analysis. Strontium and neodymium separation was carried out on aliquots of the same sample solutions used for the trace element work. Isotope ratios were measured using a Thermo-Finnigan TRITON Thermal Ionization Mass Spectrometer (TIMS; Text S1 in Supporting Information S1). Total procedural blanks for the isotopic work were <10 pg for Nd and Sr and therefore <1% of the total analyte of interest even for the smallest load size. Isotopic compositions and corresponding uncertainties are reported in Table 1 for 143Nd/144Nd and 87Sr/86Sr.
Location | Combined samples | Fraction sizea | 143Nd/144Nd | 2σmean (ppm)b | 143Nd/144Nd correctedc | 2σ (ppm)d | εNd(0)e | 87Sr/86Src | 2σmean (ppm)b | 2σ (ppm)d | ||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Glacier dust | ||||||||||||||
Franz 2020 | 8 | Coarse | 0.512186 | ± | 6 | 0.512184 | ± | 11 | −8.9 | 0.718917 | ± | 7 | ± | 11 |
Tasman 2020 | TS8 & TS2 | Coarse | 0.512201 | ± | 7 | 0.512199 | ± | 12 | −8.6 | 0.715534 | ± | 6 | ± | 11 |
Fox 2020 | 100 & 200 & 45 | Coarse | 0.512228 | ± | 6 | 0.512226 | ± | 11 | −8.0 | 0.713916 | ± | 6 | ± | 11 |
Brewster 2021 | A3 | Coarse | 0.512200 | ± | 9 | 0.512198 | ± | 13 | −8.6 | 0.717207 | ± | 6 | ± | 11 |
Brewster 2021 | C1 | Coarse | 0.512303 | ± | 7 | 0.512301 | ± | 11 | −6.6 | 0.713227 | ± | 6 | ± | 11 |
Tasman 2021 | A3 & C3 & A2 | Coarse | 0.512233 | ± | 6 | 0.512231 | ± | 11 | −7.9 | 0.713021 | ± | 7 | ± | 12 |
Tasman 2021 | B4 & B1 | Coarse | 0.512240 | ± | 7 | 0.512238 | ± | 11 | −7.8 | 0.713048 | ± | 8 | ± | 12 |
Tasman 2021 | C4 & B5 | Coarse | 0.512236 | ± | 6 | 0.512234 | ± | 11 | −7.9 | 0.713057 | ± | 5 | ± | 10 |
Tasman 2023 | 1∕3 | Coarse | 0.512321 | ± | 8 | 0.512319 | ± | 12 | −6.2 | 0.710276 | ± | 7 | ± | 11 |
Tasman 2023 | Clean 2 & Clean 3 | Coarse | 0.512324 | ± | 6 | 0.512322 | ± | 11 | −6.2 | 0.710118 | ± | 6 | ± | 11 |
Tasman 2023 | Clean 4 & Clean 5 | Coarse | 0.512325 | ± | 4 | 0.512323 | ± | 10 | −6.1 | 0.710159 | ± | 6 | ± | 11 |
Tasman 2023 | 1∕1 & 1∕2 | Coarse | 0.512330 | ± | 4 | 0.512328 | ± | 10 | −6.1 | 0.710366 | ± | 6 | ± | 11 |
Tasman 2023 | 2∕1 & 2∕2 | Coarse | 0.512323 | ± | 7 | 0.512321 | ± | 12 | −6.2 | 0.710275 | ± | 6 | ± | 11 |
Fox 2023 | 1∕3 & Clean 1 | Coarse | 0.512412 | ± | 4 | 0.512410 | ± | 10 | −4.4 | 0.709441 | ± | 6 | ± | 11 |
Fox 2023 | Clean 4 & Clean 5 | Coarse | 0.512410 | ± | 5 | 0.512408 | ± | 10 | −4.5 | 0.709478 | ± | 9 | ± | 13 |
Fox 2023 | 1∕1 & 1∕2 | Coarse | 0.512393 | ± | 8 | 0.512391 | ± | 12 | −4.8 | 0.709528 | ± | 7 | ± | 11 |
Fox 2023 | 5∕1 & 5∕2 | Coarse | 0.512413 | ± | 7 | 0.512411 | ± | 11 | −4.4 | 0.709483 | ± | 6 | ± | 11 |
Franz 2023 | 1∕3 | Coarse | 0.512344 | ± | 5 | 0.512342 | ± | 10 | −5.8 | 0.710119 | ± | 7 | ± | 11 |
Franz 2023 | Clean 1 & Clean 2 | Coarse | 0.512344 | ± | 9 | 0.512343 | ± | 13 | −5.8 | 0.709949 | ± | 6 | ± | 11 |
Franz 2023 | Clean 4 & Clean 5 | Coarse | 0.512343 | ± | 6 | 0.512341 | ± | 11 | −5.8 | 0.710026 | ± | 7 | ± | 11 |
Franz 2023 | 1∕1 | Coarse | 0.512348 | ± | 5 | 0.512346 | ± | 10 | −5.7 | 0.710225 | ± | 6 | ± | 11 |
Franz 2023 | 1∕2 | Coarse | 0.512352 | ± | 6 | 0.512350 | ± | 11 | −5.6 | 0.710066 | ± | 7 | ± | 11 |
Franz 2023 | 3∕1 & 3∕2 | Coarse | 0.512358 | ± | 7 | 0.512356 | ± | 11 | −5.5 | 0.710116 | ± | 10 | ± | 13 |
Franz 2020 | 8 | Fine | – | ± | – | – | ± | – | – | 0.713282 | ± | 52 | ± | 52 |
Tasman 2020 | TS8 & TS2 | Fine | – | ± | – | – | ± | – | – | 0.713489 | ± | 71 | ± | 72 |
Brewster 2021 | A3 | Fine | 0.512215 | ± | 24 | 0.512242 | ± | 34 | −8 | 0.715549 | ± | 53 | ± | 54 |
Brewster 2021 | C1 | Fine | 0.512245 | ± | 32 | 0.512272 | ± | 32 | −7 | 0.715024 | ± | 62 | ± | 62 |
Tasman 2021 | A3 & C3 & A2 | Fine | 0.512248 | ± | 26 | 0.512275 | ± | 36 | −7 | 0.712617 | ± | 42 | ± | 43 |
Tasman 2021 | B4 & B1 & C4 & B5 | Fine | 0.512240 | ± | 27 | 0.512267 | ± | 27 | −7 | 0.713761 | ± | 18 | ± | 21 |
Tasman 2023 | 1∕3 & Clean 2 & Clean 3 & Clean 4 & Clean 5 | Fine | 0.512347 | ± | 34 | 0.512374 | ± | 34 | −5 | 0.710339 | ± | 11 | ± | 14 |
Tasman 2023 | 1∕1 & 1∕2 & 2∕1 & 2∕2 | Fine | 0.512298 | ± | 20 | 0.512325 | ± | 20 | −6 | 0.710778 | ± | 75 | ± | 76 |
Fox 2023 | 1∕3 & Clean 1 & Clean 4 & Clean 5 | Fine | 0.512345 | ± | 27 | 0.512372 | ± | 27 | −5 | 0.709778 | ± | 26 | ± | 27 |
Fox 2023 | 1∕1 & 1∕2 & 5/1 & 5/2 | Fine | 0.512356 | ± | 24 | 0.512383 | ± | 24 | −5 | 0.709865 | ± | 17 | ± | 20 |
Franz 2023 | 1∕3 & Clean 1 & Clean 2 & Clean 4 & Clean 5 | Fine | 0.512343 | ± | 28 | 0.512370 | ± | 28 | −5 | 0.710222 | ± | 18 | ± | 21 |
Franz 2023 | 1∕1 & 1/2 & 3∕1 & 3∕2 | Fine | 0.512316 | ± | 27 | 0.512343 | ± | 27 | −6 | 0.710577 | ± | 21 | ± | 23 |
Franz 2020 | A & B | Coarse | –- | – | – | – | – | – | – | 0.718421 | ± | 5.9 | ± | 12 |
Franz 2020 | C & D | Coarse | – | – | – | – | – | – | – | 0.718997 | ± | 6.6 | ± | 13 |
Franz 2020 | E & F | Coarse | – | – | – | – | – | – | – | 0.718606 | ± | 6.1 | ± | 12 |
Fox 2020 | 6 & 13 | Coarse | – | – | – | – | – | – | – | 0.715558 | ± | 6.3 | ± | 12 |
Fox 2020 | 20 & 22 | Coarse | – | – | – | – | – | – | – | 0.718589 | ± | 6.3 | ± | 12 |
Fox 2020 | 27 & 28 | Coarse | – | – | – | – | – | – | – | 0.718619 | ± | 5.6 | ± | 12 |
Tasman 2020 | TS5 & TS9 | Coarse | – | – | – | – | – | – | – | 0.713968 | ± | 6.4 | ± | 12 |
Tasman 2020 | TS17 & TS18 | Coarse | – | – | – | – | – | – | – | 0.713222 | ± | 6.5 | ± | 13 |
Tasman 2020 | TS13 &TS14 | Coarse | – | – | – | – | – | – | – | 0.717023 | ± | 6.0 | ± | 12 |
Brewster 2021 | A4 & B3 | Coarse | – | – | – | – | – | – | – | 0.718225 | ± | 5.9 | ± | 12 |
Brewster 2021 | B1 & B4 | Coarse | – | – | – | – | – | – | – | 0.717213 | ± | 5.7 | ± | 12 |
Brewster 2021 | A2 | Coarse | – | – | – | – | – | – | – | 0.717036 | ± | 6.8 | ± | 13 |
Franz 2023 | 2/1 & 2/2 | Coarse | – | – | – | – | – | – | – | 0.709944 | ± | 6.2 | ± | 12 |
Franz 2023 | 4/1 & 4/2 | Coarse | – | – | – | – | – | – | – | 0.709916 | ± | 5.7 | ± | 12 |
Franz 2023 | 5/1 & 5/2 | Coarse | – | – | – | – | – | – | – | 0.709911 | ± | 5.5 | ± | 12 |
Fox 2023 | 2/1 & 2/2 | Coarse | – | – | – | – | – | – | – | 0.709442 | ± | 6.0 | ± | 12 |
Fox 2023 | 3/1 & 3/2 | Coarse | – | – | – | – | – | – | – | 0.709469 | ± | 5.9 | ± | 12 |
Fox 2023 | 4/1 & 4/2 | Coarse | – | – | – | – | – | – | – | 0.709495 | ± | 6.2 | ± | 12 |
Tasman 2023 | 3/1 & 3/2 | Coarse | – | – | – | – | – | – | – | 0.710166 | ± | 5.6 | ± | 12 |
Tasman 2023 | 4/1 & 4/2 | Coarse | – | – | – | – | – | – | – | 0.710151 | ± | 6.3 | ± | 12 |
Tasman 2023 | 5/1 & 5/2 | Coarse | – | – | – | – | – | – | – | 0.710210 | ± | 6.3 | ± | 13 |
Franz 2020 | A & B &C & D & E & F | Fine | – | – | – | – | – | – | – | 0.717729 | ± | 18.1 | ± | 21 |
Franz 2020 repeat | A & B &C & D & E & F | Fine | – | – | – | – | – | – | – | 0.717700 | ± | 14.2 | ± | 18 |
Fox 2020 | 6 & 13 & 20 & 22 & 27 & 28 | Fine | – | – | – | – | – | – | – | 0.716801 | ± | 15.4 | ± | 19 |
Tasman 2020 | TS5 & TS9 & TS17 & TS18 & TS13 &TS14 | Fine | – | – | – | – | – | – | – | 0.716731 | ± | 13.7 | ± | 17 |
Brewster 2021 | A4 & B3 & B1 & B4 & A2 | Fine | – | – | – | – | – | – | – | 0.716758 | ± | 11.7 | ± | 16 |
Franz 2023 | 2/1 & 2/2 & 4/1 & 4/2 & 4/1 & 4/2 | Fine | – | – | – | – | – | – | – | 0.710232 | ± | 7.5 | ± | 13 |
Fox 2023 | 2/1 & 2/2 & 3/1 & 3/2 & 4/1 & 4/2 | Fine | – | – | – | – | – | – | – | 0.709816 | ± | 13.3 | ± | 17 |
Tasman 2023 | 3/1 & 3/2 & 4/1 & 4/2 & 5/1 & 5/2 | Fine | – | – | – | – | – | – | – | 0.710255 | ± | 9.2 | ± | 14 |
Historical dust | ||||||||||||||
1939 Black Friday | – | Bulk | - | – | - | - | – | - | - | 0.715717 | ± | 6.0 | ± | 12 |
Potential source area | ||||||||||||||
Tasman Glacier | – | Coarse | - | – | - | - | – | - | - | 0.710457 | ± | 6.7 | ± | 13 |
Fox Glacier | – | Coarse | - | – | - | - | – | - | - | 0.709912 | ± | 6.1 | ± | 12 |
- a Fine 0.2–5 μm; Coarse >5 μm.
- b Internal precision, 2 standard errors of the mean.
- c Corrected to 143Nd/144Nd = 0.512115 for JNdi-1 standard (Tanaka et al., 2000) and 87Sr86Sr = 0.710250 for the NBS987 standard.
- d Uncertainty calculated by quadratic addition using the external and internal precision values.
- e Nd isotopic ratios expressed as epsilon units εNd(0) = [(143Nd/144Nd)sample/(143Nd/144Nd)CHUR-1]x104; CHUR, chondritic uniform reservoir with 143Nd/144Nd = 0.51263 (Wasserburg et al., 1981).
3 Dust Provenance
Both airmass back trajectories and dust geochemistry provide clear evidence for Australian-sourced dust deposited to the Southern Alps from the bushfire/dust event. Henceforth, “dust event” refers to the red dust collected in 2020 and 2021, while “background dust” refers to dust collected in white snow in 2023.
The observed trace elements in glacial dust are primarily of crustal origin as determined by calculating crustal enrichment factors following Winton, Edwards, et al. (2016) and using the Wedepohl (1995) compilation of the continental crust (Table S4 in Supporting Information S1). Several trace metals in the dust event (Cd, Co, Zn, Mo, Cr, Ni, Cu) were enriched indicating a pollution and/or non-crustal contribution. Enrichment of these trace metals was also observed in past Australian dust deposition events to the Southern Alps, attributed to Australia pollution sources (Marx et al., 2008), and in Australian biomass burning smoke plumes (Winton, Edwards, et al., 2016). Scanning electron microscope images of coarse angular particles further indicate the impurities are crustal in origin (Figure S4 in Supporting Information S1). We investigate the provenance of the crustal mineral dust using diagnostic geochemical fingerprinting based on Sr and Nd isotopes.
The Sr and Nd isotopic composition of dust is primarily related to lithology and geologic age of parent materials, although the Sr isotopic composition is also influenced by particle size. Australia is characterized by an extremely variable Sr and Nd isotopic composition (Delmonte et al., 2004; Gingele & De Deckker, 2005; Martin & McCulloch, 1999; Revel-Rolland et al., 2006). The 87Sr/86Sr (εNd(0)) values of New Zealand are higher (lower) than Australia allowing geochemical distinction of specific geographic source regions (Delmonte et al., 2004; Koffman et al., 2021).
The Sr and Nd isotopic composition of the glacier dust samples is reported in Table 1 and illustrated in Figure 2 with published isotopic data from New Zealand and Australian PSA. Additional glacier dust samples analyzed for 87Sr/86Sr only are illustrated in Figure S5 in Supporting Information S1. The background dust from Franz Josef, Fox and Tasman Glaciers displays isotopic compositions characteristic of the PSA material collected from those sites and the central South Island geology (Figure 2b). The geochemical fingerprint of the background dust and dust event is markedly different as evidenced by trace element ratios (Figure S6 in Supporting Information S1), and the Sr and Nd isotopic composition which provides clear evidence of different origins, that is, the Sr and Nd isotopic composition of the dust event (0.710811<87Sr/86Sr < 0.718997 and −8.9 < εNd(0) < −4.6) is isotopically distinct from the background dust (0.709441 < 87Sr/86Sr < 0.710778 and −6.2 < εNd(0) < -4.4) (Figure 2a). Size-dependent fractionation of 87Sr/86Sr between fine and coarse dust was observed, consistent with previous dust provenance studies where finer dust has relatively higher 87Sr/86Sr ratios compared to bulk samples (Figure S7 in Supporting Information S1) (e.g. Andersson et al., 1994; Winton, Dunbar, et al., 2016). This variability is less than the 87Sr/86Sr variability between the background dust and dust event. For example, the Δ87Sr/86Sr between fine and coarse Fox Glacier background dust is 0.000343, that is less than the Δ87Sr/86Sr of 0.005293 between Fox Glacier background dust and the dust event.

Strontium and neodymium isotope ratios of glacier dust and potential source areas. (a) Composition of dust samples collected as part of this study. The black line is a two-component mixing line between the end members of background dust and the dust event. (b) Comparison of the dust isotopic composition to potential source areas in New Zealand (Basile et al., 1997; Delmonte et al., 2004; Koffman et al., 2021) and Australia (De Deckker, 2019; Delmonte et al., 2004; Gingele & De Deckker, 2005; Grousset et al., 1992; Revel-Rolland et al., 2006).
The coarse dust event samples collected in 2020 fall within the isotopic envelope of an Australian source from the Murray Basin and South Australia (Figure 2b) which is consistent with airmass back trajectories crossing these two regions during the late November 2019 dust event (Figure S8 in Supporting Information S1). Similarly, the fine dust event samples collected in 2020 exhibit a Murray Basin and South Australian source signature evidenced from 87Sr/86Sr values alone (n = 6; Figure S5 in Supporting Information S1). The southeastern Australian dust signature is consistent with aerosol monitoring in New Zealand that showed PM10 concentrations were significantly higher than PM2.5 and the composition was that of crustal mineral dust (Davy & Trompetter, 2022). While both the Australian bushfires and desert dust storms contributed to large amounts of aerosol into the atmosphere, the relative difference between the coarse mineral dust and the finer bushfire aerosol impact in New Zealand can be explained by the preferential gravitational settling of the heavier mineral dust out of the atmosphere as the airmass passed over the country, while the finer particles remained suspended and were transported further east over the Pacific Southern Ocean (Tang et al., 2021; Weis et al., 2022). Taken together, this indicates that the primary driver of discoloration to the Southern Alps was coarse mineral dust sourced from a southeastern Australian desert dust storm.
The 2021 samples (0.712616 < 87Sr/86Sr < 0.718225 and −8.6 < εNd(0) < −6.6) fall along a two component mixing line between background dust and the dust event end members (Figure 2c) indicating an additional input of locally derived dust over the year that the dust event was either exposed at the surface or close enough to the surface that it was able to accumulate further deposited particulates on emergence with summer snowmelt.
The historical 1939 Black Friday dust composition (87Sr/86Sr = 0.715717) and trace element ratios (Figure S6 in Supporting Information S1) are consistent with the 2019/2020 dust event confirming an Australian source and indicating a repeated pattern of dust transport from the same geographic region. In addition, the rare earth element signature of historical Australian dust storms deposited on New Zealand glaciers, including in December 1987 (Marx et al., 2005b) and February 2000 (McGowan et al., 2005), corroborate with a southeastern Australian source. Thus, over the past 85 years, the Southern Alps episodically received southeastern Australian dust with a consistent geochemical composition. Understanding the impacts of these dust events on New Zealand glaciers is therefore critical as Australian bushfires and dust storms are increasing in frequency (e.g. Dutta et al., 2016).
4 Dust Deposition to Tasman, Fox, Franz Josef and Brewster Glaciers
Both personal and time lapse camera observations indicate that the majority of the dust event was deposited to the Southern Alps in a single event on 29 November 2019 with negligible input during the 2020 New Year event (Text S2 and Figure S1 in Supporting Information S1). While the media reported the impact on New Zealand as originating from the Australian bushfires, our observations strongly suggest that the source was Australian desert dust. A major dust storm was emitted from southeast Australia around 22 November 2019 (Li et al., 2021) and inundated Broken Hill on 29 November 2019 (Mabin, 2019). Conversely, the 2020 New Year event was dominated by biomass burning aerosol over New Zealand and remote sensing observed a particularly strong pyrocumulonimbus storm producing a gigantic smoke layer over the Tasman Sea on New Year's Eve 2019 (Zhang et al., 2021). Coincidentally, melting of snow overlying the dust, revealing its extensive distribution, also took place during this time (Figure S1 in Supporting Information S1) in accord with a reduction in albedo (Pu et al., 2021), prompting an erroneous conclusion of the timing of its deposition.
Similar meteorological conditions during the dust storm and fire event produced hot northwest winds from the desert that fanned the flames and pushed dust and smoke across the Tasman Sea toward New Zealand (Text S3, and Figures S8–S9 in Supporting Information S1).
Dust mass concentrations from the dust event ranged between 3 and 26 g m−2 and are higher than dust concentrations in a background year, which ranged between 1 and 5 g m−2 (Table S1 and Figure S10 in Supporting Information S1). The wider range of dust concentrations from the dust event compared to the background dust reflects the patchy dust accumulation on the glacier surface and regions of snow melt and bare ice which allow the dust to percolate through the snowpack as observed in the time lapse images and drone map (Figure 1b and Figure S1 in Supporting Information S1). The range of dust mass concentrations is also influenced by the particle size distribution, for example, large local dust particles increase the dust mass significantly as observed at Brewster Glacier. Higher dust mass concentrations during the dust event on the west coast glaciers of Franz Josef and Fox reflects their west facing direction further upwind toward Australia compared to the accumulation zone on the Tasman Glacier that is also west facing but downwind of the Main Divide, a significant mountain ridge (Kerr et al., 2018).
5 Distribution and Quantification of Dust Input to Southern Alps
Total dust mass deposited over permanent snow and ice on the Southern Alps is estimated to be 4,500 ± 500 Mg (metric tons) with a mean dust concentration of 6.5 ± 0.7 g m−2. Total mass deposited on each of the Fox, Franz Josef, and Tasman glaciers is estimated to be 160 ± 20 Mg, 190 ± 20 Mg, and 150 ± 20 Mg respectively (total ∼490 Mg). Figure 3 shows that dust appears to accumulate around the upper snowfields that feed into the lower glaciers, with minimal concentrations estimated for the glacier tongues (validated by visual inspection of available satellite imagery). Estimates of dust mass deposition represent minimum estimates due to removal of dust on bare ice by meltwater which cannot be accounted for, and continued burial of dust by snow at the highest altitudes that never melted to reveal the dust. The dust mass reported is best regarded as an estimate of dust that was visibly persistent at the surface, which would affect snow albedo. Overall, distribution is more even over the Fox and Franz Josef glaciers than the Tasman Glacier which features larger areas of minimal concentrations as well as patches of much higher concentrations than the other two. This is backed by mean (99th percentile) estimated dust concentrations of 5.1 g m−2 (12.7 g m−2) for Fox Glacier, 5.7 gm−2 (14.8 g m−2) for Franz Josef Glacier, and 5.7 g m−2 (17.1 g m−2) for Tasman Glacier. While dust events from Australia are episodically observed in New Zealand, the mass deposition of Australian dust to New Zealand glaciers has remained unquantified until now, making comparison of the 2019/2020 event to prior events impossible. The event falls within the range of the extreme densest mineral dust deposition events to glaciers observed globally (e.g., Bolaño-Ortiz et al. (2023); Di Mauro et al. (2015); Dumont et al. (2020). The origin and deposition of such a large quantity of terrestrial material from Australia to New Zealand also supports the rare, punctuated dispersal model proposed to explain the biogeography of many New Zealand plant species (e.g. McGlone, 2016) as well as very likely transportation of numerous microbes.

Dust mass loadings over permanent snow and ice near Aoraki (Mt Cook) and the entire extent of permanent snow and ice in the Southern Alps (insert). Blue outlines indicate the Fox, Franz Josef, and upper Tasman glaciers (in order, from left). Map base layer is a “hill shade” of a 15 m elevation model. Map projection is New Zealand Transverse Mercator (NZTM, EPSG:2193), grid line projection is WGS84 latitude/longitude.
It is difficult to reliably estimate the timing of deposition from the satellite record due to problematic levels of cloud cover which often occur during snowfall events. However, some of the heaviest estimated concentrations on the Tasman Glacier were detected on 29 November 2019. This matches with a visual assessment of the imagery from that date which shows two thick, spatially homogeneous, red patches on the lower Tasman. Unfortunately, the Fox and Franz Josef glaciers were covered in cloud on that date, and the highest dust concentration estimates for those come from an image taken on 13 March 2020. Nevertheless, satellite imagery, time lapse observations, and personal observations support a 29 November 2019 dust deposition event.
6 Conclusions
Climate conditions favorable for generating severe bushfires and desert dust storms are leading to more frequent and intense events (Abatzoglou et al., 2019; Bowman et al., 2020). Light absorbing impurities, such as mineral dust and black carbon, mobilized from such events and transported to and deposited on glacial environments, reduce albedo relative to background impurity levels, impacting the surface energy balance and summer snow melt regime. The 2019/2020 Australian bushfires and dust storm resulted in one of the most extreme mass deposition and impurity discoloration events in the Southern Alps of New Zealand; the Southern Alps permanent snow and ice region (750 km2) was blanketed in red dust. At the time, global media reported that the glacier discoloration was produced from bushfire ash on New Year 2020.
Quantification of the source, composition, mass deposition and timing of impurity deposition on the Southern Alps is necessary to understand the impacts on albedo and glacial melt. Personal observations, time lapse camera and satellite imagery confirm that the bulk of the impurities were deposited around 29 November 2019 and persisted for 2 weeks when they were buried by fresh snowfall. Patches of impurities reemerged after fresh snow disappeared and were still visible at the end of February 2020 and again in February 2023 with long lasting and multi-year impacts on albedo. We traced the impurities in snow samples collected from the Tasman, Fox and Franz Josef glaciers back to southeastern Australia using Nd and Sr isotope ratios and airmass back trajectories. The impurities were dominated by mineral dust which was five-fold above background levels. Dust mass concentrations in the Tasman, Fox and Franz Josef glaciers ranged between 3 and 26 g m−2. Using sample mass concentrations and remote sensing, we estimated the total dust deposition mass and mapped the spatial distribution across the Southern Alps.
Contrary to media reports, the glacier discoloration was caused by a desert dust storm and specific meteorological conditions that transported approximately 4,500 ± 500 tons of southeastern Australian dust across the Tasman Sea to permanent snow and ice of the Southern Alps. Australian dust deposition to New Zealand glaciers has been documented over the past decades through provenance studies, however mass deposition data for New Zealand glaciers is scarce and this study provides an important baseline to which future deposition events can be compared. Future work will assess the impact of the impurities on glacier albedo, snow melt and snow biology. Glacier mass has declined over the past few decades in New Zealand and climate conditions conducive to large Australian bushfires and dust storm events are predicted to strengthen with compounding consequences on New Zealand glacial environments.
Acknowledgments
This project was supported by the Marsden Fund Council from New Zealand Government funding, managed by Royal Society Te Apārangi (MFP-LCR2003 and MFP-UOC1804). We thank Te Rūnanga o Arowhenua, Te Rūnanga o Moeraki, Te Rūnanga o Waihao, Aoraki Environmental Consultancy Ltd, and Aukaha Ltd for cultural engagement, Heliservices and The Helicopter Line for logistical support, and J. King (Manaaki Whenua-Landcare Research) for the 1939 Golden Downs sample. T. King, K. Russell, J. Henry, M. McMillan (Te Rūnanga o Arowhenua) and R. Bellringer (Department of Conservation, Aoraki/Mt Cook) provided important field advice. Frans Gerber provided technical support for dust geochemistry. Paul Bealing flew the drone surveys and processed ortho-imagery. Samples were collected under Department of Conservation concession CA31615-OTH. We thank two anonymous reviewers for their constructive comments.
Open Research
Data Availability Statement
The dust mass concentration, trace element concentration and Sr-Nd isotopic data are available at Winton et al. (2024). A GeoTiff of dust concentration estimates and corresponding image date codes (per pixel), as well as GEOJSON vector files for the snow/ice extent and Fox, Franz Josef, and Tasman glaciers are available at Jolly (2024, https://datastore.landcareresearch.co.nz/dataset/au-dust-2019). Dust mass concentrations are also available via the Google Engine Drive (https://jollyb-landcare.projects.earthengine.app/view/red-dust-on-southern-alps). While this study used custom-processed Sentinel-2 imagery, an equivalent product widely available is the Sentinel-2 Level-2A atmospherically corrected surface reflectance (https://developers.google.com/earth-engine/datasets/catalog/COPERNICUS_S2_SR_HARMONIZED.