Large influence of dust on the Precambrian climate

The earliest land plants likely appeared between the late Cryogenian and middle Cambrian, implying that for most of Earth’s history land surfaces lacked vegetation. Without vegetation, dust emission from land would have been more widespread and intense, increasing atmospheric dust loading with potentially significant climatic effects and enhanced nutrient delivery to the oceans. Despite these implications, the dynamics of the dust cycle in the Precambrian have not been rigorously investigated, and most previous Precambrian climate modeling neglected dust. This study poses the question: how large was the dust cycle before land vegetation, and what were its climatic impacts? Using an Earth system model with a Precambrian continental configuration, the authors aim to quantify dust emissions, atmospheric loading, deposition (including to the oceans), and the resultant climate effects under varying surface erodibilities.

Dust affects climate directly through absorption and scattering of solar radiation and indirectly through cloud microphysical changes. Present-day estimates suggest dust reduces surface solar radiation by about 1–3 W m−2 due to direct effects, with indirect effects around −2 W m−2, and modern global emissions of roughly 1000–3000 Tg yr−1. Past changes in dust have influenced large-scale climate features (e.g., African monsoon, ENSO) and SSTs. Precambrian climate studies have largely omitted dust, except in contexts of snowball Earth deglaciation where dust was considered to have a warming effect by lowering surface or planetary albedo over ice. Martian climate studies underscore the importance of dust in dust-rich atmospheres. This work builds on and contrasts with prior studies by focusing on dust in an ice-free or moderately glaciated Precambrian climate and quantifying its global cooling effect and biogeochemical implications.

Model: CESM 1.2.2 was used with atmosphere (CAM4, finite-volume dynamical core), land (CLM4), sea ice (CICE4), and ocean (POP2). Horizontal resolution: atmosphere/land 1.9°×2.5°; ocean/ice ~1°. Vertical levels: atmosphere 26, ocean 60. CAM4 includes a dust scheme based on the DEAD (Dust Entrainment And Deposition) parameterization with emissions dependent on wind friction velocity, soil moisture, and vegetation/snow cover. Dust is represented in four size bins (0.1–1.0, 1.0–2.5, 2.5–5.0, 5.0–10 μm). CAM4 considers only shortwave radiative effects of dust; longwave effects and indirect cloud-aerosol interactions are not represented. CLM4 and CICE4 simulate albedo darkening by impurities in snow/ice. CLM4’s dynamic vegetation scheme (CNDV) can be enabled and includes C and N cycles and bioclimatic limits for unmanaged PFTs. Boundary conditions and configuration: A 720 Ma continental configuration (Rodinia) was used with simplified low-relief topography (mean ~400 m, sloping toward coasts) and uniform ocean depth of 4000 m plus an idealized mid-ocean ridge. Solar constant set to 94% of present (S0≈1367 W m−2). Orbital parameters as year 1990. Greenhouse gases: CO2=2000 ppmv to avoid snowball conditions; CH4=805.6 ppbv and N2O=276.7 ppbv (pre-industrial values). Other aerosols (black carbon, sulfate, organic carbon) omitted. Experimental design: A control run with uniform erodibility k=0.15 (global mean modern value) was integrated to equilibrium (≈2300 years). Branch experiments altered uniform k to 0.0 (no dust), 0.0375, 0.075, 0.15, and 0.3, each integrated ≥400 years to achieve top-of-atmosphere net flux within ±0.3 W m−2. After statistical equilibrium, an additional branch with dynamic vegetation (CNDV) was run for each nonzero k for 700 years to equilibrate vegetation and climate (TOA net flux again within ±0.3 W m−2). Analyses use the last 100 years. Diagnostics: The direct shortwave radiative effect of dust (DDSE) was diagnosed by double-calling radiation each timestep—once with and once without dust optical properties—and taking differences, over 5-year diagnostic segments. Longwave and sensible heat flux anomalies attributable to dust were estimated by paired 1-year runs with and without dust and differencing fields, acknowledging some drift and reduced precision. Additional diagnostics include partitioning of dry vs wet deposition, aerosol size-bin contributions, and spatial fields such as cloud albedo, surface albedo changes due to dust deposition, and atmospheric heating rates.

• Dust emissions and loading: In no-vegetation cases, even the lowest erodibility (k=0.0375) yields total dust emissions and atmospheric loading more than 15× and 10× present-day, respectively. Global emissions scale nearly linearly with k: 36,463 Tg yr−1 at k=0.0375 to 132,303 Tg yr−1 at k=0.3 (≈1.53× increase per doubling of k). Atmospheric dust loading increases from 312 Tg to 1114 Tg over the same range. The fraction of fine particles (0–2.5 μm) in atmospheric loading rises from 45.8% (k=0.0375) to 56.5% (k=0.3).
• Deposition: About one-third of emitted dust is deposited into the oceans in all no-vegetation experiments, representing ~10–35× modern ocean deposition. Wet deposition fraction declines slightly with increasing k (from 32.5% at k=0.0375 to 30.0% at k=0.3) due to reduced precipitation. Coastal depositional rates can correspond to sedimentation of ~10 mm kyr−1 at k=0.0375, sufficient to form thick tabular sediment beddings over million-year timescales.
• Radiative and thermal impacts: Relative to no dust (k=0), adding dust with k=0.0375 reduces global-mean surface downwelling shortwave by 22.5 W m−2 and cools the surface by 11.2 °C. Increasing k to 0.3 adds another ~9 °C of cooling (~3 °C per doubling of k). Despite increased atmospheric heating by dust (shortwave absorption), the surface cools; temperature profiles become more stable, deviating from a moist adiabat, and a near-surface inversion forms at k=0.3.
• Feedbacks mitigating surface cooling: Dust-induced atmospheric stabilization reduces cloud albedo (net warming), decreases upward sensible heat loss (increasing downward sensible component), and increases downward longwave radiation to the surface. Planetary albedo increases only slightly (from 0.383 at k=0.0375 to 0.405 at k=0.3) despite large dust loadings. Dust deposition reduces sea-ice albedo by up to ~3% at k=0.3 when climates are sufficiently cold with extensive sea ice and low snowfall.
• Hydroclimate and cryosphere: Global-mean precipitation declines from 2.9 mm day−1 (k=0) to 1.0 mm day−1 (k=0.3). Sea-ice extent expands from around 55° latitude (no dust) toward the tropics at k=0.3.
• Vegetation effects: Enabling dynamic vegetation warms the global mean surface by >6.5 °C in all cases relative to no-vegetation counterparts, implying that the emergence of land vegetation would have substantially suppressed dust emissions and warmed climate under fixed external forcing.

The simulations demonstrate that in the absence of land vegetation, Precambrian Earth likely experienced dust emissions and atmospheric loadings an order of magnitude larger than modern values, leading to strong surface cooling (~10 °C) in relatively warm background climates. These findings address the research question by quantifying a large, previously underappreciated dust-climate coupling for deep time. The results imply that Neoproterozoic climates (when not in snowball states) were probably much colder than prior simulations that neglected dust. The strong dust flux to the oceans suggests robust nutrient delivery with potential implications for marine productivity and biogeochemical cycles. Generality: Although the continental configuration aligns with late Neoproterozoic reconstructions, the authors argue that large dust loadings should be a common feature before land vegetation across different Precambrian configurations, especially given evidence that continental area since ~2.5 Ga may have been comparable to modern. The role of dust in initiating or sustaining snowball states remains uncertain: dust can cool via shortwave attenuation but may warm near-snowball conditions by lowering surface and planetary albedos over ice—a competition requiring further study. Process insights: Dust stabilizes the atmosphere, reduces cloud albedo, decreases sensible heat loss, and enhances downward longwave, partially offsetting direct shortwave surface cooling yet not enough to prevent substantial net cooling. Increased fine-mode fraction with higher k likely reflects reduced deep convection and weaker scavenging aloft. Enhanced oceanic deposition rates are consistent with observed tabular sediment beddings and may have influenced life evolution through iron fertilization. Broader implications: Post-snowball intervals may have been especially dusty due to glacial grinding, potentially stimulating primary production and organic carbon burial, contributing to oxygenation events. Compared to glacial-interglacial changes (e.g., LGM dust ~2× modern), Precambrian dust amplification is far larger, underscoring dust as a first-order climate component in deep time.

Model evidence indicates that prior to the rise of land vegetation, Earth’s dust emissions and atmospheric loading were likely ~10× modern values. This enhanced dust cycle would have: (1) cooled the global surface climate by roughly 10 °C under warm conditions; (2) substantially increased dust deposition to oceans, supplying nutrients and potentially affecting marine ecosystems and atmospheric oxygenation via enhanced organic carbon burial; and (3) contributed to widespread tabular sediment bedding. Therefore, dust must be explicitly considered in reconstructions and simulations of early Earth climate, atmospheric chemistry, and biogeochemical cycles. Future research should include improved constraints on surface erodibility and source strength through time, explicit representation of biological/physical soil crust dynamics, assessment of dust longwave and indirect cloud effects, and exploration of dust feedbacks in near-snowball and deglaciating climates.

• Uncertain surface erodibility (k): the largest source of uncertainty; spatial distribution for the Precambrian is unknown and was assumed uniform with a range of values. Despite this, even much smaller k values imply several degrees of global cooling.
• Missing physics: CAM4 neglects dust longwave radiative effects and aerosol-cloud indirect effects, which could alter quantitative results locally or regionally. Longwave and sensible heat diagnostics via paired short runs are less precise due to climate drift.
• Vegetation biases and spin-up: The dynamic vegetation scheme underestimates tundra and overestimates tree cover, and the simulations may be too short for full vegetation equilibrium, likely underestimating vegetation’s dust-suppressing and warming impacts.
• Continental configuration and topography: Uses an idealized 720 Ma configuration with simplified low-relief continents and uniform ocean depth; generalization to all Precambrian times assumes similar continent-scale characteristics.
• Soil crust processes: Potential biogenic/physical soil crusts that could reduce dust emissions are not explicitly modeled; their distribution in the Precambrian is unknown.
• Other aerosols and chemistry: Non-dust aerosols were omitted; potential chemical impacts of dust (e.g., on atmospheric composition) are not explicitly simulated.