Martian Case Study of Multivariate Correlation and Regression with Planetary Datasets

We synthesize multivariate correlation and regression methods to characterize unique relationships among compositional and physical properties of a planetary surface locally, regionally, and globally. Martian data including elemental mass fractions, areal fractions of mineral types, and thermal inertia constitute our case study. We incorporate techniques to address the effects of spatial autocorrelation and heteroscedasticity. We also utilize method and fit diagnostics. While the Mars Odyssey and Mars Global Surveyor missions provide the exploratory context in our discussion, our approach is applicable whenever the interrelationships of spatially binned data of continuous-valued planetary attributes are sought. For example, our regional-scale case study reinforces the strength of the spatial correlation among K, Th, and the dominant mineralogic type within northern low albedo regions (surface type 2) of Mars. Recent chemical and mineralogic data from the MESSENGER mission at Mercury and Dawn at Vesta may be analyzed effectively with these hierarchical regression methods to constrain geochemical processes. Likewise, our algorithm could be applied locally with the wide variety of compositional data expected from the MSL mission at Gale Crater in general, and the ChemCam sampling grids in particular.     

Satellite-derived estimates of potential carbon sequestration through afforestation of agricultural lands in the United States

Afforestation of marginal agricultural lands represents a promising option for carbon sequestration in terrestrial ecosystems. An ecosystem carbon model was used to generate new national maps of annual net primary production (NPP), one each for continuous land covers of ‘forest’, ‘crop’, and ‘rangeland’ over the entire U. S. continental area. Direct inputs of satellite “greenness” data from the Advanced Very High Resolution Radiometer (AVHRR) sensor into the NASA-CASA carbon model at 8-km spatial resolution were used to estimate spatial variability in monthly NPP and potential biomass accumulation rates in a uniquely detailed manner. The model predictions of regrowth forest production lead to a conservative national projection of 0.3 Pg C as potential carbon stored each year on relatively low-production crop or rangeland areas. On a regional level, the top five states for total crop afforestation potential were: Texas, Minnesota, Iowa, Illinois, and Missouri, whereas the top five states for total rangeland afforestation potential are: Texas, California, Montana, New Mexico, and Colorado. Afforestation at this level of intensity has the capacity to offset at least one-fifth of annual fossil fuel emission of carbon in the United States. These projected afforestation carbon gains also match or exceed recent estimates of the annual sink for atmospheric CO₂ in currently forested area of the country.     

Marine radiocarbon reservoir corrections (∆R) for Chesapeake Bay and the Middle Atlantic Coast of North America

Radiocarbon dates from known age, pre-bomb eastern oyster (Crassostrea virginica) shells provide local marine reservoir corrections (?R) for Chesapeake Bay and the Middle Atlantic coastal area of eastern North America. These data suggest subregional variability in ?R, ranging from 148 ± 46 ¹⁴C yr on the Potomac River to ? 109 ± 38 ¹⁴C yr at Swan Point, Maryland. The ?R weighted mean for the Chesapeake's Western Shore (129 ± 22 ¹⁴C yr) is substantially higher than the Eastern Shore (? 88 ± 23 ¹⁴C yr), with outer Atlantic Coast samples falling between these values (106 ± 46 and 2 ± 46 ¹⁴C yr). These differences may result from a combination of factors, including ¹⁴C-depleted freshwater that enters the bay from some if its drainages, ¹⁴C-depleted seawater that enters the bay at its mouth, and/or biological carbon recycling. We advocate using different subregional ?R corrections when calibrating ¹⁴C dates on aquatic specimens from the Chesapeake Bay and coastal Middle Atlantic region of North America.     

Assessing lithologic discontinuities and parent material uniformity within the Texas sandy mantle and implications for archaeological burial and preservation potential in upland settings

Alfisols within the Texas Gulf Coast Plain commonly exhibit textural contrasts between sandy, artifact-bearing A–E horizons (i.e., sandy mantle), and artifact-sterile clay-rich Bt (argillic) horizons. This has invoked debate about parent material uniformity and pedogenic versus geomorphic sandy mantle origins, which has implications for the scientific value of buried archaeological sites. To improve our understanding of archaeological burial in upland settings, we evaluated parent material uniformity within five pedons to distinguish pedogenically derived textural changes from geomorphologically created lithologic discontinuities. Depth trends in clay-free particle size classes and stable/immobile Ti and Zr constituents failed to reveal lithologic discontinuities between the sandy mantle and Bt horizons, and the observed textural differences are interpreted to have resulted from pedogenic processes. This interpretation is supported by clay skins, fine clay increases in Bt horizons, and micromorphological observations. Consequently, artifacts buried in upland summits have likely moved down the soil profile due to biomantle processes. Deep sandy mantle sites, non-parallel contacts between the sandy mantle and Bt horizons, and paleogullies incised into Eocene bedrock are better explained by colluvial/soil creep processes adjacent to summits, where archaeological materials may exhibit preservation potential. No single explanation can account for sandy mantle origins, and we advocate a case-by-case approach.     

A threshold in soil formation at Earth’s arid-hyperarid transition

The soils of the Atacama Desert in northern Chile have long been known to contain large quantities of unusual salts, yet the processes that form these soils are not yet fully understood. We examined the morphology and geochemistry of soils on post-Miocene fans and stream terraces along a south-to-north (27° to 24° S) rainfall transect that spans the arid to hyperarid transition (21 to ∼2 mm rain y⁻¹). Landform ages are ? 2 My based on cosmogenic radionuclide concentrations in surface boulders, and Ar isotopes in interbedded volcanic ash deposits near the driest site indicate a maximum age of 2.1 My. A chemical mass balance analysis that explicitly accounts for atmospheric additions was used to quantify net changes in mass and volume as a function of rainfall. In the arid (21 mm rain y⁻¹) soil, total mass loss to weathering of silicate alluvium and dust (−1030 kg m⁻²) is offset by net addition of salts (+170 kg m⁻²). The most hyperarid soil has accumulated 830 kg m⁻² of atmospheric salts (including 260 kg sulfate m⁻² and 90 kg chloride m⁻²), resulting in unusually high volumetric expansion (120%) for a soil of this age. The composition of both airborne particles and atmospheric deposition in passive traps indicates that the geochemistry of the driest soil reflects accumulated atmospheric influxes coupled with limited in-soil chemical transformation and loss. Long-term rates of atmospheric solute addition were derived from the ion inventories in the driest soil, divided by the landform age, and compared to measured contemporary rates. With decreasing rainfall, the soil salt inventories increase, and the retained salts are both more soluble and present at shallower depths. All soils generally exhibit vertical variation in their chemistry, suggesting slow and stochastic downward water movement, and greater climate variability over the past 2 My than is reflected in recent (∼100 y) rainfall averages. The geochemistry of these soils shows that the transition from arid to hyperarid rainfall levels marks a fundamental geochemical threshold: in wetter soils, the rate and character of chemical weathering results in net mass loss and associated volumetric collapse after 10⁵ to 10⁶ years, while continuous accumulation of atmospheric solutes in hyperarid soils over similar timescales results in dramatic volumetric expansion. The specific geochemistry of hyperarid soils is a function of atmospheric sources, and is expected to vary accordingly at other hyperarid sites. This work identifies key processes in hyperarid soil formation that are likely to be independent of location, and suggests that analogous processes may occur on Mars.     

The golden ark: arsenopyrite crystal plasticity and the retention of gold through high strain and metamorphism

Quantitative electron backscatter diffraction analysis and ion microprobe imaging of gold‐rich arsenopyrites provide the first insights into the crystal plasticity and element mobility behaviour of arsenopyrites through metamorphism (340°–460° and 2 kbar). Remarkably, the gold‐rich arsenopyrites remained structurally and chemically robust during high strain deformation. It was only during a superimposed lower strain deformation event, at a high angle to the preferred orientation of the arsenopyrites, that small amounts of crystal plasticity affected the arsenopyrites. During the low strain event, a dissolution–reprecipitation reaction resulted in loss of gold from the crystal lattice, facilitated by localised domains of recrystallisation, most likely due to fluid percolation along sub‐ and new grain boundaries. We suggest that the abundance and rheologically robust nature of gold‐rich arsenopyrite in giant gold deposits, affected by greenschist–amphibolite metamorphism, is actually critical in the preservation of those deposits.     

Environment and paleoecology of a 12 ka mid-North American Younger Dryas forest chronicled in tree rings

Until now, availability of wood from the Younger Dryas abrupt cooling event (YDE) in N. America ca. 12.9 to 11.6 ka has been insufficient to develop high-resolution chronologies for refining our understanding of YDE conditions. Here we present a multi-proxy tree-ring chronology (ring widths, “events” evidenced by microanatomy and macro features, stable isotopes) from a buried black spruce forest in the Great Lakes area (Liverpool East site), spanning 116 yr at ca. 12,000 cal yr BP. During this largely cold and wet period, the proxies convey a coherent and precise forest history including frost events, tilting, drowning and burial in estuarine sands as the Laurentide Ice Sheet deteriorated. In the middle of the period, a short mild interval appears to have launched the final and largest episode of tree recruitment. Ultimately the tops of the trees were sheared off after death, perhaps by wind-driven ice floes, culminating an interval of rising water and sediment deposition around the base of the trees. Although relative influences of the continental ice sheet and local effects from ancestral Lake Michigan are indeterminate, the tree-ring proxies provide important insight into environment and ecology of a N. American YDE boreal forest stand.     

Effect of contaminant concentration on in situ bacterial sulfate reduction and methanogenesis in phenol-contaminated groundwater

The availability of dissolved O₂ can limit biodegradation of organic compounds in aquifers. Where O₂ is depleted, biodegradation proceeds via anaerobic processes, including NO₃-, Mn(IV)-, Fe(III)- and SO₄-reduction and fermentation/methanogenesis. The environmental controls on these anaerobic processes must be understood to support implementation of management strategies such as monitored natural attenuation (MNA). In this study stable isotope analysis is used to show that the relative significance of two key anaerobic biodegradation processes (bacterial SO₄ reduction (BSR) and methanogenesis) in a phenol-contaminated sandstone aquifer is sensitive to spatial and temporal changes in total dissolved phenols concentration (TPC) (= phenol + cresols + dimethylphenols) over a 5-a period. In general, ³⁴SO₄-enrichment (characteristic of bacterial SO₄ reduction) is restricted spatially to locations where TPC < 2000 mg L⁻¹. In contrast, ¹³C-depleted CH₄ and ¹³C-enriched CO₂ isotope compositions (characteristic of methanogenesis) were measured at TPC up to 8000 mg L⁻¹. This is consistent with previous studies that demonstrate suppression of BSR at TPC of >500 mg L⁻¹, and suggests that methanogenic microorganisms may have a higher tolerance for TPC in this contaminant plume. It is concluded that isotopic enrichment trends can be used to identify conditions under which in situ biodegradation may be limited by the properties of the biodegradation substrate (in this case TPC). Such data may be used to deduce the performance of MNA for contaminated groundwater in similar settings.     

Numerical modeling of the June 24, 2015, Hongyanzi landslide generated impulse waves in Three Gorges Reservoir,China

Subaerial landslides falling into confined water bodies often generate impulsive waves. Damaging landslide tsunamis in Three Gorges Reservoir, China, have struck several times in the last 15 years. On June 24, 2015, a 23?×?10⁴ m³ slope failure occurred on the east bank of the Daning River opposite Wushan Town. The sliding mass intruded into the Three Gorges Reservoir and initiated a reservoir tsunami that resulted in two deaths and significant damage to shipping facilities. A post-event survey revealed the landslide geometry and wave run-up distribution, while an eyewitness video captured most of the landslide motion. Employing these firm constraints, we applied the Tsunami Squares method to simulate the 2015 Hongyanzi landslide and tsunami. The simulation revealed that the landslide experienced a progressive failure in the first few seconds and impacted the water with a maximum velocity of ~?16 m/s. The initial wave propagated to the opposite shore in an arch shape, and the water surface reached a maximum amplitude of ~?11 m near the landslide. Wave amplitude-time curves at four points on the river cross section show that the initial wave reached Wushan town in about 50 s with an average wave velocity of ~?30 m/s. The maximum wave run-ups on the shoreline opposite the landslide are around 6 m and attenuate to less than 1 m beyond 2-km distance. The landslide simulation matches the observed geological profile and the eyewitness video, and the numerical results coincide with the observed wave run-up heights. Nearly 80% of landslide energy is lost due to frictional resistances, but the remaining fraction imparted to the tsunami carried catastrophic consequences to a large region. The numerical results emphasize the efficiency and accuracy of Tsunami Squares method for a “Quick Look” simulation of a potential landslide.     

Grain growth in porous olivine aggregates

Grain growth experiments have been performed at 1 atm on fine grain size (<10 μm) synthetic olivine (Fo₉₁) aggregates at various temperatures (1200° to 1400° C), oxygen fugacities (10^-4 to 10^-11 atm) and total anneal times (10, 30, 60, 100 and 200 h). The rate of grain growth increased with increasing temperature and with increasing oxygen fugacity. The presence of a second phase (residual porosity), introduced during sample fabrication, has a significant effect on grain growth, with evolution in grain size paralleled by changes in the size and frequency of the pores. When the grain growth data were fit to a growth law G ⁿ ?G _O ⁿ =κ₀ tf ₀ ^m ₂e^?Q/RT, the growth exponents fall in the range of n=4 to 5, suggesting that grain growth may be controlled by the coalescence of the second phase. The evolution in pore size and frequency may occur either by the transport of the ionic species constituting olivine between the pores or by the movement of the pores themselves along the grain boundaries and edges. Thus, the rate of growth of the pores and grains is probably limited by diffusion of the slowest ionic species constituting olivine (magnesium, iron, silicon, or oxygen) moving along the fastest path for that species (through the lattice, along the grain boundaries, around the surface of the moving pores, or through the vapor phase in the pores). Activation energies for grain growth of Q=290 ± 20 kJ/mol and 345 ± 25 kJ/mol were calculated from our results for n=4 and 5, respectively. These activation energies preclude vapor-phase transport and iron diffusion along grain boundaries but do not otherwise permit a discrimination between the rate limiting species or path. The oxygen fugacity exponent of m ≈0.12 suggests that lattice diffusion does not control the grain growth. However, the lack of data for magnesium, iron, silicon and oxygen surface and grain boundary diffusion in olivine makes definitive determination of the mechanism controlling grain growth difficult.