Search

Organic carbon (OC) association with soil minerals stabilizes OC on timescales reflecting the strength of mineral-C interactions. We applied ramped thermal oxidation to subsoil B horizons with different mineral-C associations to separate OC according to increasing temperature of oxidation, i.e. thermal activation energy. Generally, OC released at lower temperatures was richer in bioavailable forms like polysaccharides, while OC released at higher temperatures was more aromatic. Organic carbon associated with pedogenic oxides was released at lower temperatures and had a narrow range of 14C content. By contrast, N-rich compounds were released at higher temperatures from samples with 2 : 1 clays and short-range ordered (SRO) amorphous minerals. Temperatures of release overlapped for SRO minerals and crystalline oxides, although the mean age of OC released was older for the SRO. In soils with more mixed mineralogy, the added presence of older OC released at temperatures greater than 450°C from clays resulted in a broader distribution of OC ages within the sample, especially for soils rich in 2 : 1 layer expandable clays such as smectite. While pedogenic setting affects mineral stability and absolute OC age, mineralogy controls the structure of OC age distribution within a sample, which may provide insight into model structures and OC dynamics under changing conditions. This article is part of the Theo Murphy meeting issue Radiocarbon in the Anthropocene.

Sequential chemical extraction has been widely used to study soil phosphorus (P) dynamics and inform nutrient management, but its efficacy for assigning P into biologically meaningful pools remains unknown. Here, we evaluated the accuracy of the modified Hedley extraction scheme using P K-edge X-ray absorption near-edge structure (XANES) spectroscopy for nine carbonate-free soil samples with diverse chemical and mineralogical properties resulting from different degrees of soil development. For most samples, the extraction markedly overestimated the pool size of calcium-bound P (Ca-P, extracted by 1 M HCl) due to (1) P redistribution during the alkaline extractions (0.5 M NaHCO3 and then 0.1 M NaOH), creating new Ca-P via formation of Ca phosphates between NaOH-desorbed phosphate and exchangeable Ca2+ and/or (2) dissolution of poorly crystalline Fe and Al oxides by 1 M HCl, releasing P occluded by these oxides into solution. The first mechanism may occur in soils rich in well-crystallized minerals and exchangeable Ca2+ regardless of the presence or absence of CaCO3, whereas the second mechanism likely operates in soils rich in poorly crystalline Fe and Al minerals. The overestimation of Ca-P simultaneously caused underestimation of the pools extracted by the alkaline solutions. Our findings identify key edaphic parameters that remarkably influenced the extractions, which will strengthen our understanding of soil P dynamics using this widely accepted procedure.

Pedogenic thresholds describe where soil properties or processes change in an abrupt/nonlinear fashion in response to small changes in environmental forcing. Contrastingly, soil process domains refer to the space between thresholds where soil properties are either unchanged, or change gradually, across a broad range of environmental forcing. Here, we test quantitatively for the presence of thresholds in patterns of soil properties across a climatic gradient on soils developed from ~20 ky old basaltic substrate on the Island of Hawai'i. From multiple soil properties, we quantitatively identified a threshold at ~750 mm/y of water balance (precipitation minus potential evapotranspiration), delineating the upper water balance boundary of soil fertility in these soils. From the threshold in the ratio of exchangeable Ca to total Ca we identified the lower water balance boundary of soil fertility in these soils at -1000 mm/y, however this threshold was qualitatively described as it lies near the limit of the climate gradient data where the statistical approach can not be applied. These two results represent the first time that pedogenic thresholds have been identified using statistically rigorous methods and the limitations of said methods, respectively. Comparing the 20 ky soils to soils that developed on basaltic substrates of 1.2 ky, 7.5 ky, 150 ky, and 4100 ky in a time-climate matrix, we found that our quantitative analysis supports previous qualitatively identified thresholds in the soils developed from older substrates. We also identified the 20 ky as the transition from kinetic to supply limitation for plant nutrients in soil in this system.

Current understanding of phosphorus (P) cycling in soils can be enhanced by integrating previously discrete findings concerning P speciation, exchange kinetics, and the underlying biological and geochemical processes. Here, we combine sequential extraction with P K-edge X-ray absorption spectroscopy and isotopic methods (33P and 18O in phosphate) to characterize P cycling on a climatic gradient in Hawaii. We link P pools to P species and estimate the turnover times for commonly considered P pools. Dissolved P turned over in seconds, resin-extractable P in minutes, NaOH-extractable inorganic P in weeks to months, and HCl-extractable P in years to millennia. Furthermore, we show that in arid-zone soils, some primary mineral P remains even after 150 ky of soil development, whereas in humid-zone soils of the same age, all P in all pools has been biologically cycled. The integrative information we provide makes possible a more dynamic, process-oriented conceptual model of P cycling in soils.

The supply of nitrogen (N) constrains primary productivity in many ecosystems, raising the question "what controls the availability and cycling of N"? As a step toward answering this question, we evaluated N cycling processes and aspects of their regulation on a climate gradient on Kohala Volcano, Hawaii, USA. The gradient extends from sites receiving 3,000 mm/yr, and the pedology and dynamics of rock-derived nutrients in soils on the gradient are well understood. In particular, there is a soil process domain at intermediate rainfall within which ongoing weathering and biological uplift have enriched total and available pools of rock-derived nutrients substantially; sites at higher rainfall than this domain are acid and infertile as a consequence of depletion of rock-derived nutrients, while sites at lower rainfall are unproductive and subject to wind erosion. We found elevated rates of potential net N mineralization in the domain where rock-derived nutrients are enriched. Higher-rainfall sites have low rates of potential net N mineralization and high rates of microbial N immobilization, despite relatively high rates of gross N mineralization. Lower-rainfall sites have moderately low potential net N mineralization, relatively low rates of gross N mineralization, and rates of microbial N immobilization sufficient to sequester almost all the mineral N produced. Bulk soil δ15 N also varied along the gradient, from +4‰ at high rainfall sites to +14‰ at low rainfall sites, indicating differences in the sources and dynamics of soil N. Our analysis shows that there is a strong association between N cycling and soil process domains that are defined using soil characteristics independent of N along this gradient, and that short-term controls of N cycling can be understood in terms of the supply of and demand for N.

The supply of nitrogen (N) constrains primary productivity in many ecosystems, raising the question "what controls the availability and cycling of N"? As a step toward answering this question, we evaluated N cycling processes and aspects of their regulation on a climate gradient on Kohala Volcano, Hawaii, USA. The gradient extends from sites receiving 3,000 mm/yr, and the pedology and dynamics of rock-derived nutrients in soils on the gradient are well understood. In particular, there is a soil process domain at intermediate rainfall within which ongoing weathering and biological uplift have enriched total and available pools of rock-derived nutrients substantially; sites at higher rainfall than this domain are acid and infertile as a consequence of depletion of rock-derived nutrients, while sites at lower rainfall are unproductive and subject to wind erosion. We found elevated rates of potential net N mineralization in the domain where rock-derived nutrients are enriched. Higher-rainfall sites have low rates of potential net N mineralization and high rates of microbial N immobilization, despite relatively high rates of gross N mineralization. Lower-rainfall sites have moderately low potential net N mineralization, relatively low rates of gross N mineralization, and rates of microbial N immobilization sufficient to sequester almost all the mineral N produced. Bulk soil δ15 N also varied along the gradient, from +4‰ at high rainfall sites to +14‰ at low rainfall sites, indicating differences in the sources and dynamics of soil N. Our analysis shows that there is a strong association between N cycling and soil process domains that are defined using soil characteristics independent of N along this gradient, and that short-term controls of N cycling can be understood in terms of the supply of and demand for N.

SUMMARYAndean grasslands (páramos) are highly valued for their role in regional water supply as well as for their biodiversity and large soil carbon stocks. Several Payment for Ecosystem Services (PES) programmes promote either afforestation or alteration of traditional burning regimes under the assumption that these land management strategies will maximize páramo ecosystem services, including carbon storage. However, knowledge of the effects of incentivized land uses is limited. In an evaluation of how afforestation and elimination of burning affect carbon storage at a site in southern Ecuador, we found the highest above-ground biomass carbon levels at afforested sites (99.3–122.0 t C ha−1), while grassland sites reached 23.9 t C ha−1after 45 years of burn exclusion. Soil carbon storage from 0–20 cm was high across all sites (172.8–201.9 t C ha−1), but was significantly lower with afforestation than with burn exclusion. These findings suggest that, although afforestation is generally favoured when carbon is the primary ecosystem service of interest, grasslands with infrequent burning have important potential as a land management strategy when both above-ground biomass and soil carbon are considered. These results are relevant to the development and adaptation of PES programmes focused on carbon as well as those focused on multiple ecosystem services.

Successful land restoration in impoverished rural environments may require adoption of new resource management strategies; however, feedbacks between local knowledge and introduced restoration technologies have rarely been articulated. We used interview scenarios to analyze the role of local knowledge in land restoration at a large-scale, long-term watershed rehabilitation and wet meadow restoration program in the highland Andes. Indigenous communities built over 30,000 check dams, terraces and infiltration ditches, and the density of erosion control structures and visible restoration varied greatly across participant communities. We developed a survey reaching across the highest restoration management intensity, lowest restoration management intensity, and non-project (control) communities. We interviewed 49 respondents using 14 scenarios based on photos depicting biophysical phenomena related to land degradation and restoration. The scenarios generated 5,828 statements that were coded into 964 distinct concepts. As expected, respondents that built more erosion control structures had more detailed knowledge of check dam construction and ecosystem development following physical interventions. More significantly, there was a shift in the conceptualization of and attitudes toward land degradation and restoration. Respondents who built more erosion control structures were more likely to: attribute wetland hydrology to groundwater recharge rather than myth constructs about seeps and springs; attribute land degradation to human rather than mythological causes; and have more proactive attitudes regarding land restoration. Evidence suggests that when addressing severe land degradation or restoring ecosystem processes not readily observable by indigenous people, such as groundwater flow and wetland recharge, restoration success will depend on combining local and scientific knowledge.

Volcanic ash soils retain the largest and most persistent soil carbon pools of any ecosystem. However, the mechanisms governing soil carbon accumulation and weathering during initial phases of ecosystem development are not well understood. We examined soil organic matter dynamics and soil development across a high-altitude (3,560-3,030 m) 20-kyr climate gradient on Mauna Kea in Hawaii. Four elevation sites were selected (~250-500 mm rainfall), which range from sparsely vegetated to sites that contain a mix of shrubs and grasses. At each site, two or three pits were dug and major diagnostic horizons down to bedrock (intact lava) were sampled. Soils were analyzed for particle size, organic C and N, soil pH, exchangeable cations, base saturation, NaF pH, phosphorous sorption, and major elements. Mass loss and pedogenic metal accumulation (hydroxlamine Fe, Al, and Si extractions) were used to measure extent of weathering, leaching, changes in soil mineralogy and carbon accumulation. Reactive-phase (SRO) minerals show a general trend of increasing abundance with increasing rainfall. However carbon accumulation patterns across the climate gradient are largely decoupled from these trends. The results suggest that after 20 kyr, pedogenic processes have altered the nature and composition of the volcanic ash such that it is capable of retaining soil C even where organic acid influences from plant material and leaching from rainfall are severely limited. Carbon storage comparisons with lower-elevation soils on Mauna Kea and other moist mesic (2,500 mm rainfall) sites on Hawaii suggest that these soils have reached only between 1% and 15% of their capacity to retain carbon. Our results suggest that, after 20 kyr in low rainfall and a cold climate, weathering was decoupled from soil carbon accumulation patterns and the associated influence of vegetation on soil development. Overall, we conclude that the rate of carbon supply to the subsoil (driven by coupling of rainfall above ground plant production) is a governing factor of forms and amount of soil organic matter accumulation, while soil mineralogy remained relatively uniform.

Our understanding of climatic conditions, and therefore forcing factors, in North America during the past two glacial cycles is limited in part by the scarcity of long, well-dated, continuous paleoclimate records. Here, we present the first, to our knowledge, continuous, millennial-resolution paleoclimate proxy record derived from millimeter-thick pedogenic carbonate clast coatings (pedothems), which are widely distributed in semiarid to arid regions worldwide. Our new multiisotope pedothem record from the Wind River Basin in Wyoming confirms a previously hypothesized period of increased transport of Gulf of Mexico moisture northward into the continental interior from 70,000 to 55,000 years ago based on oxygen and carbon isotopes determined by ion microprobe and uranium isotopes and U-Th dating by laser ablation inductively coupled plasma mass spectrometry. This pronounced meridional moisture transport, which contrasts with the dominant zonal transport of Pacific moisture into the North American interior by westerly winds before and after 70,000-55,000 years ago, may have resulted from a persistent anticyclone developed above the North American ice sheet during Marine Isotope Stage 4. We conclude that pedothems, when analyzed using microanalytical techniques, can provide high-resolution paleoclimate records that may open new avenues into understanding past terrestrial climates in regions where paleoclimate records are not otherwise available. When pedothem paleoclimate records are combined with existing records they will add complimentary soil-based perspectives on paleoclimate conditions.

Management of biodiversity and ecosystem services requires a better understanding of the factors that influence soil biodiversity. We characterized the species (or genera) richness of 10 taxonomic groups of invertebrate soil animals in replicated monocultures of 14 temperate tree species. The focal invertebrate groups ranged from microfauna to macrofauna: Lumbricidae, Nematoda, Oribatida, Gamasida, Opilionida, Araneida, Collembola, Formicidae, Carabidae, and Staphylinidae. Measurement of invertebrate richness and ancillary variables occurred ~34 years after the monocultures were planted. The richness within each taxonomic group was largely independent of richness of other groups; therefore a broad understanding of soil invertebrate diversity requires analyses that are integrated across many taxa. Using a regression-based approach and ~125 factors related to the abundance and diversity of resources, we identified a subset of predictors that were correlated with the richness of each invertebrate group and richness integrated across 9 of the groups (excluding earthworms). At least 50% of the variability in integrated richness and richness of each invertebrate group was explained by six or fewer predictors. The key predictors of soil invertebrate richness were light availability in the understory, the abundance of an epigeic earthworm species, the amount of phosphorus, nitrogen, and calcium in soil, soil acidity, and the diversity or mass of fungi, plant litter, and roots. The results are consistent with the hypothesis that resource abundance and diversity strongly regulate soil biodiversity, with increases in resources (up to a point) likely to increase the total diversity of soil invertebrates. However, the relationships between various resources and soil invertebrate diversity were taxon-specific. Similarly, diversity of all 10 invertebrate taxa was not high beneath any of the 14 tree species. Thus, changes to tree species composition and resource availability in temperate forests will likely increase the richness of some soil invertebrates while decreasing the richness of others.

Sulfur, a nutrient required by terrestrial ecosystems, is likely to be regulated by atmospheric processes in well-drained, upland settings because of its low concentration in most bedrock and generally poor retention by inorganic reactions within soils. Environmental controls on sulfur sources in unpolluted ecosystems have seldom been investigated in detail, even though the possibility of sulfur limiting primary production is much greater where atmospheric deposition of anthropogenic sulfur is low. Here we measure sulfur isotopic compositions of soils, vegetation and bulk atmospheric deposition from the Hawaiian Islands for the purpose of tracing sources of ecosystem sulfur. Hawaiian lava has a mantle-derived sulfur isotopic composition (δ(34)S VCDT) of -0.8‰. Bulk deposition on the island of Maui had a δ(34)S VCDT that varied temporally, spanned a range from +8.2 to +19.7‰, and reflected isotopic mixing from three sources: sea-salt (+21.1‰), marine biogenic emissions (+15.6‰), and volcanic emissions from active vents on Kilauea Volcano (+0.8‰). A straightforward, weathering-driven transition in ecosystem sulfur sources could be interpreted in the shift from relatively low (0.0 to +2.7‰) to relatively high (+17.8 to +19.3‰) soil δ(34)S values along a 0.3 to 4100 ka soil age-gradient, and similar patterns in associated vegetation. However, sub-kilometer scale spatial variation in soil sulfur isotopic composition was found along soil transects assumed by age and mass balance to be dominated by atmospheric sulfur inputs. Soil sulfur isotopic compositions ranged from +8.1 to +20.3‰ and generally decreased with increasing elevation (0-2000 m), distance from the coast (0-12 km), and annual rainfall (180-5000 mm). Such trends reflect the spatial variation in marine versus volcanic inputs from atmospheric deposition. Broadly, these results illustrate how the sources and magnitude of atmospheric deposition can exert controls over ecosystem sulfur biogeochemistry across relatively small spatial scales.

© 2015 US Government Tree species interact with soil biota to impact soil organic carbon (C) pools, but it is unclear how this interaction is shaped by various ecological factors. We used multiple regression to describe how ~100 variables were related to soil organic C pools in a common garden experiment with 14 temperate tree species. Potential predictor variables included: (i) the abundance, chemical composition, and decomposition rates of leaf litter and fine roots, (ii) species richness and abundance of bacteria, fungi, and invertebrate animals in soil, and (iii) measures of soil acidity and texture. The amount of organic C in the organic horizon and upper 20 cm of mineral soil (i.e. the combined C pool) was strongly negatively correlated with earthworm abundance and strongly positively correlated with the abundance of aluminum, iron, and protons in mineral soils. After accounting for these factors, we identified additional correlations with soil biota and with litter traits. Rates of leaf litter decomposition, measured as litter mass loss, were negatively correlated with size of the combined soil organic C pool. Somewhat paradoxically, the combined soil organic C pool was also negatively related to the ratio of recalcitrant compounds to nitrogen in leaf litter. These apparent effects of litter traits probably arose because two independent components of litter “quality” were controlling different aspects of decomposition. Leaf litter mass loss rates were positively related with leaf litter calcium concentrations, reflecting greater utilization and depolymerization of calcium-rich leaf litter by earthworms and other soil biota, which presumably led to greater proportional losses of litter C as CO2 or dissolved organic C. The fraction of depolymerized and metabolized litter that is ultimately lost as CO2 is an inverse function of microbial C use efficiency, which increases with litter nutrient concentrations but decreases with concentrations of recalcitrant compounds (e.g. lignin); thus, high ratios of recalcitrant compounds to nitrogen in leaf litter likely caused a greater fraction of depolymerized litter to be lost as CO2. Existing conceptual models of soil C stabilization need to reconcile the effects of litter quality on these two potentially counteracting factors: rates of litter depolymerization and microbial C use efficiency.

Mass balance models have become standard tools for characterizing element gains and losses and volumetric change during weathering and soil development. However, they rely on the assumption of complete immobility for an index element such as Ti or Zr. Here we describe a dual-phase mass balance model that eliminates the need for an assumption of immobility and in the process quantifies the contribution of aqueous versus colloidal element transfer. In the model, the high field strength elements Ti and Zr are assumed to be mobile only as suspended solids (colloids) and can therefore be used to distinguish elemental redistribution via colloids from redistribution via dissolved aqueous solutes. Calculations are based upon element concentrations in soil, parent material, and colloids dispersed from soil in the laboratory. We illustrate the utility of this model using a catena in South Africa. Traditional mass balance models systematically distort elemental gains and losses and changes in soil volume in this catena due to significant redistribution of Zr-bearing colloids. Applying the dual-phase model accounts for this colloidal redistribution and we find that the process accounts for a substantial portion of the major element (e.g., Al, Fe and Si) loss from eluvial soil. In addition, we find that in illuvial soils along this catena, gains of colloidal material significantly offset aqueous elemental loss. In other settings, processes such as accumulation of exogenous dust can mimic the geochemical effects of colloid redistribution and we suggest strategies for distinguishing between the two. The movement of clays and colloidal material is a major process in weathering and pedogenesis; the mass balance model presented here is a tool for quantifying effects of that process over time scales of soil development.

Many researchers believe that prehistoric Rapa Nui society collapsed because of centuries of unchecked population growth within a fragile environment. Recently, the notion of societal collapse has been questioned with the suggestion that extreme societal and demographic change occurred only after European contact in AD 1722. Establishing the veracity of demographic dynamics has been hindered by the lack of empirical evidence and the inability to establish a precise chronological framework. We use chronometric dates from hydrated obsidian artifacts recovered from habitation sites in regional study areas to evaluate regional land-use within Rapa Nui. The analysis suggests region-specific dynamics including precontact land use decline in some near-coastal and upland areas and postcontact increases and subsequent declines in other coastal locations. These temporal land-use patterns correlate with rainfall variation and soil quality, with poorer environmental locations declining earlier. This analysis confirms that the intensity of land use decreased substantially in some areas of the island before European contact.

We used measurements from airborne imaging spectroscopy and LiDAR to quantify the biophysical structure and composition of vegetation on a dryland substrate age gradient in Hawaii. Both vertical stature and species composition changed during primary succession, and reveal a progressive increase in vertical stature on younger substrates followed by a collapse on Pleistocene-aged flows. Tall-stature Metrosideros polymorpha woodlands dominated on the youngest substrates (hundreds of years), and were replaced by the tall-stature endemic tree species Myoporum sandwicense and Sophora chrysophylla on intermediate-aged flows (thousands of years). The oldest substrates (tens of thousands of years) were dominated by the short-stature native shrub Dodonaea viscosa and endemic grass Eragrostis atropioides. We excavated 18 macroscopic charcoal fragments from Pleistocene-aged substrates. Mean radiocarbon age was 2,002 years and ranged from < 200 to 7,730. Genus identities from four fragments indicate that Osteomeles spp. or M. polymorpha once occupied the Pleistocene-aged substrates, but neither of these species is found there today. These findings indicate the existence of fires before humans are known to have occupied the Hawaiian archipelago, and demonstrate that a collapse in vertical stature is prevalent on the oldest substrates. This work contributes to our understanding of prehistoric fires in shaping the trajectory of primary succession in Hawaiian drylands.

Lithium isotopes are a potential tracer of silicate weathering but the relationship between lithium isotope compositions and weathering state still need to be established with precision. Here, we report Li concentrations and Li isotope compositions of soils developed along a 4 million year humid-environment chronosequence in the Hawaiian Islands. Li concentrations are variable with depth and age, ranging from 0.24 to 21.3ppm, and significant Li depletions (up to 92%) relative to parent basalts are systematically enhanced towards the surface. Our calculations show that the relative contribution from atmospheric deposits to the Li soil budget remains small, with a maximum contribution from dust Li of 20% at the oldest site. This is explained by the capacity of the weathering products to retain, within the profiles, the Li coming from basalt alteration, and allows us to explore more specifically the role of alteration processes on soil Li isotope signatures. The δ7Li values display a large range between -2.5‰ and +13.9‰ The youngest soils (0.3ka) display the same δ7Li value as fresh basalt, regardless of depth, despite ~30% Li loss by leaching, indicating that there is little Li isotope fractionation during the incipient stage of weathering. δ7Li values for the older soils (≥20ka) vary non-linearly as a function of time and can be explained by progressive mineral transformations starting with the synthesis of metastable short-range order (nano-crystalline) minerals and followed by their transformation into relatively inert secondary minerals. Results highlight significant Li isotope fractionation during secondary mineral formation and in particular during Li uptake by kaolinite. Finally, we suggest that the non-monotonous evolution of the regolith δ7Li value over the last 4Ma is consistent with climatic variations, where congruent release of Li isotopes occurs during warmer periods.

The transport of organic carbon (C) to deep mineral horizons in soils can lead to long-term C stabilization. In basaltic soils, C associations with short-range-ordered (SRO) minerals often lead to colloid-sized aggregates that can be dispersed and mobilized by changes in soil solution chemistry. In the montane forest region of Hawaii, basaltic soils are exposed to high rainfall and anoxic conditions that facilitate ferric (FeIII) (oxyhydr)oxide reduction. We explored the potential of iron (Fe)-reducing conditions to mobilize C by exposing the surface mineral horizons of three soils from the Island of Hawai'i (aged 0.3, 20, and 350ky) to 21days of anoxic incubation in 1:10 soil slurries. Mobilized C was quantified by fractionating the slurries into three particle-size classes (

Question: How do canopy disturbance and soil properties structure vascular plant community species composition and resilience to encroachment by exotic species in a tropical montane wet forest? Location: Hawai'i Experimental Tropical Forest (HETF), a tropical montane wet forest, on Mauna Kea, Hawai'i Island, Hawai'i, USA. Methods: Previous studies employing airborne LiDAR were used to define three zones across an elevation gradient from 900 to 1500 m. Within each zone, a ~1000-m block transect was selected to cross two different volcanic substrates: one derived from surface lava and one derived from thick ash deposits. Non-metric multidimensional scaling (NMS) scores of vegetation data were related to independently-derived environmental NMS scores and spatial location with generalized additive models (GAM). Results: Vascular plant species composition in all elevation zones consists of three NMS axes, which are best modelled by one of three possible environmental NMS axes or by location. The first NMS axis of species composition in the lowest elevation zone (40% variance explained (VE)) is a function of location on volcanic substrates (61% deviance explained (DE)). The second lowest elevation axis (27% VE) is a function of unexplained spatial heterogeneity (31% DE). The third lowest elevation NMS axis (24% VE) is a function of the spatial mosaic of canopy disturbance (16% DE). In the middle elevation zone, species composition most strongly relates to the interaction between volcanic substrate and the condition of the soil surface for all three NMS axes (41%, 27%, 24% VE; 70%, 16%, 24% DE). The primary axis of species composition in the highest elevation zone (41% VE) corresponds with substrate and soil condition (55% DE) but the second and third axes of species composition (27% and 25% VE) relate to canopy dieback disturbance (36%, 14% DE). Counts of exotic species and 0-2 m height class native tree species respond to the type of volcanic substrate and soil surface condition in all three elevation zones. Lava-derived substrates have a higher incidence of exotic species and less native tree regeneration; whereas ash-derived substrates have higher numbers of native tree species regenerating and many fewer exotic species. Discussion: The tropical montane forests on Mauna Kea reflect a native-dominated plant community response to disturbance on both lava- and ash-derived volcanic substrates, and a higher propensity for exotic species to occur on the lava-derived substrate. Native plant communities on ash-derived soils may have higher resilience to exotic invasion than communities on lava-derived substrates. Our results indicate resource managers should explicitly account for variation in soils and substrate type when prioritizing, implementing and monitoring management interventions to foster native plant assemblages and control the spread of exotic and invasive species. © 2013 International Association for Vegetation Science.