skip to main content


The NSF Public Access Repository (NSF-PAR) system and access will be unavailable from 11:00 PM ET on Thursday, May 23 until 2:00 AM ET on Friday, May 24 due to maintenance. We apologize for the inconvenience.

Title: The role of microtopography and resident species in post‐disturbance recovery of arid habitats in Hawaiʻi

Habitat‐suitability indices (HSI) have been employed in restoration to identify optimal sites for planting native species. Often, HSI are based on abiotic variables and do not include biotic interactions, even though similar abiotic conditions can favor both native and nonnative species. Biotic interactions such as competition may be especially important in invader‐dominated habitats because invasive species often have fast growth rates and can exploit resources quickly. In this study, we test the utility of an HSI of microtopography derived from airborne LiDAR to predict post‐disturbance recovery and native planting success in native shrub‐dominated and nonnative, invasive grass‐dominated dryland habitats in Hawaiʻi. The HSI uses high‐resolution digital terrain models to classify sites' microtopography as high, medium, or low suitability, based on wind exposure and topographic position. We used a split‐plot before‐after‐control‐impact design to implement a disturbance experiment within native shrub (Dodonaea viscosa) and nonnative, invasive grass (Cenchrus clandestinus)‐dominated ecosystems across three microtopography categories. In contrast to previous studies using the same HSI, we found that microtopography was a poor predictor of pre‐disturbance conditions for soil nutrients, organic matter content, or foliar C:N, within bothDodonaeaandCenchrusvegetation types. In invader‐dominatedCenchrusplots, microtopography helped predict cover, but not as expected (i.e., highest cover would be in high‐suitability plots):D. viscosahad the greatest cover in low‐suitability andC. clandestinushad the greatest cover in medium‐suitability plots. Similarly, in native‐dominatedDodonaeaplots, microtopography was a poor predictor ofD. viscosa,C. clandestinus, and total plant cover. Although we found some evidence that microtopography helped inform post‐disturbance plant recovery ofD. viscosaand total plant cover, vegetation type was a more important predictor. Important for considering the success of plantings, percent cover ofD. viscosadecreased while percent cover ofC. clandestinusincreased within both vegetation types 20 months after disturbance. Our results are evidence that HSIs based on topographic features may prove most useful for choosing planting sites in harsh habitats or those already dominated by native species. In more productive habitats, competition from resident species may offset any benefits gained from “better” suitability sites.

more » « less
Author(s) / Creator(s):
 ;  ;  ;  ;  
Publisher / Repository:
Wiley Blackwell (John Wiley & Sons)
Date Published:
Journal Name:
Ecological Applications
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Abstract

    Improving models of community change is a fundamental goal in ecology and has renewed importance during global change and increasing human disturbance of the biosphere. Using the Mojave Desert (southwestern United States) as a model system, invaded by nonnative plants and subject to wildfire disturbances, we examined models of resilience, alternative stable states, and convergent‐divergent trajectories for 36 yr of plant community change after 31 wildfires in communities dominated by the native shrubsLarrea tridentataorColeogyne ramosissima. Perennial species richness on average was fully resilient within 23 yr after disturbance in both community types. Perennial cover was fully resilient within 25 yr in theLarreacommunity, but recovery was projected to require 52 yr in theColeogynecommunity. Species composition shifts were persistent, and in theColeogynecommunity, the projected compositional recovery time of 550 yr and increasing resembled a deflected trajectory toward potential alternative states. Disturbed sites contained a perennial species composition of predominately short‐statured forbs, subshrubs, and grasses, contrasting with the larger‐statured shrub and tree structure of undisturbed sites. Auxiliary data sets characterizing species recruitment, annual plants including nonnative grasses, biocrust communities, and soils showed persistent differences between disturbed and undisturbed sites consistent with positive feedbacks potentially contributing to alternative stable states. Resprouting produced limited resilience for the large shrubsL. tridentataandYuccaspp. important to population persistence but did not forestall long‐term reduced abundance of the species. The nonnative annual grassBromus rubensincreased on disturbed sites over time, suggesting persistently abundant nonnative plant fuels and reburn potential. Biocrust cover on disturbed sites was half and species richness a third of amounts on undisturbed sites. Soil nitrogen was 30% greater on disturbed sites and no significant trend was evident for it to decline on even the oldest burns. Disturbed desert plant communities simultaneously supported all three models of resilience, alternative stable states, and convergent‐divergent trajectories among community measures (e.g., species richness, composition), timeframes since disturbance, and spatial resolutions. Accommodating expression within ecosystems of multiple models, including those opposing each other, may help broaden theoretical models of ecosystem change.

    more » « less
  2. Abstract

    Harsh habitats dominated by invasive species are difficult to restore. Invasive grasses in arid environments slow succession toward more desired composition, yet grass removal exacerbates high light and temperature, making the use of “nurse plants” an appealing strategy. In this study of degraded subtropical woodlands dominated by alien grasses in Hawai'i, we evaluated whether individuals of two native (Dodonaea viscosa,Leptocophylla tameiameia) and one non‐native (Morella faya) woody species (1) act as natural nodes of recruitment for native woody species and (2) can be used to enhance survivorship of outplanted native woody species. To address these questions, we quantified the presence and persistence of seedlings naturally recruiting beneath adult nurse shrubs and compared survival and growth of experimentally outplanted seedlings of seven native woody species under the nurse species compared to intact and cleared alien‐grass plots. We found that the two native nurse shrubs recruit their own offspring, but do not act as establishment nodes for other species.Morella fayarecruited even fewer seedlings than native shrubs. Thus, outplanting will be necessary to increase abundance and diversity of native woody species. Outplant survival was the highest under shrubs compared to away from them with few differences between nurse species. The worst habitat for native seedling survival and growth was within the unmanaged invasive grass matrix. Although the two native nurse species did not differentially affect outplant survival,D. viscosais the most widespread and easily propagated and is thus more likely to be useful as an initial nurse species. The outplanted species showed variable responses to nurse habitats that we attribute to resource requirements resulting from their typical successional stage and nitrogen fixation capability.

    more » « less
  3. null (Ed.)
    Abstract Reducing invasive species abundance near the leading edge of invasions is important for maintaining diverse, high-functioning ecosystems, but it can be hard to remove invasives present at low levels within desirable plant communities. Focusing on an invasive annual grass, Bromus tectorum , near the edge of its range in the southern Colorado Plateau, we used an observational study to ask what plant community components were associated with lower levels of B. tectorum , and a manipulative experiment to ask if targeted spring grazing or seeding native competitors were effective for reversing low-level invasion. The observational study found that higher C 3 perennial grass cover and shrub cover were associated with lower B. tectorum abundance, and adult Poa fendleriana and Pascopyrum smithii plants had the fewest B. tectorum individuals within 50 cm. Our manipulative experiment used a randomized, hierarchical design to test the relative effectiveness of seeding native perennial grasses using different spatial planting arrangements, seeding rates, seed enhancements, and targeted spring grazing. Two years after seeding, seeded species establishment was 36% greater in high seed rate than unseeded plots, and high rate plots also had lower B. tectorum cover. One season after targeted spring grazing (a single, 2-week spring-grazing treatment 17 months post-seeding), grazed paddocks displayed trends towards higher seeded species densities and lower B. tectorum biomass in certain seeding treatments, compared to ungrazed paddocks. Results suggest high rate native grass seedings may be effective and short-duration spring grazing should be further evaluated as potential tools for preventing ecosystem conversion along invasion fronts. 
    more » « less
  4. The impact of a biological invasion on native communities is expected to be uneven across invaded landscapes due to differences in local abiotic conditions, invader abundance, and traits and composition of the native community. One way to improve predictive ability about the impact of an invasive species given variable conditions is to exploit known mechanisms driving invasive species' success. Invasive plants frequently exhibit allelopathic traits, which can be directly toxic to plants or indirectly impact them via disruption of root symbionts, including mycorrhizal fungi. The indirect mechanism – mutualism disruption – is predicted to impact plants that rely on mycorrhizas but not affect non‐mycorrhizal plant species. To assess whether invader‐driven mutualism disruption explains observed changes in native plant communities, we analyzed long‐term (1998–2018) plant cover data from forest plots across the state of Illinois. We evaluated native plant communities experiencing a range of abundance of invasive allelopathic garlic mustardAlliaria petiolataand varying environmental conditions. Consistent with the mutualism disruption hypothesis, we showed that as garlic mustard abundance increased over time in 0.25 m2sampling quadrats, the abundance of mycorrhizal plant species decreased, but non‐mycorrhizal plant species did not. Over space and time, garlic mustard abundance predicted plant abundances and diversity at the quadrat level, but this relationship was not present at a larger scale when quadrats were aggregated within sites. Garlic mustard's impact on the plant community was highly localized, yet it was as important as abiotic variables for predicting local plant diversity. We showed that garlic mustard abundance was a key predictor of patterns of plant diversity across invasion intensity and environmental heterogeneity in a way that is consistent with mutualism disruption. Our work indicates that the mutualism disruption hypothesis can provide generalizable predictions of the impacts of allelopathic invasive plants that are evident at a broad spatial scale.

    more » « less
  5. Abstract

    Nonnative conifers are widespread in the southern hemisphere, where their use as plantation species has led to adverse ecosystem impacts sometimes intensified by invasion. Mechanical removal is a common strategy used to reduce or eliminate the negative impacts of nonnative conifers, and encourage native regeneration. However, a variety of factors may preclude active ecological restoration following removal. As a result, passive restoration – unassisted natural vegetation regeneration – is common following conifer removal. We asked, ‘what is the response of understorey cover to removal of nonnative conifer stands followed by passive restoration?' We sampled understorey cover in three site types: two‐ to ten‐year‐old clearcuts, native forest and current plantations. We then grouped understorey species by origin (native/nonnative) and growth form, and compared proportion and per cent cover of these groups as well as of bare ground and litter between the three site types. For clearcuts, we also analysed the effect of time since clearcut on the studied variables. We found that clearcuts had a significantly higher average proportion of nonnative understorey vegetation cover than native forest sites, where nonnative vegetation was nearly absent. The understorey of clearcut sites also averaged more overall vegetation cover and more nonnative vegetation cover (in particular nonnative shrubs and herbaceous species) than either plantation or native forest sites. Notably, 99% of nonnative shrub cover in clearcuts was the invasive nonnative species Scotch broom (Cytisus scoparius). After ten years of passive recovery since clearcutting, the proportion of understorey vegetation cover that is native has not increased and remains far below the proportion observed in native forest sites. Reduced natural regeneration capacity of the native ecosystem, presence of invasive species in the surrounding landscape and land‐use legacies from plantation forestry may inhibit native vegetation recovery and benefit opportunistic invasives, limiting the effectiveness of passive restoration in this context.

    Abstract in Spanish is available with online material.

    more » « less