The genus Ribes is distinguished within the Grossulariaceae by the presence of dense glandular trichomes on all above-ground organs—leaves, stems, buds, and fruit—that constitutively secrete a complex mixture of volatile terpenoids. Gas Chromatography-Mass Spectrometry (GC-MS) analysis of R. nigrum bud essential oil has identified 19 monoterpenes, 20 sesquiterpenes, 12 carbonyl compounds, 11 esters, and 34 alcohols [1]. Headspace analyses of intact leaves and berries confirm that monoterpene hydrocarbons and oxygenated monoterpenes are the dominant volatile classes in constitutive emissions [8]. Critically, emission is constitutive—it does not require wounding, herbivory, or fruiting—meaning that an established Ribes planting generates a persistent, continuous scent field throughout the growing season. VOC emissions are modulated by temperature and radiation, suggesting that emission intensity will be elevated during the warm Australian growing season coinciding with peak nesting periods for the target fauna. The characteristic “catty” sulfurous note of R. nigrum stems—attributable to 4-methoxy-2-methyl-2-butanethiol—is co-emitted with the terpenoid suite, potentially broadening the olfactory signal relevant to mammalian deterrence.

It should be noted that the majority of published VOC characterization data cited above derives from R. nigrum and other Ribes species rather than from R. odoratum directly. The hypothesis that R. odoratum functions as an olfactory deterrent therefore involves a degree of genus-level inference. This inference is supported by the shared glandular trichome architecture and constitutive terpenoid emission characteristic of the genus [1][7], and is strengthened by direct evidence specific toR. odoratum: its VOC profile is documented to include a prominent eugenol and phenylpropanoid fraction absent or minor in R. nigrum [9], and the species is recorded as deer-resistant across its native North American range [10], consistent with a functional deterrent effect on large browsing mammals. The present study is designed in part to generate the first direct field evidence for R. odoratum as a mammalian deterrent, replacing genus-level inference with site-specific empirical data.

Rooting mammals—including Sus scrofa, feral deer, and rabbits—locate buried food resources primarily through olfaction, using ground-level scent detection via direct snout-soil contact. Laboratory and field studies have demonstrated that strong olfactory masking agents can significantly reduce rooting frequency and excavation success. A castor oil, ginger, and garlic mixture has been reported to achieve up to approximately 95% deterrence efficacy against wild boar in field conditions [6], establishing proof of concept for olfactory deterrence. Capsaicin-based and essential-oil-based repellents have demonstrated efficacy against white-tailed deer (Odocoileus virginianus) in agricultural contexts [11].

The mechanistic basis for monoterpene-mediated deterrence is not fully characterised, but proposed pathways include: 1) olfactory masking of the target food scent signature at ground surface; 2) trigeminal irritation via volatile terpenoids at ground-surface concentrations, potentially mediated by TRPA1 ion channel activation as recently described for Piperales VOCs in large herbivores [12]; and 3) conditioned aversion following repeated exposure to a novel, persistent, and non-food-associated scent landscape. These mechanisms are not mutually exclusive and may operate simultaneously at a living plant barrier.

Ribes odoratum Pursh (syn. R.aureumvar.villosum) is a thornless, deciduous shrub native to the central-western plains of North America, growing to 2 - 2.5 m height and width. Its USDA hardiness zone range of 3 - 9 reflects tolerance of both severe cold (to −25˚C) and hot, dry summers—a combination that renders it uniquely suitable for deployment in the semi-arid mallee and warm-temperate coastal zones of southern Australia where R. nigrum’s high chilling requirement (1200 - 2500 hours below 7˚C) makes it unviable [13]. The species produces strongly scented foliage and intensely clove-scented flowers—a VOC profile dominated by eugenol and related phenylpropanoids in addition to the terpenoid suite characteristic of the genus. It tolerates poor soils and moderate drought, spreads by root suckers to form dense thickets, and is not listed as a declared weed in any Australian jurisdiction [14]. The absence of Pinus strobus (white pine) populations in Australia renders the white pine blister rust (Cronartiumribicola) pathway—the basis for US cultivation restrictions on Ribes—entirely irrelevant in the Australian context.

While aromatic shrubs such as Lavandula spp. and Rosmarinus officinalis are widely recognized as drought-tolerant and VOC-emitting, their utility as mammalian olfactory barriers at ecologically meaningful scales is constrained by growth form. Both reach maximum heights of approximately 0.5 - 1.2 m under field conditions, limiting total above-ground biomass and the vertical extent of any olfactory boundary they could sustain. By contrast, R. odoratum achieves a mature height and spread of approximately 1.8 - 2.4 m [9], produces a constitutive phenylpropanoid VOC load—principally eugenol—distributed across all above-ground tissues including stems, leaves, buds, and flowers [7], and is documented as deer-resistant in its native range [10], suggesting that its VOC profile already functions as a deterrent to large browsing mammals. No comparable combination of canopy height, multi-tissue constitutive emission, and confirmed large-mammal aversion is available in a single alternative aromatic species suited to the moisture-variable conditions of southern Australia.

4.1.1. Ecological Context

The malleefowl (Leipoaocellata Gould, 1840) is listed as Vulnerable under the Commonwealth Environment Protection and Biodiversity Conservation (EPBC) Act and Threatened under the Victorian Flora and Fauna Guarantee Act [15]. As one of only three mound-building megapodes in Australia and the sole species inhabiting arid and semi-arid zones, the malleefowl constructs large incubation mounds—up to 1 m in height and 4 m in diameter—from leaf litter and sandy substrate, within which eggs are incubated for approximately 60 days with daily male thermal regulation to within 1˚C of 33˚C - 34˚C. Feral deer have been formally identified as an emerging threat in South Australia through physical disruption of mound architecture, compromising thermal regulation essential for embryo viability. Feral pigs disturb mounds through rooting for invertebrates in the organic litter that forms the thermal core. The mallee and mulga scrublands of the Grampians-Little Desert-Mallee corridor experience hot, dry summers with cool winters accumulating moderate frost—conditions within the full viability range of R. odoratum.

4.1.2. Experimental Design

Twenty active malleefowl mounds will be identified in partnership with the Malleefowl Recovery Group [16] and Australian Wildlife Conservancy sanctuaries (Dakalanta, Scotia, Mt Gibson). Mounds will be paired by vegetation type, mound age, and estimated activity level, generating ten matched pairs. Within each pair, one mound will be randomly assigned to treatment (olfactory barrier planting) and one to unplanted control. Treatment mounds will receive a perimeter planting of R. odoratum at 5 m radius from mound centre, using established bare-root stock planted at 1 m intervals to form a continuous shrub ring (approximately 31 plants per mound). Plantings will be established 12 months prior to data collection to allow root establishment and initiation of full foliar VOC emission.

4.1.3. Outcome Measures

Camera-trap recordings: frequency of feral deer, feral pig, and rabbit visits per 24-hour period at each mound.Mound structural integrity: monthly photographic assessment and volumetric measurement; all disturbance events recorded with date, estimated cause, and extent.Nesting success: number of incubation attempts and emergence events recorded on camera; ratio of successful to failed incubation attempts.VOC gradient sampling: passive SPME samplers at 0.5 m, 2 m, 5 m, and 10 m from planting edge to characterize the effective scent field radius.

4.2.1. Ecological Context

The Eastern Ground Parrot (Pezoporuswallicuswallicus Kerr, 1792) is listed as Endangered under the EPBC Act and is one of only two fully ground-dwelling parrot species in Australia, nesting in shallow scrapes directly beneath coastal heathland vegetation. This nesting strategy renders eggs and chicks acutely vulnerable to rooting disturbance by feral pigs, which probe through heath understorey for roots, corms, and invertebrates—precisely the foraging action that would expose and destroy a ground parrot nest. Ground parrot populations are now confined to isolated heathland patches along the coast of southeastern Victoria and eastern South Australia, where mild cool winters provide adequate but not excessive chilling for R. odoratum establishment at heathland margins.

4.2.2. Experimental Design

In collaboration with Parks Victoria and the Department for Environment and Water (South Australia), sixteen ground parrot territories (identified by dawn call surveys) will be selected across four heathland reserves. Eight territories will receive treatment plantings of R. odoratum as linear barrier rows along the heathland-access track boundary—the primary pig ingress vector—planted at 2 m spacing; eight will serve as unplanted controls. Territorial boundaries will be mapped by GPS tracking of radio-tagged individuals where ethics permits.

4.2.3. Outcome Measures

Feral pig activity index: rooting scar frequency and area within each territory per monitoring interval using standardized monthly transect surveys.Nest survival (secondary outcome): egg fate (survived/depredated/abandoned) will be recorded via miniaturized nest cameras only where nests can be located without disturbance; if ethical nest location is not achievable, this outcome will be excluded and territory occupancy and pig activity index will serve as the primary proxies for breeding success.Territory occupancy: dawn call survey counts before, during, and after breeding season as a proxy for breeding activity.VOC ambient concentration at heathland boundary: portable PTR-ToF-MS measurements or equivalent during peak growth.

4.3.1. Ecological Context

Multiple EPBC-listed freshwater turtle species nest along exposed riparian and lacustrine banks across southeastern Queensland and northeastern New South Wales, including the Critically Endangered Bellinger River Snapping Turtle (Myuchelysgeorgesi Wells & Wellington, 1985) and the Endangered Manning River Helmeted Turtle (Flaviemyspurvisi Wells & Wellington, 1985). Females deposit eggs in spatially predictable sandy or loamy bank substrate, rendering nest sites highly susceptible to repeated predation by feral pigs that learn locations through olfactory detection of egg scent. Feral pig populations in this region are among the densest in eastern Australia, concentrated in the wetland and riparian habitats freshwater turtles depend on. The warm, humid winters of this region place it at the upper warm boundary of R.odoratum’s viable range, making this the most climatically marginal trial site and the most important for testing VOC emission and plant performance under warm-winter conditions.

4.3.2. Experimental Design

Twelve known nesting bank sections (minimum 50 m length, confirmed by previous surveys or egg remains) will be identified on three river systems in collaboration with state freshwater turtle recovery teams. Sections will be stratified by river system and randomly allocated: six to treatment (linear R. odoratum plantings along bank crest at 1.5 m intervals, 3 m setback from active nesting substrate edge) and six to unplanted control. The crest planting position targets the primary pig approach vector while minimizing interference with nesting substrate.

4.3.3. Outcome Measures

Pig rooting disturbance: frequency of rooting events per nesting bank section per week, from camera traps covering the full nesting zone.Nest fate: GPS-flagged nest locations monitored for egg survival across the full incubation period (60 - 90 days depending on species).Pig approach behaviour: camera-trap recordings of head-raising (scenting), direction changes, and retreat frequency at treatment versus control banks.Plant establishment and VOC output: quarterly plant health assessments and SPME VOC sampling to confirm establishment success and ongoing emission under subtropical conditions.

Bare-root or container-grown R. odoratum stock will be sourced from certified Australian nurseries or propagated from hardwood cuttings taken from established parent plants 18 months prior to trial planting. Provenance selection will favour ecotypes originating from warm-summer continental climates (Kansas, Nebraska, Oklahoma) to maximize performance in Australian conditions. All plant material will be propagation-certified, screened for Phytophthora and blister rust compliance, and approved for planting by state biosecurity authorities prior to deployment.

Concurrent background VOC samples will be collected at each sampling interval outside the planting boundary to establish ambient baseline concentrations independent of the Ribes odoratum planting. Ideally, background sampling will commence prior to planting establishment to capture pre-treatment baseline VOC conditions at each site. Wind speed, temperature, and relative humidity will be recorded at each sampling interval to account for microclimatic variation in VOC dispersal and emission rate. Foliar VOC emissions will be characterized at each site using passive SPME fibre samplers (Supelco PDMS/DVB, 65 µm film thickness) exposed for 24-hour periods at characterized heights (5 cm, 50 cm, and 100 cm above ground level) and at distances of 0.5, 2, 5, and 10 m from the planting edge. Fibres will be desorbed and analyzed by GC-MS (Shimadzu GCMS-QP2020 or equivalent) with identification against the NIST 2020 mass spectral library. Quantification will use external standard calibration for the 15 most abundant compounds identified in the pilot characterization phase.

Reconyx HyperFire 2 (or equivalent) camera traps will be deployed at each nesting site in paired configurations: one camera oriented toward the nesting structure or substrate and one oriented to capture the approach vector from the planting edge. Cameras will be set to continuous video mode at motion trigger, 30 fps, infrared illumination. Footage will be reviewed by two independent coders blind to treatment allocation, coding: species identity, visit duration, rooting and digging events, scenting behaviour (head orientation, snout-ground contact), and approach or retreat decisions at the planting boundary.

Primary outcomes will be analyzed separately for each site using generalized linear mixed models (GLMMs), with treatment as a fixed effect, site-specific random effects, and season and year as covariates. A secondary cross-site pooled GLMM will then be fitted incorporating site-by-treatment interaction terms to formally test whether treatment efficacy varies across the three trial landscapes. This combined approach preserves site-level interpretability—important given the ecological distinctiveness of each site—while providing a synthesized cross-site estimate of overall deterrent efficacy. All analyses will be conducted in R v.4.3+ using the lme4 and glmmTMB packages.

Sample sizes were determined based on practical and ecological constraints at each site. For Site 1, twenty malleefowl mounds in ten matched pairs provide adequate replication given the logistical constraints of locating active mounds across the Grampians-Little Desert-Mallee corridor. For Site 2, sixteen ground parrot territories across four reserves balance statistical replication with the limited availability of confirmed territories in accessible heathland. For Site 3, twelve freshwater turtle nesting bank sections across three river systems reflect the availability of confirmed nesting sites identified by recovery teams. Baseline event rates for rooting disturbance frequency, nesting success, and territory occupancy are expected to vary across sites and will be empirically established during the pre-treatment monitoring period prior to data collection, at which point power estimates will be refined accordingly.

The minimum treatment effect the study is designed to detect is a 35% reduction in rooting event frequency in treatment versus control plots, representing a conservative and ecologically meaningful threshold for a first-of-its-kind passive deterrence trial. As a secondary supporting measure, deterrent behaviour—defined as head-raising, direction change, or retreat at the planting boundary recorded by camera trap—will be reported as a qualitative indicator of olfactory deterrence independent of rooting frequency outcomes.

Ribes emits a constitutive blend of phenylpropanoid and terpenoid VOCs—principally eugenol and associated monoterpenes in R. nigrum—from leaves, stems, buds, and flowers across all above-ground tissues throughout the growing season [7]. Because VOC output scales with photosynthetically active biomass, and because olfactory deterrence requires sustained atmospheric concentrations above the detection threshold of the target species, planting date alone is an insufficient basis for defining treatment onset of R. odoratum. A planted individual was therefore classified as an active treatment unit only upon meeting all of the following conditions concurrently: completion of a minimum of two full growing seasons post-transplant, at which point Ribes plants will generally have achieved their first fruiting and an established multi-stem architecture [9]; attainment of sufficient foliar biomass to sustain constitutive VOC emission above local background during the foliated period [7]; and, in barrier configurations, canopy contact or near-contact with adjacent individuals. Given a growth rate of approximately 30 - 45 cm per year and a mature spread of 0.9 - 1.5 m [10], with full mature dimensions of approximately 2 m height and spread [9], a planting spacing of 1.5 - 2.0 m between centres is projected to achieve canopy continuity within three to four growing seasons under the trial conditions described below. Barrier segments in which one or more individuals had not yet reached active treatment status were excluded from deterrence assessments for that measurement period.