Predicting trajectories of dryland wetland vegetation transformation under climate change: a case study of the northern Murray–Darling Basin, Australia

Introduction

Dryland wetlands are an ecologically important ecosystem as they support a range of biodiversity in otherwise harsh landscapes with low and patchily distributed resource availability (Sheldon et al. 2010; Bino et al. 2015; Reynolds et al. 2016; Haig et al. 2019). The formation of dryland wetland vegetation communities relies on the capacity for plants to establish and survive hydrological input (i.e. rainfall or flooding) and, subsequently, persist during prolonged periods without this input (i.e. drought) (Webb et al. 2012). Many plant species in these habitats utilise strategies such as seed dormancy to persist over long periods until germination is stimulated by environmental factors such as rain, flooding, fire or temperature fluctuations (Casanova and Brock 2000; Ooi et al. 2009; Brock 2011; Capon and Reid 2016). Hydrological input and variability are highly influential over resident biota in dryland wetlands, often shaping composition through boom-and-bust cycles and structuring floodplain populations and assemblages spatially (Bunn et al. 2006; Capon and Brock 2006).

Climate change is driving changes to the duration, frequency, and severity of hydrological inputs, creating novel challenges for dryland wetland ecosystems (Sandi et al. 2020a). These impacts can manifest as extreme and intense periods of drought (Capon and Reid 2016) and increasingly severe and unpredictable flooding events (Ballinger and Nally 2006), as well as rising temperatures and more frequent and intense heatwaves. These impacts of climate change are likely to cause changes to river and wetland flow regimes (e.g. Murray et al. 2012; Mosner et al. 2015; Colloff et al. 2016).

As a result of its effects on hydrological inputs, climate change has the potential to profoundly affect the way wetland vegetation is able to disperse, establish and persist. The effects of extreme drought in dryland wetlands range from forest die-back to community composition shifts (Broich et al. 2014; Doody et al. 2014). In many instances globally, temperature increase has begun facilitating plant invasions (e.g. Zhang R et al. 2011; Sheppard et al. 2012). Prolonged, excessive, or unseasonal flooding can negatively affect vegetation in every life phase, including mortality in both seedlings and mature plants (Stokes et al. 2010; Mac Nally et al. 2011). Additionally, the periodic flooding of dryland wetlands is typically necessary to encourage germination and maintain vegetation condition, and to alleviate the impacts of intensive land use (Dawson et al. 2017; Broich et al. 2018).

Changes to the natural flow regime are often a result of land-use demands and will be further exacerbated by the impacts of climate change (Nielsen and Brock 2009; Lough and Hobday 2011; Quijano-Baron et al. 2022). Principally, the alteration of natural flow regimes creates novel conditions for which resident biota may be maladapted (Chen et al. 2021). For example, increased permanence of flows compared with an intermittent flow regime, or the drying out of systems, leading to negative outcomes for wetland vegetation (Greet et al. 2011, 2012; Webb et al. 2012; Shaeri Karimi et al. 2021). Other impacts of flow-regime alteration are contextual but can include the uncharacteristic movement of sediment and propagules, increases or decreases in overbank flows, and the disruption of stream morphology (Greet et al. 2012; Rolls et al. 2012; Swirepik et al. 2016; Shaeri Karimi et al. 2021). The result of the changing of such conditions can include shifts in native resident plant communities including via the arrival and dominance of non-native species (Catford et al. 2011). The combination of drought and augmented flow regime is associated with plant invasions (Catford et al. 2014).

The sensitivity of dryland wetlands to climate change is exacerbated by disturbance regimes, many of which are associated with land-use modification. In particular, agricultural practices have exposed dryland wetland ecosystems to prolonged and drastic disturbance in a variety of global contexts (e.g. Davis and Froend 1999; Finlayson and Rea 1999; Fluet-Chouinard et al. 2023) These disturbances include broadscale land clearing, cultivation (replacing native vegetation with homogenised crops), trampling, and selective grazing by livestock (Dorrough et al. 2004; Horner et al. 2012), nutrient influxes (e.g. from fertilisers, pesticides) (Blumenthal 2006) and the introduction of invasive plant species (either deliberately for fodder or by zoochorous propagule dispersal) (Dawson et al. 2017).

To persist through the challenges imposed by climatic and localised disturbance regimes, dryland ecosystems require resilience (Colloff and Baldwin 2010). For dryland wetlands, this resilience includes the capacity to transition between wet and dry phases, where vegetation in the landscape is expressed differently in each phase. Some plants, such as Duma florenta (Meisn. T.M.Schust: lignum), which is common in swampy wetlands across arid Australia, exhibit physical dormancy in its vegetative state, allowing the species to survive for extended periods of time under harsh conditions (Freestone et al. 2017). Plants without physical-dormancy strategies may instead store seed within the landscape, commonly within soil, the canopy or leaf litter (Zivec et al. 2021). Seeds can remain dormant until germination is triggered by environmental cues, which are commonly associated with hydrological input (i.e. flooding, rainfall) and can result in broadscale regeneration events (Capon 2007; Capon and Reid 2016).

Projected changes to the global climate will alter the timing and duration of wet and dry phases in semiarid regions by increasing temperatures, decreasing precipitation and increasing evapotranspiration (CSIRO and The Bureau of Meteorology 2015; Sandi et al. 2020a). Understanding the resilience of dryland wetlands in the face of these changes, including the potential shifts in vegetation expression, is important for managing and maintaining these systems into the future. Here, we (1) explore the observed and projected impacts of climate change to dryland wetlands, with a focus on the northern Murray–Darling Basin, (2) present some plausible trajectories of vegetation change for this region, and (3) discuss the role of adaptation and management for these wetlands into the future.

Case study: northern Murray–Darling Basin, eastern Australia

The northern Murray–Darling Basin (nMDB) encompasses over 500,000 km² of eastern Australia (Department of Climate Change, Energy, the Environment and Water 2022). Aboriginal peoples have inhabited the nMDB for tens of thousands of years prior to European invasion, and through this time have forged deep and ongoing spiritual connections with the natural environment (Davies et al. 2021). Traditionally, in dryland floodplain systems, the boom-and-bust nature of water within landscapes is associated with many aspects of Aboriginal peoples’ culture, including storytelling, diet, exchange of resources and interactions among nations (Davies et al. 2021; Moggridge and Thompson 2021). After European settlement, the nMDB has become an important economic asset, with a substantial allocation of agricultural land, especially for cropping (both dryland and irrigated) and livestock grazing (Gell et al. 2019). A diverse range of natural ecosystem types, including dryland wetlands, forests, woodlands, and grasslands, is also found in the nMDB (Kirsch et al. 2022). The nMDB dryland wetlands include sites of global biodiversity significance, e.g. the Narran Lakes, Gwydir wetlands and Macquarie Marshes Ramsar sites (Ley 2003), and are home to many endemic, threatened and iconic plant species, such as the river red gum (Eucalyptus camaldulensis) (Fu et al. 2015).

The waterways of dryland wetlands in the nMDB are commonly fringed by E. camaldulensis canopy, with associated mid-storey species such as Acacia stenophylla, Eremophila spp., and a varied herbaceous groundcover. Further from waterways, and onto floodplains, vegetation communities are typically characterised by Eucalyptus coolabah, along with many species of Chenopodiaceae comprising shrub and groundcover communities. Many plant species are associated with both landscape positions, such as Duma florenta (Murray et al. 2019). The composition of dryland wetland vegetation communities is dependent on interplay between hydrological variability (i.e. flooding or rainfall) and site-based dynamics such as resource availability, resident biota and local climatic variability (Casanova and Brock 2000; Webb et al. 2012; Wen et al. 2018). For example, plant species that have evolved in dryland wetlands are often capable of forming persistent seed banks that remain dormant until hydrological inputs generate conditions suitable for germination and establishment (Brock 2011). The longevity of these germinants may then depend on retained soil moisture and other local climatic factors (Rodriguez-Iturbe et al. 2007). Vegetation phenology, often used as an indicator for vegetation function or condition, also has strong links to flow regimes in drylands wetlands of the nMDB (Mac Nally et al. 2011; Webb et al. 2012; Broich et al. 2014; Thapa et al. 2016).

Climate-change impacts in the nMDB

Historically, water availability (rainfall and runoff) in the nMDB has been highly variable, characterised by dry periods interspersed with intense rainfall (Callaghan 2019). Rainfall in the northern basin has an association with tropical weather patterns, particularly landfalling cyclones that can result in large summer floods and La Niña events such as the extreme wet seasons of 2010–2012 and 2020–2022 (Speer et al. 2022). Climate change has altered these patterns and made the seasonality of flooding events less predictable (Callaghan 2019). The average temperature across the MDB has increased by 1°C since 1910, which has contributed to the drying of the landscape (Whetton and Chiew 2021). In general, the traditionally cooler months have become much drier, and autumn rainfall has decreased markedly (Whetton and Chiew 2021). Such is the extent of drying; the ecological benefits of rainfall are sometimes nullified because of enhanced plant evapotranspiration associated with high temperatures (Speer et al. 2022).

Significantly, the nMDB has begun experiencing extreme weather events in association with climate change, notably the Millennium Drought (1997–2009), which is considered to be the worst drought on record for south-eastern Australia (Whetton and Chiew 2021). The impact of the Millennium Drought was realised in the dryland wetlands of the nMDB, including Ramsar-listed sites, which during this period were subject to declines in vegetation quality without adequate flows (Catelotti et al. 2015; Jiao et al. 2020). In recent decades, some of the lowest ever water-runoff recordings have been documented in the nMDB (Potter et al. 2010; Whetton and Chiew 2021). Periods of drought and extremely low rainfall have driven keystone plant species (e.g. E. camaldulensis and E. coolabah) towards dieback and into states of poor health (Doody et al. 2015). Understorey vegetation can also be negatively affected under extremely hot, dry conditions; however, these assemblages are more resilient with the capacity to respond quickly to hydrological input, largely because of life-history strategies of vegetative growth and seed dormancy (Capon and Reid 2016; Zivec et al. 2021). In the Macquarie Marshes, a contraction of inundation area and reduced duration, in conjunction with a drying landscape has led to widespread die-back of mature trees and a degradation in several wetland types (Catelotti et al. 2015; Sandi et al. 2020b). Conversely, Saintilan et al. (2021) uncovered a pattern of woody encroachment into grasslands (i.e. an increase in recruitment), and suggested that this trend is likely to be exacerbated under climate-change scenarios, which alter current regimes of wet–dry phases. In the Gwydir River system (including the Ramsar-listed wetlands), an array of vegetation communities, including shrublands, marsh club-rush sedgelands, terrestrial, and aquatic flora communities have all faced condition declines in response to diminished hydrological input (Kingsford 2000). In other instances, especially wet or dry conditions have facilitated plant invasions (e.g. Eichhornia crassipes and Salix spp.) (Taylor and Ganf 2005; Nolan et al. 2021). Conversely, the frequency of extreme flooding events has also increased in parts of the nMDB since the late 19th Century, particularly in regions closer to the coast (e.g. 2010 was the wettest year on record for much of the MDB) (Power and Callaghan 2016). Knowledge about the impact of extreme or unseasonal flooding on the dryland wetlands of the nMDB is limited, with much of the existing literature being focused on the positive impacts that periodic flooding can have for plant recruitment and survival of native herbaceous cover and canopy species (Capon 2007; Catelotti et al. 2015). There is also growing concern around the role of flooding in the dispersal of non-native plant species into dryland wetlands, particularly considering the potential impacts of invading species, such as Parthenium hysterophorus, which is toxic to cattle (Howard and Harley 1997; Mao et al. 2019), or Phyla sp. (lippia), which is an extremely invasive herbaceous groundcover species (Price et al. 2011).

Future climate change in the nMDB

The ongoing changes to the climate projected for the nMDB are likely to manifest as increased temperatures and more consecutive hot days, changes to rainfall patterns with increased summer and autumn rainfall, but decreased winter rain, and more frequent or severe extreme weather events (AdaptNSW, see https://www.climatechange.environment.nsw.gov.au/projections-map). Murray–Darling Basin Authority has adopted three potential scenarios for climate change in the Basin, which are the outcome of scenario modelling that considers paleoclimate, historical records, and future climate models, and which underpin future water planning (Zhang L et al. 2020). These scenarios can be summarised as (1) warmer and wetter, (2) warmer and drier, and (3) warmer and drier with more severe droughts (Zhang L et al. 2020) (Table 1).

Vegetation change under a warmer and drier climate

Wet-phase vegetation communities can persist through extended dry periods through seeds stored in the soil seed bank, which can also include dry phase and terrestrial species (Capon and Reid 2016). This strategy will enable taxa to persist through periods of climate change-induced drought and recover when wet conditions return. Wetland trees could persist through dry periods, although the conditions experienced during the Millennium Drought reduced reproduction and regeneration of river red gums (E. camaldulensis) (Woods et al. 2012; Sandi et al. 2020a, 2020b; Ngugi et al. 2022) A short-term recovery response, indicated by change in remotely sensed vegetation indices, occurred as rapidly as within 1 year in the Macquarie Marshes (Sandi et al. 2020b), although modelling suggests that further deterioration could occur ~40% of the time in a 30-year period (Quijano-Baron et al. 2022). Further, the ability of forests to recover is reduced with the number, intensity, and duration of consecutive droughts (Jiao et al. 2021). The persistence of trees through drought in some areas can at least partially be attributed to their use of groundwater resources when surface water evaporates (Doody et al. 2015), as well as other adaptive responses, e.g. leaf shedding. However, this reliance on groundwater during dry times may be impeded as a result of the increasing demands of agricultural and domestic activities on groundwater extraction (Colloff et al. 2016). Further, with droughts expected to increase in severity and frequency, stands of E. camaldulensis may experience even more widespread dieback and, eventually, be replaced by more drought-tolerant species (Sandi et al. 2020a). During prolonged dry periods, non-woody wetlands (e.g. sedgelands, herb lands) may experience encroachment by both native and non-native woody wetland and terrestrial species. Understorey communities could transition from grasslands and herb lands dominated by amphibious or flood-responding species, towards communities dominated by terrestrial shrubs and grasses or towards floodplain woodlands, depending on the flood frequency (Saintilan et al. 2021). However, when floods return, the encroaching terrestrial vegetation may experience rapid mortality, making room for wetland plants to re-establish from soil seed banks while conditions remain wet. Or woody terrestrial vegetation may survive the flooding, leading to the expansion of woody stands (e.g. E. camuldulensis forests) in place of the formerly ‘wetter’ habitats. After floods, conditions could rapidly transition back into drought, with increased evaporation under warmer temperatures (Speer et al. 2022).

This rapid transition between wet and dry conditions could reduce the success of reproduction because plants do not have time to reach maturity under appropriate conditions. Under these conditions, species that can reproduce through vegetative mechanisms (e.g. lignum, Phragmites australis, lippia) may expand. The presence of dense exotic understorey vegetation in some degraded wetlands and farmland could restrict the transformation towards terrestrial communities under drier conditions, although the suitable habitat for these exotic species is also restricted and terrestrialisation could occur in the long term (Murray et al. 2012). Overall, a warmer and drier climate future could result in further degradation of dryland wetlands in the nMDB, although interspersed wet periods would allow these systems to partially recover.

Vegetation change under a warmer and drier climate with more extreme events

Although wetland vegetation in the nMDB is inherently resilient to periods of dry conditions, the severity and length of drought projected under the ‘warmer and drier with more severe drought’ scenario could reduce this resilience. In addition to the impacts discussed in the ‘warmer and drier’ scenario, fire may become a more prominent threat to wetlands. The likely increase in extreme fire risk could see wildfires extend into wetland areas that do not typically burn (e.g. red gum and lignum vegetation communities). In the Riverina bioregion within the southern MDB, Zhang and Lim (2019) identified that fires burned more frequently in areas that are inundated frequently because biomass production is greater; however, fires were less likely in areas where water is more permanent. The more severe and prolonged drying that is likely under this scenario would lead to the drying of these permanently inundated or high soil-moisture areas in the nMDB, potentially resulting in burning of areas not seen before, or at higher intensities, with likely negative outcomes for wetland vegetation.

Even though dryland wetland vegetation is inherently resilient to drought, the antecedent conditions from severe drought may challenge the capacity of ecosystems to recover. For example, Doody et al. (2015) outlined a 9-year period between flooding events, whereby E. camaldulensis was able to rely on groundwater. These authors also found that there is the potential for E. camaldulensis to survive up to 15 years in drought; however, in such a scenario productivity would be significantly reduced (Doody et al. 2015). Availability of groundwater to support the surviving trees, and broader ecosystem resilience, can be affected by extraction for agricultural purposes, extraction and contamination from mining and fracking (Doody et al. 2019), and lack of rainfall inputs (Colloff et al. 2016). Recruitment of tree seedlings is affected by drought conditions, particularly the more rapid onset of dry conditions after rain or flood event, exacerbated by increased evaporation, can significantly reduce the number of surviving seedlings despite abundant germination (Ngugi et al. 2022). Flood events that occur when tree communities are severely drought affected can drown out any remaining seedlings and adult trees if inundated for long periods (Colloff et al. 2014). Low recruitment and tree die-back have been observed for much of the northern basin over the past decade, with growing concern for these tree communities into the future (Woods et al. 2012; Ngugi et al. 2022). Conditions that result in wetland degradation of the Macquarie Marshes could occur up to 95% of the time under drier climate scenarios (Quijano-Baron et al. 2022). Furthermore, soil seed banks may become depleted if conditions, such as extreme heat and evaporation, post-inundation do not support the successful growth and reproduction of emerging plants. In general, a future with warmer, drier, and more extreme conditions would result in widely degraded wetland systems, allowing for the invasion of drought-tolerant exotics and terrestrial vegetation (Sandi et al. 2020a).

Vegetation change under a warmer and wetter climate

It is also possible that the nMDB could experience a warmer (+2°C) and wetter climate overall, compared with historical conditions. In this scenario, rainfall is likely to increase in the autumn and summer months (Zhang L et al. 2020) and may also include an increased frequency of severe wet events leading to flooding. Flooding or wetter conditions generally promote the regeneration of E. camaldulensis communities, including both recruitment and rejuvenation of the existing canopy, as has been observed in areas of the southern basin in response to environmental watering actions (Mac Nally et al. 2011). However, in the nMDB under this scenario, the extent of this regeneration or condition, improvement could be only minor in wetlands that are already drought stressed, e.g. the Macquarie Marshes; however, smaller but more frequent flows achieve better conditions than does a single large flood (Mason et al. 2022; Quijano-Baron et al. 2022). Where woodland communities are degraded, transition to alternate wetland states may occur (e.g. reed swamps or lignum shrublands) facilitated by vegetative regeneration mechanisms. Alteration to the timing and duration of inundation events can influence the composition of species that emerge from soil seed banks after the event (Capon 2007). For example, annual and perennial monocots emerged in greater abundance after summer flooding, whereas annual forbs were more abundant when soils remained waterlogged for an extended duration such as that associated with winter flooding (Capon 2007). However, frequent, or consecutive flooding events may further deplete the soil seed bank, by removing affected species from the vegetation assemblage. Wetter conditions could also prove favourable for the expansion of riparian and floodplain weeds, particularly species that have become abundant in the region because of their association with more constant altered flow regimes, e.g. Salix and Phyla (Murray et al. 2012).

Transforming wetland management in the nMDB

Ongoing and predicted effects of climate change demand urgent and transformative adaptation of current wetland policy and management because of the significant impacts and changes that can be expected at both local and landscape scales (Finlayson et al. 2017). Adaptation will require reconsideration of wetland management objectives and targets, planning approaches, implementation of management actions, and monitoring and evaluation as well as governance, to enable such transformation to occur.

Priority actions to future-proof freshwater biodiversity are well-established and globally applicable (Lynch et al. 2023), encompassing environmental flows, water-quality improvements, protection and restoration of critical habitats, managed exploitation of freshwater species and resources, management of species invasions, and protection and restoration of hydrological connectivity. Minimisation and mitigation of existing threats and pressures, such as grazing, hydrological modifications and topographic disturbance, will be particularly important in the nMDB, so as to allow wetland biodiversity, ecosystems, and landscapes to adapt autonomously to climate change (Capon et al. 2013). This may include, for example, provision of buffer zones around wetlands, as well as habitat linkages between these, in which grazing is restricted to minimise deleterious effects on vegetation and habitat quality.

Planning, implementing, and evaluating such management actions under climate change requires consideration of the risks and uncertainty involved (e.g. Horne et al. 2022). Vulnerability assessments that consider the likely exposure, sensitivity and adaptive capacity of wetland components and processes to changes in climatic stimuli, for example, are useful for identifying and prioritising adaptation actions (Capon et al. 2013). Potential for maladaptation and perverse outcomes of management actions, as well as stranding of infrastructure (e.g. environmental works and measures), also need to be considered.

The broadscale and significant changes to wetland composition and structure that are anticipated under a changing climate mean that many current wetland management objectives and targets will no longer be tenable, especially those that are backwards looking and do not account for irreversible alterations to the landscape (e.g. pre-European settlement or pre-development conditions). Restoring wetlands to past conditions or even maintaining them in current conditions may not be feasible, given the magnitude and pace of change expected. In many parts of the world, wetland managers are embracing the resist–accept–direct (RAD) framework (Lynch et al. 2021), which explicitly recognises the potential for ecological transformation and is underpinned by a conceptualisation of ecosystem processes that acknowledges possible rapid, irreversible change. Under this framework, wetland managers might shift from ‘resisting’ any ecological change to either ‘accepting’ inevitable or unmanageable changes or attempting to ‘direct’ these changes, so that key values and functions are protected.

In the nMDB, current management is largely within the ‘resist’ category, with objectives and targets often associated with maintaining species diversity or extent. However, climate-change impacts may necessitate acceptance of some ecological changes, especially where these are very technically or logistically difficult, expensive, and risky to prevent or reverse. For example, shifts in dominant plant species across wetland systems in response to hydrological change, such as the migration of vegetation zones towards waterways following reductions in flooding, may be inevitable. More dramatically, climate change may trigger some wetlands to shift in character sufficiently to become different wetland types altogether, e.g. semi-permanent wetlands that become more ephemeral in response to a drying climate. Accepting or directing such changes may enable managers to prioritise protection or adaptation of key values and minimise risk of maladaptation or failure, e.g. by managed relocation of threatened species populations.

Adaptation of wetland management, including developing ‘climate-ready’ objectives and targets, necessitates careful and equitable consideration of the diverse array of wetland functions and values across multiple scales and identification of what is being managed (Campbell et al. 2021). A transformative approach might focus management objectives on ecosystem functions and services, for example, rather than just composition and structure (Capon and Pettit 2018). Additionally, wetland management objectives must look beyond local, site-based goals, such as those typically associated with Ramsar sites, to consider landscape and greater scales (Finlayson et al. 2017). For example, regularly providing environmental water to the same wetlands or wetland types could reduce landscape-scale heterogeneity in the long term and, in turn, ecological resilience and adaptive capacity. Instead, greater attention might be given to watering a range of wetland types or states at a regional scale to promote landscape heterogeneity rather than prioritising maintenance of a selected suite of wetlands in their current states.

Crucial to any transformation of wetland management will be ensuring equitable and transparent decision-making, underpinned by reflexive and flexible institutions. Community involvement will be critical to facilitating respectful consideration of a diversity of values and perspectives concerning wetlands and promoting a shared understanding of adaptation choices. Monitoring, evaluation, and research, including citizen science, are therefore more important than ever as a basis for making evidence-based and participatory adaptation choices, particularly when all signs suggest that there are difficult decisions ahead.