Climate Change Facilitates the Formation of Natural Barriers in Low-Inflow Estuaries, Altering Environmental Conditions and Faunal Assemblages

Abstract

Climate change in Mediterranean regions is projected to cause declines in rainfall and higher temperatures and evaporation, which will enhance the formation of barriers at the mouth of low-inflow estuaries and potentially also in the riverine reaches. This review uses data from estuaries in south-western Australia across a rainfall gradient to describe how these barriers form and the effects they have on environmental conditions and biotic communities. The formation of barriers disconnects the estuary from adjacent freshwater and marine environments, prohibiting the movements of fauna and lowering taxonomic and functional diversity. Moreover, the longer periods of bar closure can result in increased frequency and magnitude of hypersalinity, hypoxia and nutrient enrichment. These conditions, in turn, act as stressors, often synergistically, on the floral and faunal communities. In some cases, mass mortality events occur, and some estuaries dry completely. To ensure the functioning of such systems in the future, regular monitoring across a wide range of estuaries is needed to understand how climate change is impacting different types of estuaries. A range of management options are discussed that may help mitigate the effects of increased barrier formation but should be employed as part of a whole-of-catchment approach and regularly evaluated.

1. Introduction

Estuaries are typically regarded as transitional bodies of water where freshwater derived from land drainage meets saline water from the ocean [1,2]. The highly dynamic environments in these ecosystems are influenced by three main control factors, i.e., the extent of rivers, tides and waves, with local climate and geology acting as secondary factors [3,4]. There are many classifications of estuaries based on different criteria, such as geomorphology [5,6], water circulation [7] and tidal range [8]. The scheme developed by Dalrymple et al. [3] used the relative influence of the three main control factors and recognised two distinct (but intergradational) types of estuaries, i.e., wave- and tide-dominated. Wave-dominated estuaries occur in areas where tidal influence is small and the mouth of the estuary experiences relatively high wave energy. These waves, together with tidal currents, transport sediment alongshore and onshore into the mouth, forming a sand bar, which prevents some wave energy from entering the estuary. This type of estuary is similar to barrier, bar-built and blind estuaries [5,6]. At the seaward end, relatively coarse marine sand forms flood and ebb tidal deltas, in the central basin fine sediments accumulate and, in the upper reaches, catchment-derived sediment is deposited forming a bayhead (fluvial) delta (Figure 1) [3,9].

During times when riverine inflow (including groundwater) is less dominant than wave and tidal energy, the rate of wave-driven sediment deposition exceeds the rate of removal by outflowing currents, leading to the formation of a sand bar (berm) across the mouth of the estuary, temporarily disconnecting it from the marine environment [10,11]. Such estuaries are known by various terms, including intermittently closed/open lakes and lagoons (ICOLLs), intermittently open/closed estuaries, temporarily open/closed estuaries (TOCEs), seasonally-open estuaries, bar-built estuaries and blind estuaries [6,12,13]. McSweeney et al. [14] estimated that 1477 of the 53,618 estuaries globally (91% coverage) are closed to the ocean periodically. Subsequent work by Khojasteh et al. [15] using satellite imagery with greater spatial and temporal coverage increased this number to at least 2245 (4.2% of estuaries globally) and found that the catchments of these estuaries supported 55 million people. These estuaries receive low and/or episodic/seasonal inflow of freshwater, i.e., low-inflow estuaries [16] and commonly occur along microtidal and mesotidal coasts [14]. Countries with the highest number of intermittent estuaries were Australia, South Africa, Mexico, mainland USA and Madagascar [15]. Moreover, most estuaries in south-western Australia and South Africa are disconnected from the ocean for periods [13,17,18]. Closure of the estuary can lead to marked changes in environmental conditions and alter faunal composition [9,19].

Despite the well-recognised value and ecosystem goods and services that estuaries provide [20] and the prevalence of intermittent estuaries globally, studies on these estuaries represent only 0.5% of the peer-reviewed scientific literature on estuaries [15]. Furthermore, even when including all types of estuaries, only 1% of climate change research focuses on estuarine environments [21]. Due to the role of freshwater inflow in determining the extent of connectivity between intermittent estuaries and the ocean, they are particularly vulnerable to climate change [22]. Estuaries in south-western Australia are relatively well-studied compared to other regions of the world [15] and this region has a pronounced rainfall and freshwater discharge gradient, which results in estuaries having different extent of connectivity with the ocean (Figure 2) [23,24]. This facilitates the use of a space-for-time approach, i.e., the sampling of various sites that have differing physical and biological characteristics in the same period of time, to estimate the longer-term impacts of climate change and offers lessons for slower warming regions [25]. As such, this review aims to describe how climate change may influence the connectivity of south-western Australian estuaries to adjacent freshwater and marine ecosystems through the formation of natural barriers and the potential effects of these barriers on the fauna. A range of potential management strategies is discussed, which are broadly applicable to low-inflow estuaries and/or those with an intermittent connection to the ocean.

2. Study Area

Mediterranean climatic regions such as south-western Australia, California, the western USA, the Mediterranean basin and parts of South Africa are characterised by hot/warm dry summers and cool wet winters [26,27,28]. Such regions are recognised as climate change hotspots due to their responsiveness to climatic changes [29]. For example, they are experiencing warming at rates higher than other regions globally, annual precipitation is projected to decline substantially and more frequent and intense extreme weather events (e.g., droughts and heatwaves) are also expected to occur [30,31].

South-western Australia experiences highly seasonal rainfall, with >70% occurring between May and October [25]. This coincides with the time of year that the bars of most estuaries would break [24,32]. Rainfall during this time of year comes primarily from the cold fronts and low-pressure systems that cross the region from the west, while summer rainfall is linked to the west coast trough, thunderstorms and deteriorating cyclones moving down from the tropics [24,33]. Mean annual rainfall changes along a gradient ranging from ~1400 mm per year in the south-western corner and becomes lower and more variable north and particularly eastwards to a minimum of ~400 mm [34]. In the western part of this region, annual rainfall has decreased by 16% since the 1970s, which is among the greatest of any region in Australia and has led to a reduction in freshwater flow of >50% [35,36,37]. Climate models predict that winter rainfall will decline in the future due to a southward shift in winter storms [25,38]. Moreover, recent modelling results have suggested that few, if any, very wet years will occur between 2023 and 2100, even if strong cuts to global emissions are made [33]. Generalised changes in summer rainfall are less clear; however, very large fluctuations in summer rainfall intensity are predicted due to less frequent but more intense storms [25].

Mean air temperatures in south-western Australia increased by 1.1 °C between 1910 and 2013 and are predicted to increase further, accompanied by an increase in days with hot (>35 °C) and extreme (>40 °C) temperatures [25]. There has also been an increase of 0.02 °C y⁻¹ in sea surface temperatures since the 1950s and multiple marine heatwaves, e.g., 2010/11, 2016 and 2024/25 [39]. Evaporation, i.e., 1676, 1362 and 1754 mm in Perth, Albany and Esperance, respectively, generally exceeds average annual rainfall, i.e., 723, 922 and 619 mm in Perth, Albany and Esperance, respectively, between 1991 and 2020, and is also greater than the depth of many estuaries in the region [24,40]. Trends in pan evaporation in the region are not linear. Values initially decreased from the 1970s to the 1990s due to decreasing wind speeds, but have since increased due to greater solar radiation and vapour pressure deficit [41]. Evaporation during the long, hot summer/autumn period leads to increases in salinity and water temperature, both of which lower the saturation of oxygen in water [32].

Tides in the regions are mainly diurnal and microtidal with mean and maximum daily ranges of 0.4 m and 1.1 m, respectively [24]. Sea level is rising (1.4 mm y⁻¹ at Fremantle between 1966 and 2009) and the magnitude of this is expected to increase, together with the frequency of extreme sea level and storm surge events [25,42].

In terms of local biota, south-western Australia is 1 of 36 biodiversity hotspots [43]. It has a relatively depauperate but highly endemic freshwater fish fauna [44] and a relatively rich coastal fauna [45]. The richness of the estuarine-associated fish fauna is relatively low and distinct compared to other areas of the Indian Ocean, particularly temperate South Africa, and dominated by atherinids, gobiids, mugllids and sparids [9,46]. The benthic macroinvertebrate fauna present in south-western Australian estuaries mainly comprise annelids (e.g., spionid, capitellid and nereidid polychaetes), crustaceans (e.g., amphipods) and bivalve and gastropod molluscs [47].

3. Types of Barriers

3.1. Sand Bar at the Mouth of the Estuary

There are several types of natural sedimentary barriers that can occur in wave-dominated estuaries, particularly those that have an intermittent connection to the ocean. At the downstream end of the estuary a sand bar may be present. The sand bar is formed by the longshore and onshore movement of sediment that accumulates during times of low riverine flow and can completely disconnect the estuary from the ocean [48,49]. Estuaries in south-western Australia are broadly categorised as (i) permanently-open, i.e., bars never (or rarely) close, (ii) seasonally-open, i.e., the bars open and close annually (typically bars open with river flow in winter and close in summer), (iii) normally-closed, i.e., bars remain closed for years at a time and (iv) permanently-closed, i.e., bars never (or rarely) open (Figure 3) [24]. Systems in the last category are not regarded as estuaries, due to the lack of a regular connection to the ocean [50]. Note that some authors also include intermittently-open estuaries, which open and close multiple times a year [9].

The volume of riverine inflow is the major factor in maintaining a connection between the estuary and the ocean. For example, annual mean discharge of ~14 m³ s⁻¹ for permanently-open estuaries, ~2 m³ s⁻¹ for seasonally-open estuaries and only 0.10 and 0.13 m³ s⁻¹ for normally-closed and permanently-closed estuaries, respectively [9,34]. Similarly, in California, the Carmel Estuary was open on 99% of the days when riverine flow was >0.5 m³ s⁻¹ [10] and in South Africa flow of 2 m³ s⁻¹ was needed on the lower-energy south-western Cape coast and 5–10 m³ s⁻¹ on the high-energy KwaZulu-Natal coast [53]. As most rainfall in south-western Australia occurs during winter, the bars of estuaries most frequently breach during this time of year [32,54]. However, climate modelling predicts that this rainfall is likely to decline in the future [38], which will decrease the frequency with which breaches occur. Thus, some estuaries that used to breach annually may have periods of several years where no breaching occurs and therefore function more like normally-closed estuaries. Likewise, some normally-closed estuaries may become permanently-closed. Systems such as the currently named “Lake Nameless” and “Jerdacuttup Lakes” were once estuaries but became closed in the Holocene marine transgression ~3500 years before present [24]. The sand bar at the former mouth of Jerdacuttup is 12 m above sea levels and 60–120 m wide at its narrowest point and salinities can range from 10 to >200 ppt [55,56]. One estuary, Culham Inlet, has undergone the reverse trend, initially being classified as permanently-closed and having not breached between 1918 and 1993 [57]. However, land clearing in the catchment has increased run-off and also the frequency of breaching, making it more similar to normally-closed estuaries such as the Hamersley and Stokes inlets [32].

The combination of protracted bar closure and high evaporation leads to the evapoconcentration and occurrence of hypersalinity, i.e., salinities > 40 ppt [58]. A systematic literature review of hypersaline estuaries globally identified south-western Australia as a hotspot [58]. Moreover, estuaries in this region are the most hypersaline globally with salinities of 313 and 345 ppt, recorded in the Culham and Hamserly inlets, respectively [32,59]. The extreme salinities recorded in many of these estuaries is, in part, influenced by saline riverine input primarily due to secondary salinisation [60,61]. For example, based on the estimated average annual flow of selected rivers, 44% of the volume was fresh (<0.5 ppt), 10% marginal (0.5–1 ppt), 21% brackish (1–2 ppt), 25% moderately to highly saline (2–35 ppt) [62]. Moreover, between 1960 and 2002, rivers in south-western Australia were estimated to have discharged 4700 GL of water and exported about 7.5 million tonnes of salt annually.

Not all estuaries in south-western Australia become disconnected from the ocean. Some have been modified to remain permanently-open. Examples of this include the Swan-Canning Estuary, where a rock bar at the mouth of the estuary was blasted and dredged in the late 1800s and the Leschenault and Peel-Harvey estuaries [34]. In the case of the latter estuary, a second entrance channel (Dawesville Cut) was constructed to improve tidal water exchange [63]. Bandy Creek was a seasonally-open estuary until 1982, when two breakwaters, seawalls and groynes were constructed at the mouth to prevent closure and a boat harbour was dredged [64]. However, riverine flow is insufficient to remove accumulations of sand, and thus, dredging is needed. Extreme rainfall (221 mm over two days) as a result of cyclone Isobel led to the deposition of ~150,000 m³ of sediment into the harbour, which required almost 18 months of dredging to remove [65]. Due to the westward (longshore) movement of sediment on the south coast of Western Australia, permanently-open estuaries where the western side of the mouth is sheltered by a headland (e.g., Walpole-Nornalup Estuary and Jorndee Creek) are more resilient to reductions in riverine flow and bar closure [64]. Waychinicup Estuary is unique in south-western Australia being tide-dominated. The system flows through a gorge carved through a fault in the granite which reaches 22 m deep at the mouth [66].

Reductions in rainfall and riverine discharge associated with climate change are likely to increase the prevalence of barriers at the mouth; however, the extent to which this occurs will depend on sea level rise. In some areas of South Africa, sea level rise is predicted to prolong periods of open mouth conditions as it will increase the tidal prism [67]. Furthermore, in the Laguna Madre (USA/Mexico), sea level rise is expected to decrease the occurrences of elevated salinities, and previous dredging to deepen a canal in the estuary reduced the proportion of salinity measurements >50 ppt from 33 to 38% to only 2.6–2.8% [68]. Increasing sea level rise is unlikely to have a major influence on bar breaching in the short term due to the height of the sand bars. For example, the sand bars of estuaries near Albany and Esperance are between 1 and 5 m above sea level and up to several hundred metres wide [32,61], while sea level at Esperance increased by an average of 3.1 mm yr⁻¹ between 1993 and 2009 and is predicted to increase by 8–18 cm and 30–65 cm by 2030 and 2090, respectively, under the IPCC scenario RCP4.5 [69]. However, the increased frequency of extreme sea level events may alter salinity levels, including reducing the magnitude of any hypersalinity, via overtopping [25]. Such events may also result in the transport of large amounts of sediment onshore, increasing the height of the bar, increasing the resilience of the estuary to breaching and/or causing the premature closure of any breach [58].

3.2. Barriers in Riverine Reaches

In areas of estuaries where the riverine reaches flow into the basins, the flow velocity decreases, and the loss of energy leads to the deposition of coarse sediment (Figure 1). When the rate of sedimentation from riverine sources is greater than sea level rise, sediments can accumulate, forming bayhead (fluvial) deltas [70,71]. While these structures reduce the width of the channel, they typically do not prevent the exchange of water between these two regions of an estuary. However, during times when the bar at the mouth of the estuary is closed to the ocean, evaporation can lower the water level in the basin, disconnecting these regions (Figure 4).

In the riverine regions, decreases in flow velocity can also cause the deposition of sediment forming point and mid-channel bars [72]. The formation of these bars is dependent on flow. During periods of low riverine flow, the increasing width-to-depth ratio exposes higher riverbeds to the surface, giving rise to new river bars at lower elevations [73]. Due to the highly seasonal rainfall in south-western Australia, many of the rivers and streams are non-perennial. Such temporary rivers go through three phases based on hydrological conditions, i.e., flowing, isolated pools and dry pools, which represent lotic, lentic and terrestrial habitats, respectively [74]. Isolated pools form in many of the rivers in south-western Australia during times of low (or no) surface riverine flow, and in those located further east along the south coast, the riverine reaches are carved through impervious bedrock (e.g., archaean granite). Pools range in size, but some can be several kilometres long, up to 5 m deep and even contain fish [75,76]. Without access to groundwater, these pools are known as “perched isolated pools” and experience the most pronounced changes in environmental conditions. For example, small pools in the Philips River (Culham Inlet) had salinities of 50–100 ppt due to the evapoconcentration of saline river flow (Figure 5) [77]. Other types of isolated pools and how they differ from perched pools have been explained in detail in a review conducted by Bonada et al. [74].

4. Effects of Barriers

4.1. Impediments to Movement

Different species utilise estuaries for different purposes and/or at different stages in their lifecycle [78,79]. While some species are obligate users, e.g., those that complete their lifecycle within the estuary (estuarine residents) or use it as a migration route (e.g., diadromous species), others are more facultative, such as marine estuarine-opportunist species, whose juveniles use these systems as their preferred nursery habitat but can also use alternative environments, e.g., marine embayments. The partial and complete formation of bars has a number of impacts on species, firstly by acting as a physical barrier preventing movement and secondly by reducing cues for immigration, e.g., the discharge of estuarine water into the ocean.

In south-western Australia, the only native diadromous fish species are the anadromous pouched lamprey (Geotria australis), semi-anadromous Perth Herring (Nematalosa vlaminghi) and semi-catadromous common galaxias (Galaxias maculatus). Lampreys have an antitropical distribution, with the incipient lethal temperature of ammocoetes of the pouched lamprey estimated to be 28 °C [80]. Declining riverine flow and increasing air temperatures may result in mass mortality events and the poleward range shift, as has occurred in the Pacific Lamprey (Entosphenus tridentatus) in the USA [81]. Pouched lamprey are able to use their oral disc to help climb vertical hard surfaces (e.g., isolated pool) and traverse land to overcome instream obstacles during their migration [82], which may limit the effect of estuarine barriers. That being said, traversing these structures does come at an energetic cost, and the upstream migration is conducted by non-trophic individuals [83]. Perth Herring are a predominantly tropical species and thus are currently only found in the estuaries in the northern parts of south-western Australia (e.g., Swan-Canning and Peel-Harvey estuaries), which are permanently-open [84,85,86]. With increasing sea surface temperatures the distribution of this species may extend polewards, in which case immigration into the estuary may be limited by sand bars. For example, downstream migration of the catadromous short-finned eel (Anguilla australis) in south-eastern Australia has been prevented by the closure of estuary bars [87]. Migrations in the common galaxias vary across with wide geographic distribution of this species [88]. For example, in New Zealand, individuals are amphidromous, but landlocked populations in south-western Australia have altered their life-history strategies to avoid a marine phase and be semi-catadromous [89].

The extent of connectivity between the estuary and the ocean may also affect facultative users of estuaries. For example, the marine estuarine-opportunist sea mullet (Mugil cephalus) spawns in marine waters and the juveniles recruit to the upper reaches of estuaries and even rivers between March and September, before returning to the ocean to spawn [90,91]. Krispyn et al. [61] found that the size structure of this species varied in estuaries with different extents of connectivity. This was due to bar closure restricting the immigration of 0+ recruits into the estuary and also the emigrations of sexually-mature individuals to the ocean to spawn. These latter individuals resorb their gonads and shift from reproductive to somatic growth [92]. Impediments to movement also influence invertebrates such as the Western king prawn (Penaeus latisulcatus). While studies have not been conducted in Western Australia, the congeneric Eastern king prawn (Penaeus plebejus) is considered recruitment-limited in some estuaries in New South Wales. Larvae of this penaeid are transported southwards by a boundary current, but recruitment into estuarine nursery areas can be inhibited by sand bars [93]. Moreover, as this species is targeted by recreational fishers, they are highly susceptible to fishing pressure [93]. Although there is limited knowledge on the lifecycle of estuarine invertebrates in south-western Australia, freshwater caridean larvae require substantial flow from the Mfolozi River in South Africa to help enter the St Lucia Estuary to be able to breed [94].

In general, the fish species found in the shallow (<1.5 m deep) and deeper (>1.5 m deep) waters of south-western Australian estuaries [9,76,95,96,97], as with many other microtidal estuaries [98,99,100], is dominated by marine species, i.e., marine estuarine-opportunists and marine stragglers (sensu [78,79]). As these types of species typically require an open connection with the ocean in order to immigrate into the estuary, decreasing extents of connectivity have a marked influence on faunal composition. For example, the number of species in shallow waters declined from a maximum of up to 71 in permanently-open estuaries (34–71) to the least in normally-closed estuaries (6–18; Figure 6a). Moreover, there is a pronounced sequential decline in the proportion of marine species, being on average 67% in permanently-open estuaries to 52 and 47% in intermittently open and seasonally-open estuaries, respectively, to only 15% in normally-closed estuaries (Figure 6b). The reverse trend was exhibited by estuarine species, which contributed >80% of the species in Culham and Stokes. Estuarine species also dominate the fauna when based on the proportion of individuals, particularly in seasonally-open and normally-closed estuaries (Figure 6c). This reflects the fact that these species are able to spawn in estuaries and that those populations are not sustained by recruitment from adjacent marine waters. The richness of the fish fauna in deeper waters, which mostly comprise marine species, is also related to the extent of connectivity with the ocean [9,101]. In a study of eight estuaries along 130 km of the south coast of Western Australia, the composition of the fish fauna in shallow and deep waters was found to be related mostly strongly to the number of days the sand bar was open and then salinity, due to increases in marine fish [102].

Although benthic macroinvertebrates are less mobile than fish, and many present in south-western Australia, e.g., species of nereids, capitellids, amphipods, tanaids and molluscs such as Hydorccocus brazieri and Arthritica semen undergo direct development (including brooding) rather than produce the planktivorous larvae [110,111,112]. There is a clear pattern in richness among estuaries of different types (Figure 7). Using an identical sampling effort along eight estuaries across 130 km of coastline, 84 and 67 species were recorded in the permanently-open Oyster Harbour and Waychinicup Estuary, respectively, between 27 and 33 in five seasonally-open estuaries (Torbay, Taylor, Normans, Cordinup and Cheyne inlets) and only 9 species in the normally-closed Beaufort Inlet [113]. The substantial variability in the number of species in the normally-closed estuaries in Figure 7 highlights the influence of bar state and salinity (see also Section 4.2). For example, 16 taxa were recorded in Beaufort Inlet when the sand bar was closed and the estuary was hypersaline, only 5 when the estuary had breached due to flushing and 25 after the sand bar had reformed but salinities were estuarine [114]. The influence of estuary type has been found in both eastern Australia [115] and South Africa [116], where greater numbers of species have been recorded in permanently-open than intermittently-open estuaries.

Movement can be used by fish and larger invertebrates as a way to prevent mortalities due to poor water quality [132], and barriers to that can result in mortality events [133,134]. For example, in three normally-closed estuaries (i.e., the Beaufort, Hamersley and Culham inlets), rock bars in the riverine reaches of the estuary (Figure 5) prevented fish from moving further upstream to refuges where salinities were lower, resulting in mass mortality events. Such events can be substantial, for example, ~1.3 million individuals of the long-lived and estuarine-resident sparid black bream (Acanthopagrus butcheri) died in the Phillips River (Culham Inlet) when salinities reached 83–85 ppt [135]. Larger invertebrates such as crayfish can be highly sensitive to poor water quality conditions (i.e., high salinities, low dissolved oxygen and/or high temperatures) and are known to move over land to nearby water bodies where conditions are more suitable [136].

4.2. Hypersalinity

South-western Australia is a global hotspot for hypersaline estuaries, due to the low inflow, prolonged sand bar closure, high rates of evaporation and, for some systems, also saline riverine inflow [58,61,62]. The extent to which salinity increases is dependent on several factors, including the size of the catchment and the amount of precipitation it receives, the volume of riverine inflow, depth and size of the estuary and dimensions of the sand bar [16,32,54]. While salinities in permanently-open estuaries are typically <40 ppt, areas of the Peel-Harvey and Leschenault estuaries do become hypersaline during summer and autumn [109,137,138]. Seasonally-open estuaries can also become moderately hypersaline, e.g., 54 ppt in Cheyne Inlet during [61], but the most severe hypersalinity occurs in normally-closed estuaries. Salinities of ~120 ppt have been recorded in Wellstead Estuary and Beaufort Inlet, 180 ppt in Gordon Inlet, 240 ppt in St Mary Inlet, 313 ppt in Culham Inlet and 345 ppt in Hamersley Inlet, all typically after years of bar closure and low rainfall [32,59,77,114,139].

Elevated salinities can result in the mortality of habitat-forming macrophytes such as the seagrass. Among the common seagrasses present in south-western Australian estuaries, Ruppia megacarpa and Ruppia polycarpa seeds failed to germinate at salinities above 30 and 80 ppt, respectively [140,141], and Halophila ovalis has a salinity tolerance of 9–52 ppt [142]. Thus, extreme hypersalinity will lead to the loss of macrophytes and potentially the transiton to a phytoplankton-dominated system, as occurred in Beaufort Inlet during a protracted period of hypersalinity (~60–120 ppt) [102,114].

The fauna present in estuaries have evolved physiological and lifecycle adaptations to specific salinity ranges. When exposed to salinities outside their tolerance range, it can disrupt homeostasis, leading to physiological stress, reduced fitness, altered behaviour and even mortality [143,144]. Hypersalinity has been documented to alter faunal communities by reducing the number of species, reducing taxonomic and functional diversity, simplifying community composition and telescoping the food chain [145,146,147,148]. Moreover, the loss of macrophytes and shifts in microbial and invertebrate communities alter nutrient and redox dynamics, exacerbating eutrophication [149,150].

Although few fish species globally can survive in salinities > 50 ppt [151], several of the endemic atherinids are highly euryhaline. For example, the western hardyhead Leptatherina wallacei and silver fish Leptatherina presbyteroides can osmoregulate in salinities of at least 85 and 71 ppt, respectively [152], and the elongate hardyhead Atherinosoma elongatum has been recorded in a salinity of 147 ppt [76]. Furthermore, the estuarine-resident southern longfin goby Favonigobius lateralis and bluespot goby Pseudogobius olorum have also been found in salinities of 119 and 114 ppt [76], and the catadromous common galaxias is relatively euryhaline with an upper lethal dose to 50% mortality of 45 ppt after direct transfer and 62 ppt after gradual acclimation [153]. Among marine species, Mugil cephalus were recorded in Beaufort Inlet in salinities > 100 ppt for four months, which was thought to be facilitated by their detritivorous diet as the abundance of benthic macroinvertebrates prey also declined in the high salinities [61]. However, this species and all others (except A. elongatum) eventually died. Increasing salinities in normally-closed estuaries have led to reduced numbers of species and shifts in faunal composition, and, in some cases, also mass mortalities [59,101,135,139].

Using a comprehensive dataset of 257 benthic macroinvertebrate taxa from 12 estuaries in south-western Australia with salinities ranging from ~0 to 122 ppt [154,155] demonstrated how aspects of the diversity and composition changed with increasing hypersalinity (Figure 8). Almost two-thirds of taxa recorded were euryhaline (i.e., found in a salinity range ≥ 10 ppt), with five being classified as holohaline. However, compared to models of species richness vs. salinity (e.g., the Remane diagram sensu [156,157,158]), the loss of species in hypersaline conditions were more pronounced. In salinities between 0 and 49 ppt the faunal community was relatively diverse with numerous species of polychaetes, bivalves, crustaceans, gastropods and insects (mainly larvae) [155]. However, as salinity increased further, polychaetes, insects and gastropods dominated until 110 ppt, where only insect larvae remained. In salinities > 60 ppt, the fauna resembled that of a saline wetland or salt lake rather than an estuary [114].

4.3. Drying

Estuaries are ephemeral ecosystems on geological timescales and naturally act as sediment traps (Figure 1) [24]. However, clearing of native vegetation in the catchments has exacerbated the infilling process, reducing the depth of the estuary. This is particularly prevalent when estuaries are closed and thus suspended sediment is retained in the estuary rather than exported. Storm events lead to additional transport of sediment into estuaries, e.g., the estimated 100,000 and 150,000 m³ of sediment that was deposited in Stokes Inlet in 1982 [159] and Bandy Creek in 2007 [65], respectively. The predicted increase in the intensity of storms (albeit occurring less frequently) may result in sedimentation events and would decrease the depths, reducing the volume of the estuary and resulting in parts of estuaries (and in some cases entire estuaries) drying out (Figure 9) [77]. Shallow water depths can also increase predation risk. For example, Lane et al. [160] reported feeding frenzies of piscivorous water birds, including pelicans, egrets, herons, cormorants and gulls, on fish trapped in shallow pools in the Vasse-Wonnerup Estuary as water levels decreased during summer when the sand bar was closed, preventing the inflow of marine water.

4.4. Nutrient Levels

The extent of nutrient input into an estuary is largely governed by catchment land use, with the nutrients transported via riverine input. Restricted connectivity with the ocean due to a sand bar will increase the residence time of the water, reduce flushing and lead to greater retention of nutrients, facilitating algal blooms [9,25]. Although the frequency of unseasonal rainfall events is uncertain, the intensity of any events is likely to increase [38]. Two large summer rainfall events led to 270 GL of riverine flow containing an estimated 800 t of Nitrogen and 300 t of phosphorus entering the permanently-open Swan-Canning Estuary [161]. Subsequently, high water temperatures and solar irradiation promoted rapid phytoplankton growth of the cyanobacterium Microcystis aeruginosa with cell counts of 100,000 mL⁻¹ in the water column and scums of 1,300,000 cells mL⁻¹ in sheltered bays [161,162,163]. The bloom subsided when tidal water input increased salinities to above that which M. aeruginosa can tolerate. While this event occurred in a permanently-open estuary, the magnitude of the flow, which was five times greater than the volume of the estuary, would have breached the bar of any estuary that was closed. The breaching of the seasonally-open Wilson Inlet was shown to lead to the loss of dissolved nutrients from the estuary, but this mainly occurred when there was the outflow of estuarine water to the ocean and not when there was tidal exchange [164].

However, lower volumes of flow can transport nutrients into an estuary but are insufficient to cause a breach. In these instances, evaporation can result in an increase in nutrient concentrations [165], albeit other factors may also be influential [166]. In the seasonally-open Great Brak Estuary (South Africa), low inflow and closed mouth conditions facilitated blooms of the opportunistic macroalga Cladophora glomerata, and, due to the relatively long water residence time, also blooms of phytoplankton [167].

The effects of increased nutrients on the ecology of a south-western Australian estuary is best described by the Peel-Harvey Estuary [63]. This system became increasingly eutrophic between the 1960s and 1980s, leading to growths of macroalgae (which replaced seagrasses) and blooms of the toxic cyanobacteria Nodularia spumigena [168,169]. Lower abundances of fish were present at N. spumigena affected sites than unaffected sites, likely due to the movement away from these areas by mobile teleosts, whereas there were mass mortalities of less mobile benthic and demersal fish and crabs [170]. Cyanobacterial blooms were prevented by increasing the salinity in the estuary through the construction of a second artificial entrance channel in 1995, which tripled tidal water exchange and increased flushing to 10% of the volume of the estuary each day, increasing salinities and reducing nutrient concentrations [171,172]. This radical engineering solution had pronounced changes in the invertebrate and fish faunas [120,173]. However, due to declining rainfall and riverine inflow, climate change, rather than the artificial channel, is now the dominant driver of hydrological change in this estuary [174].

4.5. Low Dissolved Oxygen Concentrations

Intermittent estuaries are particularly susceptible to hypoxia and anoxia. Events can occur due to nutrient accumulation in prolonged periods of closure and the decomposition of material from blooms of macroalgae and phytoplankton [167]. Stratification-included hypoxia can occur due to riverine inflow from precipitation [175,176] and also marine water inflow from artificial sand bar breaches [177]. Moreover, oxygen concentrations are often inversely correlated with salinity (e.g., Hoeksema et al. [32]), as the solubility of oxygen decreases with increases in the amount of dissolved solids [178,179]. Similarly, warmer waters can hold less oxygen than cooler waters [180]. A review of climate change in estuaries in New South Wales found that water temperatures were increasing, particularly in systems that were shallow and had limited water exchange with the ocean [181]. Thus, the predicted increase in the salinity and temperature of waters in south-western Australian estuaries will reduce oxygen concentrations and increase the frequency and/or severity of hypoxic events [25].

The microtidal nature of estuaries in south-western Australia makes them particularly susceptible to hypoxia, and thus many of the invertebrates recorded are relatively tolerant of perturbations (e.g., nutrient enrichment and hypoxia) [47,182]. For example, larval culicids (mosquitoes) can respire aerially, and some larval chironomids (non-biting midges) possess respiratory pigments allowing them to tolerate hypoxic conditions [183,184]. Some fish species have adaptations, e.g., the bluespot goby and bridled goby Arenigobius bifrenatus, which occur in silty sediment in the upper reaches of estuaries [185,186], and they can hold a buccal gas bubble to engage in aquatic surface respiration to ventilate gills during times of hypoxia [187].

Several hypoxia events have been studied in south-western Australia. In the Vasse-Wonnerup Estuary, a combination of a closed sand bar, unseasonally warm temperatures and ~30 mm of rain over 48 h resulted in the stratification-induced hypoxia/anoxia and the mortality of >30,000 large-bodied fish in April 2013 [188]. Multiple hypoxic events have occurred in the lower reaches of this estuary and have been linked to the management of a tidal exclusion barrier. While studies on fish movement through this barrier [133,189] have helped to inform decision-making to help reduce the frequency and magnitude of any fish kills, this highlights the roles instream barriers can play in mass mortality events. During times when the tidal exclusion barrier is closed, the area on the upstream side, which comprises nutrient-rich fine sediments, can become hypoxic due to salinity and/or temperature stratification [176]. Sampling of the benthic macroinvertebrate community recorded a depauperate fauna with low values for a suite of taxonomic and functional measures of diversity [176].

In the Swan-Canning Estuary, two unseasonal summer rainfall events have led to hypoxia. During three months of persistent hypoxia in the upper reaches of this estuary in 2010, the number of species, total density, Simpson’s evenness index and taxonomic distinctness of the benthic macroinvertebrate fauna declined markedly, and there was a shift in community composition [190]. This was due to the absence of crustaceans, which are sensitive to hypoxia [191,192], and the densities of annelids (mainly large-bodied species) and molluscs had declined slightly. Similar trends were recorded in Chesapeake Bay (USA), where seasonal hypoxia has been linked to the decline in abundance of larger, more long-lived benthic species and the dominance by smaller, short-lived species [193]. The prevalence of these small-bodied species does limit their ability to provide ecosystem services such as nutrient processing [194]. A storm in the summer of 2017 also led to pronounced hypoxia in the Swan-Canning Estuary and was implicated in a marked decline in abundance of the Western school prawn (Metapenaeus dalli) [195].

The drying climate in south-western Australia is also likely to result in increased occurrences of bushfires. Rainfall post-fire can wash sediment, ash and debris (including nutrients, polycyclic aromatic hydrocarbons and metals) into rivers and estuaries, which can lead to rapid depletions in dissolved oxygen and result in the mass mortality of fish and crustaceans [196,197]. As with algal blooms, these impacts would be greater in closed estuaries due to reduced flushing and tidal water exchange [9]. In addition, evidence from the 2019/20 bushfires in New South Wales shows that fires can burn riparian vegetation, such as salt marshes and mangroves [198], which will alter the nutrient cycling and fauna of estuaries.

5. Management

Estuaries are regarded as naturally-stressed ecosystems due to the substantial changes in environmental conditions that occur over various temporal scales [199]. However, ref. [22] argue that estuaries in microtidal areas and, especially, those in Mediterranean climatic regions that become disconnected from the ocean are particularly vulnerable to anthropogenic stressors, including climate change. Given the value of the ecosystem goods and services that estuaries provide, management interventions are needed to ensure the functioning of estuaries in the future. Management approaches should be adaptive but recognise the social and human context and thus also involve stakeholders [200,201]. Several potential management options are discussed below.

As pointed out by Khojasteh et al. [15], there is a lack of information globally on intermittent estuaries, and even though these estuaries are relatively well researched in south-western Australia, focus has often been placed on those near population centres or located in conservation estates. Given the size of the region and sparse population, future monitoring could utilise remote sensing and data loggers. Given the influence of sedimentation [159], airborne LiDAR and satellite imaging could be used to generate bathymetric maps of the estuary [202,203] and potentially also sand bar morphology [204]. Satellite imagery has also been used to detect breaches [205] and, together with bathymetric, rainfall and/or flow data, can be used to predict how the timing and frequency of breaching may change in the future. Although most readily available data loggers only record salinity (conductivity) up to 55,000 uS/cm, and this is insufficient to capture the full extent of potential variation in normally-closed estuaries, some companies (e.g., Aquamonix) have developed sensors that can measure up to 300,000 uS/cm.

Given that bar breaching is heavily dependent on the timing and magnitude of riverine inflow, the removal of any instream barriers and reduction in water abstraction would be beneficial. Flow modification has been highlighted as a major threat to estuaries in South Africa [206] and a key component of restoration efforts [207]. One of the most common and longstanding practices in bar management is the artificial breaching of sand bars [206,208]. This is often conducted to reduce flooding risk, improve water quality, prevent fish kills, facilitate recruitment/immigration of fishery species and enhance recreational use [209,210]. However, breaching when water levels are lower than those required for a natural breach can result in poor flushing, stratification, hypoxia, fish kills and hypersalinity [167,177,211,212]. Given these potential negative implications, any artificial breaches require careful planning and monitoring, but during times of moderate hypersalinity during summer/autumn when rainfall is not forthcoming, there may be value in trailing breaches to increase the volume of water and lower salinities. Such an action is considered for the Seekoei Estuary (South Africa) when salinity reaches 45–50 ppt to avoid mass mortalities of fish and invertebrates [213]. Based on the limited available studies in southern Australia, a threshold of ~60 ppt may be appropriate [114,148,155].

During times of low inflow, extensive areas of some estuaries dry, with water remaining only in the small areas of the entrance channel and upper riverine reaches, which are scoured and thus the deepest [77]. Work in freshwater environments has highlighted the value of artificial refuge pools (e.g., fire water points) in supporting biodiversity [214] and the use of excavation to remove nutrient-rich sediment and deepen areas [215,216]. This could be trialled as a way of preserving faunal assemblages in those estuaries that have become disconnected for extended periods and may become dry.

In areas that were previously dry or extremely hypersaline, recolonisation of the fish fauna may occur relatively rapidly (i.e., hours to days) following the “removal” of any barriers due to their mobility [188]. Due to many benthic macroinvertebrates having direct development, recolonisation of defaunated sediments may be slower [114]. Transportation of sediment (and associated invertebrate fauna) from nearby areas may help repopulate and remediate sediment by reducing the concentration of toxins (e.g., sulphide and ammonium) and promoting oxic conditions [217,218]. Restoration of shellfish has been undertaken in some south-western Australian estuaries and may help to mitigate the effects of eutrophication and extended residence time caused by reduced riverine inflow [219,220,221]. The environmental tolerances of the candidate shellfish species would need to be assessed against predicted conditions in the estuaries during periods of bar closure. Aquaculture-based enhancement, i.e., the release of hatchery-reared individuals into a system, can be used to overcome recruitment bottlenecks (e.g., variable survival of eggs/larvae and bar closure preventing recruitment) and also increase populations following fish kills [93,222]. However, although some enhancement programmes for estuarine-resident species have been successful [223,224], episodic climatic events have resulted in massive population decline [195,225].

6. Conclusions

Estuaries in south-western Australia and other Mediterranean climatic regions are vulnerable to the effects of climate change. Decreasing riverine flows will lead to the formation of natural barriers in estuaries, isolating them from adjacent freshwater and marine environments and making their conditions less hospitable, e.g., more saline, eutrophic and less oxic. Due to geographic variations in rainfall, south-western Australia is an ideal location to conduct long-term studies and space-for-time comparisons to understand the impacts of climate change on estuaries and their biotic communities, and how this might influence the goods and services these ecosystems provide, and to provide advice to management to ensure their functioning in the future. Given the transitional nature of estuaries and the factors that influence them, whole-of-catchment approaches are needed that incorporate input from a diverse range of stakeholders.