Climate Change Implications for River Restoration in Global Biodiversity Hotspots


  • Peter M. Davies

    Corresponding author
    1. Centre of Excellence in Natural Resource Management, The University of Western Australia, 1 Foreshore House, Albany 6330, Australia
    Search for more papers by this author

P. M. Davies, email


Global biodiversity hotspots contain exceptional concentrations of endemic species in areas of escalating habitat loss. However, most hotspots are geographically constrained and consequently vulnerable to climate change as there is limited ability for the movement of species to less hostile conditions. Predicted changes to rainfall and temperature will undoubtedly further impact on freshwater ecosystems in these hotspots. Southwestern Australia is a biodiversity hotspot and, as one of the first to experience significant climate change, is an example and potentially a global bellwether for issues associated with river restoration. In this hotspot, current and predicted water temperatures may exceed thermal tolerances of aquatic fauna. Gondwanic aquatic fauna, characteristic of southwestern Australia, are typically cold stenotherms and consequently intolerant of elevated temperatures. The hotspot in southwestern Australia is geographically restricted being surrounded by ocean and desert, and many important national parks are located on the extreme south coast, where the landscape is relatively flat. Consequently, fauna cannot change their distribution southwards or with altitude as a response to increasing temperatures. Therefore, any mitigation responses need to be in situ to produce a suitable biophysical envelope to enhance species' resilience. This could be through “over restoration” by increased riparian replanting at a catchment scale. A rule-of-thumb of a 10% increase in riparian cover would be required to reduce water temperatures by 1°C. These restoration techniques are considered applicable to other global biodiversity hotspots where geography constrains species' movement and the present condition is the desired restoration endpoint.


Global biodiversity hotspots contain exceptional concentrations of endemic species in regions undergoing significant habitat loss (Myers et al. 2000). Global climate change is an additional threatening process with the capacity to further impact on hotspot biodiversity. Southwestern Australia is the only biodiversity hotspot present on the Australian continent (Hopper & Gioia 2004). The freshwater fauna of the region, for temperate Australia, is comparatively rich, with 24 amphibians and 8 species of freshwater fish (Pusey et al. 1989; Morgan et al. 1998). Over 300 freshwater invertebrate species have been identified (Horwitz et al. 2008), with high levels of regional endemism of species with Gondwanan affinities (Edward 1989; Bunn & Davies 1990). Over 5000 species of vascular plants occur in the region with over half considered endemic (Horwitz et al. 2008).

Global warming has been highlighted as an escalating threat to global biodiversity hotpots (Malcolm et al. 2006) and southwestern Australia was one of the first regions to undergo significant climate change (CSIRO 2007). This has been predominantly due to reduced rainfall and subsequent runoff (IOCI 2002) and increases in air temperature (IPCC 2007a). Across Australia including the southwest, average temperatures have increased by almost 1°C since 1910 (CSIRO 2007).

Climate predictions for the southwestern region are of increasingly dry and hot condition. These are dependent on the level of global intervention (based on Special Report on Emission Scenarios [SRES]; IPCC 2000) as A2, politically weak; and B1, politically strong (Nicholls 2006; IPCC 2007b). The A2 scenario describes a global response to mitigate climate change which is only marginally successful; but not the worse “business as usual” case. The B1 scenario represents a strong global response from early and aggressive commitment to both stabilization and subsequent reduction of greenhouse gas emissions (IPCC 2007b). The current and predicted climate conditions therefore represent a threat to biodiversity to southwestern Australia which is considered particularly vulnerable on a global scale (COAG 2007). Climate change in this region and restoration options serve as an example and bellwether to underpin planning in other hotspots.


The climate of temperate southwestern Australia is Mediterranean with hot, dry summers and cool, wet winters (Seddon 1972). Rainfall and streamflow are both highly seasonal and predictable (Bunn et al. 1986), leading to aquatic communities that are highly structured and temporally concordant among years (Bunn & Davies 2000). The region has experienced a substantial decline in rainfall since 1950 (CSIRO 2007; BOM 2007) with the larger changes occurring since 1965 (Li et al. 2005). The decline in rainfall was manifest in a shift to increased summer/spring rainfall and an associated decrease during autumn/winter (Cowan & Cai 2006). This is considered to be a consequence of changes in the Southern Hemisphere circulation decreasing the likelihood of storm development over southwestern Australia (IOCI 2002) partly due to disproportionate heating of the Indian Ocean off Western Australia (CSIRO & BOM 2007).

In Australia, a reduction in rainfall typically results in a disproportionately larger fall in streamflows (Chiew 2006). At present in southwestern Australia, a 20% reduction in rainfall since the mid-twentieth century has been accompanied by almost a 50% reduction in runoff (IPCC 2007a). The predictions to 2050 (and beyond) are of increased drying and heating (CSIRO 2007) with the intensity depending on the intervention scenario (IPCC 2007b). However, the drying effect of temperature alone on river flows is often neglected; the Murray Darling Basin drought was considered by some to be largely the consequence of temperature rather than reduced rainfall due to increased evapotranspiration rates (Cai & Cowan 2008a). After consideration of interdependencies, including the effect of rainfall and clouds on minimum temperatures, Cai and Cowan (2008b) concluded that in the Murray Darling Basin, a 1°C increase in maximum temperature resulted in a almost a 15% decrease in river flow.


Anthropogenic forcing has led to global warming since the mid-twentieth century (IPCC 2007a) with mean surface temperatures increasing by 0.76°C since 1850 (IPCC 2007a). During the last 50 years, rates of warming have increased at an average of 0.13°C per decade with 11 of the past 12 years the warmest on record (IPCC 2007b). Current global temperatures are similar to the Holocene maximum and within about 1°C of the maximum temperature experienced in the past million years (Hansen et al. 2006).

Temperature increases have not been globally uniform and using a statistical downscaling of 0.8 for Australia (CSIRO 2007), averages have increased by 1–2°C since 1910 with warming greatest at night and in more inland locations. Warming of autumn/winter rather than summer/spring and increased minimums rather than maximums has led to reduced diurnal ranges (BOM 2007; CSIRO 2007). In southwestern Australia, temperatures have risen and predictions are a 0.2°C increase per decade for the next 30 years and a likely acceleration after this period under both A2 and B1 scenarios (Hennessy et al. 2007; Swan River Trust 2007).

Extreme Events

Disturbance is an important defining characteristic of stream and river communities (Lake et al. 2007). However, predicted increases in the frequency and intensity of extreme events due to climate change (IPCC 2007a) will expose river fauna and ecosystem processes to conditions probably not historically experienced potentially inhibiting recovery. In addition, following restoration, the increased frequency and intensity of extreme events including floods, droughts, and fires may limit revegetation efforts and damage associated infrastructure.

Impact on Freshwater Fauna

In southwestern Australia, flows in streams and rivers are both highly seasonal and predictable (Bunn et al. 1986). Freshwater fauna of the region show distinct responses to these predictable flow conditions with community structure fundamentally different between perennial and temporary systems (Bunn et al. 1986) and between seasons with distinctive summer/autumn and winter/spring assemblages (Bunn et al. 1986). Native freshwater fish undergo a seasonal reproductive migration tightly linked to flow with early spring “freshers,” a cue to initiate movement (Pen et al. 1991a). This has been shown for native fish including the western minnow (Pen et al. 1991b), pygmy perch (Pen & Potter 1991c), and nightfish (Pen & Potter 1990).

A reduction in flows or changes to the seasonality of flows may therefore result in the loss of migratory triggers and unseasonally low flows can expose in-stream barriers to migration (e.g. large woody material; Walker 1985). The historic flow paradigm (Poff et al. 1997) emphasizes the importance of natural (or historic) flows for the maintenance of aquatic biodiversity and ecological processes. Departures from natural flows, either through regulation and/or climate change, may therefore impact a range of aquatic species and supporting ecosystem processes. In addition, longitudinal connectivity maintained by flow is an important hydrological feature where downstream ecosystems are subsidized by carbon inputs from forested upland reaches (Davies & Bunn 1999).

Increasing water temperature can impact on freshwater systems by reducing dissolved oxygen levels and by increasing benthic respiration (Bunn et al. 1999). These two processes typically combine to drive river pools anoxic and can ultimately lead to fish kills (Bunn & Davies 1992). Dissolved oxygen at concentrations <2.0 mg/L is considered a critical threshold at which respiration becomes difficult for native fish (ANZECC/ARMCANZ 2000).

Increasing water temperature may also exceed thermal limits of freshwater fauna. Many freshwater species in southwestern Australia are Gondwanic in origin (Edward 1989; Horwitz 1997; Zwick 2000) and considered cold stenotherms (cool water species) (Bunn & Davies 1990). The thermal limits of the fauna of southwestern Australia have been determined by LD50 testing (Davies et al. 2004). A temperature of about 21°C was considered the upper limit for a range of sensitive freshwater insect taxa (e.g. Ephemeroptera, Plecoptera, and Trichoptera; Davies et al. 2004). This critical temperature was also suggested for a range of temperate species from studies elsewhere (e.g. De Kowzlowski & Bunting 1981; Quinn et al. 1994; Cox & Rutherford 2000). In southwestern Australia, this threshold temperature is often exceeded in upland streams flowing through cleared catchments (Rutherford et al. 2004) where the lack of riparian vegetation increases the irradiance into streams (Davies et al. 2002).

In small streams in southwestern Australia, dense riparian shade of 600–960 m along the streamside reduced water temperature by 4°C and cleared reaches resulted in similar levels of heating (Rutherford et al. 2004). Cooler water can be transported downstream improving conditions in more open or cleared areas. In southwestern Australia, summer temperature stress will be more exaggerated as high air temperatures co-occur with seasonal low flows.

Climate change predictions for southwestern Australia are of a 2°C increase by 2050. These temperatures may exceed historic conditions experienced by the biota and therefore outside the range of species' resilience. For example, recent bioclimatic modeling showed that many of the 92 species of Dryandra (terrestrial plants) on the extreme south coast of Western Australia are vulnerable to a temperature rise of as small as 1°C (Pouliquen-Young & Newman 2000). Increased temperature may also have other effects including influencing sex determination in species including turtles (e.g. McCarty 2001) leading to skewed sex ratios. Other temperature induced impacts include a mismatch of interactions between predator–prey, host–pathogen, and pollinators–flowering plants (McCarty 2001).

Lapse Rates

In the Northern Hemisphere, recent climate change has shifted the distribution of key biota (Root et al. 2003). This has been toward the pole for: plants (Crumpacker et al. 2001); butterflies (Parmesan et al. 1999; Crozier 2004); and birds (Hitch & Leberb 2006). Other movement has been to higher altitudes (e.g. Franco et al. 2006). Some impacts to amphibian species may be indirect thorough changes in the distribution of their pathogens (e.g. chytrid fungus; Pounds et al. 2006). Where movement has not been possible due to barriers, some species have become locally extinct (McCarty 2001).

Lapse rates are the average reduction in air temperature with increasing movement toward the poles or with increasing altitude. An environmental lapse rate with increasing altitude is about 6.5°C/km (Allaby 2004) which can vary with region. The large changes of temperature associated with altitude result in the observed rapid changes in montane communities (Grabherr et al. 1994). Toward the poles, a lapse rate of 1°C/145 km is the accepted reduction in temperature (Trapasso 2005). Consequently, if a 2°C increase in temperature in southwestern Australia is accepted a medium-term scenario, then for thermally sensitive species this is a movement of about 200 km further south or an increase in elevation of about 300 m.

Barriers to Movement

Distributional shifts of species away from unsuitable conditions can be hampered by both natural and artificial barriers. Landscape fragmentation would also limit distributional changes of species as a response to increasing temperature. In southwestern Australia, important national parks (e.g. the Stirling Range and the Fitzgerald River) are on the extreme south coast; limiting further southerly movement. Southwestern Australia is also characterized by an ancient weathered landscape with limited relief (Bettenay & Mulcahy 1972); the dominant topographic feature, the Darling Scarp, is only 250–300 m above sea level. Consequently, there is no capacity for movement further south and limited ability for movement to higher altitudes.

Some southerly movement of biota may be possible from the more northerly regions of southwestern Australia. However, as most streams and rivers in this region run east–west, there is little opportunity for wholly aquatic species (e.g. fish and crustaceans) to cross catchment divides. Aquatic fauna with a terrestrial adult phase (e.g. holometabolous insects) would have some ability to cross catchment boundaries. However, this is limited by the extremely strong and consistent southwesterly winds which characterize southwestern Australia (Masselink & Pattiartchi 2001). These winds would limit southerly movement of the generally weak-flying terrestrial insects.


The already fragmented landscape of southwestern Australia is predicted to increase under climate change (Horwitz et al. 2008). The hotter and more arid climate will also alter the distribution of native vegetation and probably current land use practices in the region. Secondary salinization which currently impacts 30% of the region (NLWRA 2001) will increase due to clearing (Bari & Ruprecht 2003) to the extent where about one-third of the freshwater fauna will be at risk (Halse et al. 2003). Existing freshwater fauna are sensitive to small changes in salinity (Edward et al. 1994) and there is often a shift in salt-impacted rivers to salt lake fauna (Bunn & Davies 1992).


Southwestern Australia is surrounded by either ocean or deserts, with little relief. Consequently in situ restoration of rivers and streams is a practical response to climate change in the short to medium term. This may be partly achieved through better reserve design (Dunlop & Brown 2008), maintenance of off-reserve biodiversity, reducing other threats (e.g. secondary salinization; Halse et al. 2003) or through restoration that increases overall resilience. Another response is increasing the connectivity in the landscape through direct corridors or habitat “stepping stones.” This has been recommended for Dryandra species in the southwest (Pouliquen-Young & Newman 2000).

River Restoration

River restoration is a practical response to climate change both on and off reserves. However, using historic references or spatially fixed protected areas (Rutherford et al. 1999) for restoration targets may fail in the context of rapid climate shift (Harris et al. 2006). Therefore, appropriate restoration needs to consider future conditions to produce a suitable “bioclimatic envelope” within species' tolerances (see Harris et al. 2006).

In southwestern Australia, a model using data on vegetation, climatic conditions, water depth, bed materials, and flow (STREAMLINE; Rutherford et al. 1997) showed shade, provided by both riparian vegetation and topography (i.e. stream banks) could maintain upland streams below the 21°C threshold of sensitive freshwater fauna (Davies et al. 2004). The modeling showed that about 50–70% shade from riparian zones was required in the more southerly (Albany) to the more northerly (Perth) respectively, under typical flow conditions, to maintain rivers and streams below the thermal tolerance limits of aquatic fauna (Davies et al. 2002, 2004; Rutherford et al. 2004). As a rule-of-thumb, about 10% of increased shade is required for every 1°C rise in temperature. Replanting riparian zones, when targeted at upland streams, also leads to improvements in more open downstream reaches (Rutherford et al. 2004). Focusing revegetation on east–west running streams is also a priority as the sun “tracks” along this path increasing irradiance (see Davies et al. 2008) and planting on streambanks as close as practicable to the channel will maximize shade benefits. As southwestern Australia is in the Southern Hemisphere, the northern bank of east–west streams should be targeted for restoration as the effects on reducing irradiance into streams will be greater than replanting the southern banks (Davies et al. 2004, 2008).

Riparian revegetation is not an isolated response to increased temperature but will have other benefits for ecosystem function which are underpinned by existing theory (e.g. Lake 2001; Lake et al. 2007). Streams and rivers of southwestern Australia are reliant on intact riparian zones for; shade to limit nuisance algal growth (Bunn & Davies 2000), the input of terrestrial carbon to support aquatic food webs (Bunn et al. 1999; Davies & Bunn 1999), bank stability (Rutherfurd et al. 2004), and for woody material which is important aquatic habitat particularly for fish (Davies & Storey 1998). Any form of vegetation would provide shade and consequently temperature benefits, however other ecological benefits of riparian vegetation, including both the nature and timing of litter input (Bunn et al. 1986), emphasizes the value of replanting using local provenance species.

While riparian restoration will have immediate and with enhanced revegetation, preadapt the rivers for an increasingly hot future, many important ecological assets in southwestern Australia may require other methods of intervention including flow restoration (Table 1).

Table 1.  Climate change trends in southwestern Australia, probability of anthropogenic forcing, potential scenarios under different levels of political intervention, possible consequences, the potential ecological response, and proposed restoration actions.
Issue and Direction of the TrendEvidence of Anthropogenic ForcingFuture Trend Under A2 and B1 ScenariosPossible ConsequenceEcological ResponseRestoration Actions
  1. Modified after Swan River Trust (2007). The “issue and direction of the trend” based on CSIRO (2007). Categories in “Evidence of anthropogenic forcing” after IPCC (2007a). The A2 and B1 scenarios after the SRES of IPCC (2000). Environmental flows are shown as e-flows, biological oxygen demand as BOD.

Decrease in autumn rainfallVery likely >90%Likely >66%Pool anoxiaFish kills, release of some metals and nutrients from river sedimentsE-flows, fencing, and riparian revegetation
   Loss of longitudinal hydrological connectivityLoss of upstream carbon subsidyRemoval of artificial barriers
Decrease in winter rainfallVery likely >90%Extremely likely >95%Loss of reproductive triggersLocal extirpation of fishE-flows
   Loss of variability encouraging invasivesIncreased populations of competitive dominantsE-flows
   Exposing barriers to fish migrationLocal extirpation of fishRemoval of artificial barriers, weed management
Decrease in winter streamflowVery likely >90%Extremely likely >95%Reduced pool scouringLoss of pool habitatE-flows, pool dredging
   Build up of organic materialIncreased BOD, fish kills (during summer)E-flows
   Reduced riparian inundation and subsequent seed-setLonger-term degradation of riparian vegetationE-flows, fencing, and riparian revegetation
Escalating atmospheric warmingVirtually certain >99%Virtually certain >99%Exceedance thermal limits of sensitive faunaLocal extirpation of faunaFencing and riparian revegetation
   Increased benthic respirationIncreased BOD, anoxiaE-flows, fencing, and riparian revegetation
   Decreased dissolved oxygen levelsFish killsE-flows, fencing, and riparian revegetation
Increased evapotranspiration ratesMore likely than not >50%More likely than not >50%Reduced streamflow; shift to a temporary systemLoss of permanent river faunaE-flows
   Loss of hydrological connectivityLoss of upstream carbon subsidyE-flows
Increased fire hazardMore likely than not >50%Extremely likely >95%Loss of riparian vegetation leading to increased water temperatureLoss of sensitive speciesFencing and riparian revegetation
   Exceedance of thermal limitsLoss of sensitive speciesFencing and riparian revegetation
   Liberation of phosphorus; increased likelihood of algal bloomsDecreased water qualityFencing and riparian revegetation, fire management
Decrease in the intensity of winter flood flowsExtremely likely >95%Likely >66%Reduced pool scouringLoss of pool habitatE-flows, pool dredging
   Increased pool aggradation  
   Loss of longitudinal hydrological connectivityLoss upstream carbon subsidyE-flows
   Weed encroachment into channelsLoss of habitat, reduced high flow conveyanceWeed control
   Increased invasives (e.g. Typha, Gambusia)Increased competitive dominantsE-flows, weed control
Increased frequency of droughtVery likely >90%Extremely likely >95%Shift to temporary systemsLoss permanent river faunaE-flows
   Loss of refugia (pool water quality)Local extirpation of faunaFencing (pools) and riparian revegetation, e-flows
Increased sea and estuary levelsVirtually certain >99%Virtually certain >99%Loss of floodplain habitatReduced fish recruitmentReserve design, removal of artificial barriers
Increased frequency of warm spells/ heat wavesLikely >66%Extremely likely >95%Pool anoxiaLocal loss of pool faunaFencing and riparian revegetation, e-flows
   Exceedance thermal limitsLocal extirpation of sensitive speciesFencing and riparian revegetation, e-flows

Flow Restoration

Many important processes in rivers are regulated by the magnitude, frequency, and seasonality of flows (Poff et al. 1997). Restoration responses including environmental flows (e-flows) can maintain important ecosystem processes and biodiversity (Table 1). Changes to the natural flow regime may impact on cues for fish migration, seasonality of insect emergence, floodplain inundation, and riparian recruitment. E-flows, such as reservoir releases, could be used to reestablish important components of the historic flow regime. However, in a drying environment, it has proved difficult to secure e-flows and the distribution of dams and reservoirs in global hotspots is typically limited and usually restricted to the main stem of rivers minimizing benefits to tributaries.

Translation to Other Hotspots

Southwestern Australia was the first global biodiversity hotspot to experience significant climate change and consequently serves as a bellwether and an example of potential restoration responses. Most biodiversity hotspots will experience global warming (see IPCC 2007a) and all are geographically restricted. For example, many hotspots are islands or island chains (i.e. Madagascar-Indian Ocean Islands, New Zealand, the Caribbean, New Caledonia, Japan, and Polynesia–Micronesia). Other hotspots of Gondwanic origin (e.g. Western Ghats/Sri Lanka and the Cape Floristic Province) share common origins with southwestern Australia and consequently potentially the temperature tolerances of the cold stenotherm fauna.

In summary, in situ “over restoration” is a practical response for climate change to establish a short–medium-term biophysical envelope in biodiversity hotspots where the desired ecological endpoint is the present condition.

Implications for Practice

  • Climate change in the southwestern Australian biodiversity hotspot may serve as a bellwether for other global hotspots.
  • Due to increased drying and warming, restoration of rivers and streams in some global hotspots will require interventions including targeted riparian replanting and environmental flows.
  • Riparian replanting will increase shade; decreasing water temperatures to levels below thermal limits of sensitive aquatic fauna.
  • Improvements in water temperature may be “transported” downstream to lower, more open reaches.
  • “Over restoration” (through riparian revegetation) using local provenance species is recommended for in situ maintenance of aquatic biodiversity and ecosystem processes.
  • Restoration of east–west upland streams, particularly the north bank (in the Southern Hemisphere), is considered a priority.
  • In southwestern Australia, a 10% increase in riparian revegetation is required for a 1°C decrease in water temperature. This is considered transferable to other global hotspots, particularly those of similar Gondwanan origin.


This research was supported by the Australian Water Research Facility (AWRF) of AusAID and the Tropical Knowledge and Coastal Knowledge (TRaCK) program.