Urbanisation has profound effects on environmental water fluxes and storages, which can be further compounded by climate change and management responses. This study shows how climate change influences the type and amount of recharge in urban areas with different housing densities under Perth, Western Australia. We demonstrate that consequent water conservation measures aimed to reduce irrigation may be having unintended negative consequences by increasing urban heating. Groundwater recharge was measured under Perth, Western Australia, in 1980-81 and again in 2017-18, a 38-year climate period which saw annual rainfall fall by 150 mm; months with heavy rain reduce by half; maximum annual temperatures increase by more than 1 degrees C and potential evaporation increase by 15%. The consequent reduction in direct recharge through soils due to climate change was enhanced by regulated reductions in sprinkler irrigation, whilst interception losses via tree canopies increased as rain fell as lighter showers rather than prolonged events. Indirect recharge off roofs and roads, however, was less affected and now dominates recharge to the shallow, unconfined aquifer. The primary management response to falling groundwater levels has been to reduce extraction from the aquifer. We propose an additional viable strategy would be managed aquifer recharge, using either water from main drains that remove groundwater, or secondary-treated wastewater, which is currently pumped to the ocean. Increasing, or at least maintaining, local irrigation could provide multiple positive benefits, including evaporative cooling over large urban areas to help reverse urban heating. Increased urbanisation worldwide generally increases urban runoff and generates urban heat islands. In cities, such as Perth, built over transmissive sand aquifers, however, increasing local recharge can provide valuable non-potable water and help urban cooling. Identifying and installing additional recharge systems, however, is difficult for managers as it requires innovation and additional resourcing compared with purely restricting licensed extraction.

The response of groundwater recharge to climate change needs to be understood to enable sustainable management of groundwater systems today and in the future, yet observations of recharge over long-enough time periods to reveal responses to climate trends are scarce. Here we present a meta-analysis of 60 years of recharge studies over the Gnangara Groundwater System of South-West Western Australia, covering a period of sustained drying consistent with climate change projections. The recharge process in the area is defined by a wet winter during which rain saturates a deep, highly permeable soil profile with very low water storage capacity. Measurements of recharge since the 1960s show near-linear reductions in potential recharge of 50%, in response to a 20% reduction in rainfall. For the best-represented land cover in the dataset (Banksia woodland), the reduction in potential recharge was closer to 70%. A simple analytical model suggests that reductions in the duration of winter, coupled with a decreased frequency of winter storms, were most responsible for these declines, and reveals the potential for nonlinear relationships between the recharge fraction (recharge/precipitation) and climatic variables such as mean storm frequency, mean storm depth, and the length of the winter wet season. Overall, results suggest that recharge declines in drying Mediterranean groundwater systems are likely to outstrip the declines in rainfall, and that leveraging existing observation networks worldwide to characterise recharge responses to changing climate is needed to overcome existing interpretation challenges created by inconsistent sites, methods and durations of recharge estimation.

Accurately describing the hydrology of intermittent rivers is a critical step in improving our understanding and management of freshwater ecosystems. Traditional approaches such as using gauged discharge data provide little information once flow ceases and no insight into the location, morphology, or persistence of river pools. However, multispectral images can be used to describe surface water, characterize hydrology, and provide insight into ecological functioning. A multispectral approach is highly cost-effective and well suited to remote intermittent rivers with little or no gauging infrastructure. Here, we develop an algorithm to extract hydrological attributes (i.e., pool area, length, perimeter, and mean width) from multispectral imagery (Sentinel-2) and use these attributes to create a suite of ecologically relevant hydrological metrics. We describe changes in attributes and metrics in a large lowland intermittent river as it transitions from wet to dry over a four-year period. We also describe temporal changes in attributes and metrics among five river sections with contrasting hydrological persistence and fragmentation. Our algorithm successfully identified surface water in the main channel and the adjacent floodplain, the centerline of pools, and their upstream and downstream ends. Metrics proved effective at describing seasonal patterns in hydrology; revealing how the size, complexity, and elongation of surface water features (e.g., pools) decreased as the study river transitioned from wet to dry and how fragmentation increased. Metrics also successfully differentiated the river sections with varied hydrological persistence. Ecohydrological metrics derived from multispectral imagery have the potential to provide meaningful insights into riverine morphology, resilience, and ecological functioning. Our spatial approach represents a significant advancement in the ability to characterize and manage intermittent rivers, which are increasingly threatened by water resource development and a drying climate.

Mapping surface water using remotely sensed optical imagery is a particular challenge in intermittent rivers because water contracts down to narrow linear features and isolated pools, which require accurate water detection methods and reliable image datasets. Of the many methods that use optical sensors to identify water, the Water Detect algorithm stands out as one of the best options due to its classification accuracy, open-source code, and because it does not require ancillary data. However, in the original study, the Water Detect algorithm was only tested with Sentinel-2 imagery. High-resolution and high-frequency imagery, such as Planetscope, combined with sharpening and band synthesizing techniques have the potential to improve the accuracy of surface water mapping, but their benefit to the Water Detect algorithm remains unknown. Uncertainty also exists about the extent to which different input parameters (i.e. maximum clustering and regularization) influence the accuracy of Water Detect. Practitioners seeking to map surface water in intermittent rivers need guidance on a best-practice approach to improve the accuracy of Water Detect. To meet this need, we automated an existing method for sharpening and synthesizing bands and applied it to a series of multispectral Sentinel-2 and Planetscope images. We then developed a sensitivity analysis algorithm that compared the accuracy for all possible combinations of input parameters in a given range for the water detection process - enabling optimal parameters to be identified. We applied this workflow to an 81 km stretch of the lower Fitzroy River (Western Australia) to periods when spatial water extent varied markedly, i.e. mid-wet (February), early-dry (June), and late-dry season (October), across three years with variable wet season flow. We found that the ability to accurately detect surface water using multispectral imagery was increased by using input parameters identified by the sensitivity analysis and using Visible + Near-infrared (VNIR) bands, with relatively little gained by image sharpening unless the area of interest was burnt or experienced considerable shading. Also, the regularization parameter exerted less influence on results than maximum clustering. Importantly, the accuracy of the Water Detect algorithm can vary drastically if input parameters are not calibrated to local conditions. Results also revealed that our approach was adept at detecting linear features in intermittent rivers. We recommend that practitioners using Water Detect to identify surface water undertake a workflow similar to that described here to improve the accuracy of the Water Detect algorithm. The automated routines provided by this study will significantly assist practitioners in doing so. Increasing the accuracy with which we detect and map water in intermittent rivers will improve our understanding and management of these important systems which are under increasing threat.

As the human population grows and the demand for freshwater intensifies, river systems previously overlooked for water production are increasingly being earmarked for development. Many of these rivers are intermittent or in remote locations, and most are insufficiently instrumented (i.e., few to no gauging stations) and poorly understood, even though sustainable water resource development and the effective management of these systems relies on a thorough understanding of their hydrology. Remote sensing of surface water has been posed as a viable method for describing the morphology, resilience and fragmentation of river hydrology, much like metrics generated from discharge data. The low cost and increasing frequency and quality of spectral images means that this approach has a great potential to characterise hydrology at fine scales across large areas. However, the approach remains largely untested with previous research only examining small river sections. This study aimed to characterise a hydrologically diverse 400 km reach of an intermittent lowland river using newly developed, spatially derived, ecohydrological metrics. We used multidimensional clustering to identify river sections with similar hydrology, examined their spatial arrangement along the river and compared their characteristics with discharge data from four gauging stations. Metric clustering revealed four hydrological types (zones) that spanned a continuum from highly intermittent to highly persistent. Spatially, zones were at the scale of pool-run/ riffle geomorphic units that alternated along the study area. Density maps of hydrological zones revealed that the mid-section of the study reach had higher persistence and longitudinal connectivity, a finding that aligned with a pre-existing groundwater discharge map generated from field sampling of environmental tracers. Limited gauging station coverage constrained comparisons, but available stations in Persistent or Refuge zones exhibited similar hydrograph responses. The ability to map hydrology continuously along the length of a river is a significant advance compared to a gauged approach, which can only classify hydrology at a single point (i.e., the gauge). Continuous classification increases our ability to describe spatial patterns in hydrology, which could markedly affect how a river is managed.

Many regions of the world have experienced rapid hydro-climatic changes in recent decades. Mediterranean areas have especially experienced prolonged droughts, changes to runoff, and low-flow periods. Many methods and assumptions for assessing trends in rainfall and runoff are available, however, a comprehensive framework is required that can consider variations at large temporal and spatial scales, long-term persistence (LTP), seasonal patterns, and multiple change points. An analysis framework has been developed in R and applied to a network of 107 rainfall gauges (with about 100 yrs data) and 90 streamflow-water quality gauges (30 - 50 yrs data) across South-west Western Australia, a region with a typical Mediterranean climate. Trends were estimated using the Original Mann-Kendall (OMK) family of tests including using the Mann-Kendall Test under the Scaling Hypothesis (MKLTP) and Seasonal Mann-Kendall (SMK). Single and multiple change points were calculated by the non-parametric Lanzante's test (LAT), E-Agglomerative algorithm (ECP), and Energy Divisive (EDP). Trend tests show annual rainfall has decreased by 2.6 mm/yr in the Western areas and increased by 0.9 mm/yr in the Eastern areas with corresponding change points in the 1940 s-1960 s and from 1995 to 2000 although 30 % of the annual rainfall and 100 % of the standardized precipitation index (SPI) time series were found to be under LTP. A regional flow trend analysis shows catchments are yielding much less runoff in more than 90 % of gauged stations. The runoff in perennial and ephemeral streams has shown a decline with the rate of 6 (54 %) and 2 (50 %) mm/yr, respectively. Runoff and baseflow decline occurred as step changes lagged after rainfall, and the rate of decline in baseflow was more than for runoff, suggesting that groundwater connectivity with stream inverts has fallen substantially. Stream salinities increased by 64 mg/L/yr (or 15 %) overall which was highly significant. The framework provides a structured approach to analyse changes in rainfall, runoff and stream salinity for a Mediterranean climate that is projected to become even warmer and drier.

The level of water resource development was not at a critically high level when climate change dramatically reduced runoff and recharge in South-west Western Australia. This was because of the state’s relatively low population for its size, a small irrigation industry (mainly based on self-supply groundwater) and the low level of secondary industry. Few resources were heavily over allocated despite the need to significantly de-rate water supply systems. The strong uni-directional drying signal in the south-west of WA has been an advantage in that new water supplies have been developed and/or demands reduced as conditions continued to dry, reinforcing the decisions made. Strong government leadership in adapting to a drying climate in the south-west, through accelerated and alternate water source developments has been extremely successful in mitigating dire impacts, especially in the Perth-Peel region. However, as drying and warming trends continue, there is less opportunity to develop new sources and a need to consider a wider array of sources, especially use in water resource planning. More nuanced planning requires stronger governance settings and increased engagement of stakeholders. A step change in the governance and sophistication of water planning and management could be achieved through new legislation in coming years.

Climate change has profoundly affected the hydrology of south-western Australia since at least 1975. It took over a decade before the signal could be detected from annual variability. The impacts of rainfall reductions were exacerbated by higher temperatures and a decrease in wet periods when most recharge and runoff occurred. As a rule-of-thumb, runoff and recharge reduced by 3 percent for each percent reduction in rainfall. Reductions in runoff were driven by falling groundwater levels. Stream- and dryland-salinity required levels be monitored, otherwise this driver would have gone unnoticed. Runoff into reservoirs has almost ceased as processes irreversibly changed. Using historical records to estimate future runoff had limited application because of non-stationary processes. While water resources have diminished, the threats posed by dryland salinity, stream salinity, flooding and waterlogging have decreased. While winter flood risks have dramatically reduced, summer flood risks appear to have increased. Almost all GCMs project an even drier and warmer future. Perth (population 2m) has avoided a ‘Day Zero’ by the rapid expansion of shallow- and deep-groundwater extraction, and seawater desalination. Highly treated wastewater has started to be added to augment drinking water aquifers. Recharge under tree canopies have been most reduced. This is due to greater interception losses because showers have largely replaced heavy rain, and trees using a higher proportion of rainfall. Rainfall intensities, at least for long durations, have decreased despite the fear that higher sea surface temperatures (SST) and a warmer atmosphere will result in more intense rainfall. While SSTs have started to rise, there are complications related to El Niño– Southern Oscillation, the Indian Ocean Dipole and the warm Leeuwin Current that flows down the coast of Western Australia. This current results in much higher rainfall than would be expected and may weaken if El Niño becomes stronger and/or more frequent. As well as impacting water resources and rates of land degradation, climate change has affected ecosystems and industries. Abnormally hot and dry years have resulted in the deaths of trees able to withstand harsh Mediterranean summers. Wetlands have dried and groundwater-dependent ecosystems have been lost. Cereal crops are now grown in regions that used to be severely affected by soil waterlogging. Tree plantations have become unviable due to slow wood growth and deaths. Water restriction may have exacerbated urban heat islands as outdoor areas are irrigated less often, losing evaporative cooling. Fortunately, there are opportunities for diverting stormwater and treated wastewater to urban aquifers that provide a non-potable source of water for self-supply. Government regulations and planning that have been set during the pre-1975 climate are struggling to keep pace with changes in understanding and future predictions. Restrictions tackling old problems are not being replaced with those needed for new issues. It is difficult to allocate water on a fixed volumetric basis when runoff and recharge are highly impacted. Society is also having to accept water reuse more quickly than is ideal. Lessons learned in SW Australia may be applicable to other Mediterranean climate zones.

Current water governance systems and processes are often insufficient to deal with the challenge of climate uncertainty. Adapting to climate uncertainty requires trialing out new experiments in water governance and establishing processes to learn from those experiments. Social learning is regarded as an important aspect that supports the transformation of water governance systems. Understanding where and how social learning is occurring is critical to improving adaptation outcomes of water management in an uncertain climate. Yet, there is little research asserting the scope of social learning processes in a wide variety of empirical contexts. This article examines social learning in water management in Western Australia. Two specific case studies involving the state-of-the-art experiments in surface and groundwater management are used to illustrate and to understand the institutional dynamics and barriers of social learning. The cases depict single-loop learning to a larger extent and double-loop learning to a lesser extent. The findings emphasize the importance of dealing with learning externalities, the mismatch of ecological and administrative scales, and facilitating actor networks and multilevel interaction in order to embed social learning in climate adaptation. Water governance reforms should facilitate institutional configurations that enhance social learning opportunities.

The rainfall-runoff response in arid and semi-arid regions is very different from that in humid or temperate regions due to factors including the relative difference in magnitude of rainfall and potential evaporation and the absence of sustained base flow. Despite being infrequent and of short duration, rainfall and runoff events sustain regional ecosystems through flooding and groundwater recharge. Understanding spatial and temporal hydrological patterns and their changes together with inter-relations of hydrological variables can provide insight into the current and future water resource in these regions. However, scarcity of data, small numbers of flow events and access difficulties make study of arid zone hydrology difficult. This study carried out quantitative hydrological analysis using regional-scale data from a semi-arid region in northern Western Australia and compared the findings with those from other similar regions of the world. We found that the regional runoff showed a spatial pattern that does not correlate closely with the rainfall across the region. The mean peak annual floods per unit catchment area were similar to those from many other arid and semi-arid regions of the world, despite cyclonic rainfall. Furthermore, the annual runoff showed high temporal and spatial variations with little to no carry-over from the past at an annual time scale. We found statistically significant covariance between the hydrological variables such as mean annual runoff, number of flow days, high flows and drought-like conditions. The mean annual runoff and runoff ratio however showed no significant correlation with catchment area. Despite increasing trends in annual and seasonal rainfall in recent years, the annual and seasonal flow trends were quite small and exhibited much lower rainfall elasticity of runoff than in other semi-arid regions across the world. However, part of the region had a 44% increase in rainfall during a 7-year period, which nearly doubled the runoff. Annual and seasonal runoff trends can be considerably different from corresponding rainfall trends for a variety of reasons. Runoff spatial variability patterns may not follow the rainfall spatial variability. Mean annual runoff can be uncorrelated to catchment area or mean annual rainfall in semi-arid regions. Despite prevalence of cyclonic rainfalls, mean annual peak flows in the Pilbara semi-arid study region are similar to those in most other semi-arid regions elsewhere. Key findings

A warming and drying climate in the South West of Australia since about 1975, and especially from 2000, has reduced flooding and lessened salinisation risks, but also has reduced fresh water supplies. As well as continuous trends, there have been abrupt changes in hydrological processes as groundwater levels have receded from valleys. In some cleared inland wheatbelt areas, saline groundwaters are still rising where the watertable is deep, clearing has been recent and/or the reduction in rainfall has been limited. Over time, groundwater level changes will better reflect the drying climate. The reduction in rainfall and higher potential evaporation rates has dried catchments overall, and greatly reduced runoff and major flooding, even in catchments where salinity is still expanding across valley floors. Major flooding after rare storms may occur, but now is more likely in summer than winter. It is now driven by infiltration-excess rather than saturation-excess runoff processes, depending on landscape position, and rainfall amount and duration. Episodic events, such as occurred in 2017, may increase salinity and flooding for a period despite the overall decreasing trend in such risks. In the largely cleared Zone of Rejuvenated Drainage and vegetated Darling Range, groundwater levels in cleared areas are close to reaching a new equilibrium, with the drier climate reducing salinisation risks. In the largely forested Darling Range, groundwater levels are falling below stream beds thereby substantially reducing runoff into dams in the western part of the zone where groundwaters are fresh. As a result, Perth (population two million) has transitioned from being almost entirely dependent on such runoff for its drinking water, to not having any usable runoff in some years. In the Perth Basin, groundwater levels are falling within sedimentary strata as the predominantly perennial vegetation uses a high proportion of incoming rainfall. The less intense and more intermittent rainfall is also increasing canopy interception and unsaturated-zone water losses. Cleared areas with high watertables are least affected because a reduction in recharge may be offset by less rejected runoff, lower drain flows and evaporation from vegetation tapping into groundwater. This buffering will continue until groundwater levels fall beneath drain inverts and plant rooting depths. The impact of the drying climate on groundwater levels has been masked to date by increasing recharge due to clearing and urbanisation. Streams connected to strata in the Perth and Collie basins usually gain fresher water from unconfined aquifers. With falling groundwater levels drainages are transitioning from gaining- to losing-streams, with reducing surface water flows and increasing risk of aquifer salinisation where the cross-cutting streams are saline. Climate projections indicate a continuing drying trend is likely with increased temperatures and possibly a greater proportion of annual rainfall, and therefore flood risks, in summer. Water yields in dams and aquifers will continue to decline if these projections are correct. However, risks associated with too much water (salinisation, flooding, soil waterlogging) will probably continue to abate unless the amount and/or intensity of rainfall increases in the future.

A proposal to increase wastewater infiltration at the Kwinana Wastewater Treatment Plant (KWTP) to an unconfined aquifer near a groundwater flow-through wetland (The Spectacles) in Western Australia, prompted a study of the groundwater geochemistry and water levels to identify the water types and to develop a conceptual model for groundwater flow and geochemical evolution. Within the catchment, the water table has dramatically declined (e.g up to -70 to - 90 mm/yr over three decades) due to the drying climate and increasing demand for groundwater. Annual rainfall has declined by 9 mm/decade on average or about 1.4% of the annual mean over the last three decades Wastewater infiltration began in 1975 and has increased to the current rate of 4.7 thousand cubic meters per day. Groundwater-surface water interactions were interpreted using multivariate statistical analysis of major ion and stable isotope (delta O-18, delta D) data. Rainwater infiltrating through the aquifer interacts with carbonate minerals while flowing westward to produce a Ca-HCO3 type of water (Group 1). Below the wetland, groundwater has an evaporative signature (Group 2) that has higher concentrations of TDS (> 1 g/L) and distinctively higher delta O-18 (> 2.5 per mil) than groundwater derived from wastewater infiltration (Group 3). The ratio of ions (K:Cl) was used to understand the mixing of water types and to quantify the proportions of wastewater at the site. Water samples collected in 2014 reveal a high proportion ( > 34%) of wastewater within the uppermost 30 m of aquifer and generally decreasing proportions with distance away from the source. Further investigation of The Spectacles water balance and hydrochemistry is needed to optimally manage and prevent degradation of the water quality. Proposed increases in wastewater recharge may continue to help offset the impact of the drying climate on lake levels.

The Pilbara region is one of the most important mining hubs in Australia. It is also a region characterised by an extreme climate, featuring environmental assets of national significance, and considered a valued land by indigenous people. Given the arid conditions, surface water is scarce, shows large variability, and is an unreliable source of water for drinking and industrial/mining purposes. In such conditions, groundwater has become a strategic resource in the Pilbara region. To date, however, an integrated regional characterization and conceptualization of the occurrence of groundwater resources in this region were missing. This article addresses this gap by integrating disperse knowledge, collating available data on aquifer properties, by reviewing groundwater systems (aquifer types) present in the region and identifying their potential, and proposing conceptualizations for the occurrence and functioning of the groundwater systems identified. Results show that aquifers across the Pilbara Region vary substantially and can be classified in seven main types: coastal alluvial systems, concealed channel iron deposits, inland valley-fill aquifers, karstified dolomites, sandstone aquifers (West Canning Basin), Permian/Cenozoic Paleochannels, and Fractured Rock aquifers. Coastal alluvial systems show the greatest regional potential as water sources and are currently intensively utilised. Conceptually, the main recharge processes are infiltration of precipitation associated with cyclonic events and the interaction with streamflows during summer season, whereas the main discharge mechanisms correspond to evapotranspiration from riverine and coastal vegetation, discharge into the Indian Ocean, and dewatering of iron-ore bodies to facilitate mining activities. Important gaps in the knowledge relate to aquifer connectivity and accurate quantification of recharge/discharge mechanisms.

This paper describes the development of a system for decimetre-scale monitoring of land-surface and land-cover in urban and peri-urban environments. We describe our methodology that comprises the application of highly automated processing and analysis methods to digital aerial photography. The approach described in this paper addresses a monitoring need by providing the ability to generate change information at a spatial resolution suitable for urban, peri-urban and coastal areas, where an increasing percentage of the worlds' population dwells. These areas are dynamic, with many environmental issues associated with planning, service provision, resource management and allocation, as well as monitoring regulatory compliance. We present a system based on standardised data and methods, which is able to track and communicate changes in features of interest in a way that has not been previously possible. We describe the methodology and then demonstrate its feasibility by applying it to geographic areas of planning and policy relevant size (the order of tens of thousands of square kilometres). We demonstrate the approach by applying it to the problem of urban forest assessment.