Monitoring groundwater-dependent ecosystems using synthetic aperture radar (SAR) imagery in Australia

GDES

Go back

Download the practice

People in charge of the innovative practice :

Not known

Goundwater Dependent Ecosystems are highly sensitive environments that provide several services from extreme events mitigation to recreational purposes and can be heavily affected by human activities such as groundwater abstractions or contamination. To properly identify GDEs and their relations with surrounding ecosystems is fundamental to monitor environmental impacts to achieve an effective integrated water resources management.
The present factsheet will investigate the use of Sentinel-1A satellite Synthetic Aperture Radar (SAR) imagery in the remote identification of GDEs in two Australian study areas (Castellazzy et al., 2019). The technique’s authors presented acceptable results in comparison to the previously developed Australian GDE Atlas, a continental scale GDE database based on field works and literature. Further developments are expected in this field with the latest advances in SAR-based imagery acquiring methods with increased resolution and lower assembling costs coupled with enhanced computational power. This type of information has provided good results e different applications from forestry to coastal surveillance even in regions prone to cloud formation which ultimately do not allow to observe Earth’s surface variations. Further technique fine calibration may allow for application in other regions and the high coverage continuous acquisition of SAR imagery

Responsible entity

The use of Synthetic Aperture Radar (SAR) is today relatively widespread. In the case study presented in this factsheet, the entity that supported the use of such method for Groundwater Dependent Ecosystems (GDEs) detection is the Commonwealth Scientific and Industrial Research Organisation (CSIRO), Australia’s national science agency founded in 1949. This Government agency is divided into eight research areas/business units, which include Land and Water. It is in the later business unit that groundwater related research is developed.

The Groundwater Management group seeks to study and evaluate groundwater resources and develop fit-for-purpose technologies. This multi-disciplinary team has been developing work in the following areas:

  • Adaptation of remote sensing techniques to solving groundwater management problems.
  • Data analytics making use of “big data” and machine learning.
  • Groundwater flow modelling and integration of environmental and social aspects into groundwater management solutions.
  • Assess the feasibility of groundwater resources enhancement methods such as Managed Aquifer Recharge.

Concerning GDE’s, CSIRO through the Bureau of Meteorology (Fig. 1) is building a comprehensive and reliable picture of Australia’s water resources to support policy and planning – the GDE Atlas.

Commonwealth Scientific and Industrial Research Organisation (CSIRO) (left) and Australian Bureau of Meteorology (BOM) (right) logos
Fig. 1 – Commonwealth Scientific and Industrial Research Organisation (CSIRO) (left) and Australian Bureau of Meteorology (BOM) (right) logos

It collates and manages water data and information as part of its water information role and responsibilities under the Water Act 2007. The GDE Atlas was initially developed with funding from the National Water Commission and significant support from State and Territory water agencies. The GDE Atlas is now maintained by the Bureau of Meteorology and updated with new data from State and Territory water agencies (Doody et al., 2017).

GDE Atlas was constructed by combining already identified GDEs, available literature, geospatial layers and remote sensing data making use of robust GIS technologies (Merz, 2012). It represents the most exhaustive inventory of GDEs that has been accomplished at a continental scale (Pérez Hoyos et al., 2016). This web-based application allows to visualise, analyse and download GDE information for an area of interest without specialised software. Data from this platform was used to validate the SAR based approach to define GDEs.

Detailed explanation

Groundwater-Dependent Ecosystems (GDEs) are those that require access to groundwater at some stage in their life cycle in order to maintain structure and function (Dabovic et al., 2019). Those include (1) terrestrial ecosystems that seasonally rely on groundwater, (2) aquatic and riparian ecosystems dependent on the input groundwater baseflows especially in dry seasons, (3) cave ecosystems, (4) groundwater-dependent wetlands and (5) estuarine and near-shore marine ecosystems that rely on groundwater discharge (Murray et al., 2003; Eamus and Froend, 2006).

GDEs provide many ecological and socio-economical values – biodiversity, flood mitigation, erosion prevention, fishing, forestry, agriculture, recreation and tourism.

Location of Sites and footprints of the 90 Sentinel-1 scenes composed of three spatial subsets
Fig. 2 – Location of Sites and footprints of the 90 Sentinel-1 scenes composed of three spatial subsets (SUB) (adapted from Castellazzi et al., 2019)

Successful maintenance can only be achieved by understanding the distribution of GDEs while assessing water requirements within management plans (Doody, et al., 2017). In many parts of Australia, there is an increasing pressure on groundwater resources from activities including agriculture, mining, urban and commercial developments. GDEs can be degraded by the modification of flow regimes and salinisation or pollution of groundwater as a result of these activities (Kuginis et al., 2016). GDE detection can be a challenging and lengthy process and remote detection methods can significantly speed it up.

The Synthetic Aperture Radar is a high-resolution spaceborne sensor source of information that allows earth observation in different weather conditions (Moreira, 2007). It is used in a wide range of applications from environmental monitoring to security and reconnaissance uses.

The SAR active antenna sends an EM wave to Earth and the measured response is the strength of the signal bouncing back to the satellite along the same angular plane. This produces one or two “like-polarized” images: horizontal-horizontal (HH) and/or vertical-vertical (VV) bands. Some sensors can also record the signal returned along a perpendicular angular plane, producing “cross-polarized” images: horizontal-vertical (HV) or vertical-horizontal (VH) bands. Global and automatic acquiring SAR systems like Sentinel-1, used in the presented study, are dual-polarized, acquiring simultaneously VV and VH bands (Castellazzy et al. 2019).

InSAR is formed by interfering radar signals from two spatially or temporally separated antennas. An interferogram is created by coregistering two SAR images and calculating the difference between their corresponding phase values on a pixel-by-pixel basis. The change in the interferogram is caused by five effects: (1) differences in the satellite orbits when the two SAR images were acquired, (2) landscape topography, (3) ground deformation, (4) atmospheric propagation delays, and (5) systematic and environmental noises (Lu et al., 2007).

Castellazzi et al. (2019) proposed an index derived from SAR observation data (SARGDE) for capturing vegetation reliance on groundwater during dry periods. This analysis is based on the condition that due to GDEs ability to supplement natural water needs using groundwater in water deficit periods, vegetation is expected to have a permanent canopy over longer periods of time compared with non-GDE associated vegetation. This results in the assumption that the proportions of volumetric, soil and double-bounce scattering mechanisms are expected to be relatively stable in time (Fig. 3).

Simplified SAR signal scattering mechanisms in vegetation (a) 1, canopy-only direct scattering;
Fig. 3 – Simplified SAR signal scattering mechanisms in vegetation (a) 1, canopy-only direct scattering; 2, soil-trunk or trunk-soil double-bounce scattering; and 3, soil-only scattering, (b) volumetric scattering (adapted from Castellazzi et al., 2019)

Ninety Sentinel-1A interferometric wide images were used, acquired in 2017. Each image is composed of 30 consecutive SAR acquisitions – one image every twelve days for each of the three sub-datasets in the two study areas (Fig. 2).

All information extracted from SAR imagery is structured into stacks, where every pixel of each SAR acquisition is projected to the same ground footprint, forming a 3D data matrix [Space] x [Space] x [Time], referred to as data-cube. Images were processed using SARSCAPE software. Coherence matrixes, derived by comparing the phase of two polarized bands of two subsequent acquisitions, were computed using a regular matrix grid with approximately 30 m resolution. All data cubes were analysed to generate those coherence matrixes and later normalized (mean value = 10, st. dev. = 1), easing the interpretation by spreading values over a domain in which statistics offer equal weighting as a final index.

To help explore possible varying spatial and temporal patterns in data-cubes a machine-learning technique was used to reduce the time-series into a 2D map – Generative Topographic Mapping (GTM) which considers non-linear structured datasets. Generated results are compared with the Australian GDE Atlas to compare the coherence of GDE marked areas (Fig. 4). This allowed the authors to understand that in floodplains (high possibility of GDE occurrence) seasonal patterns are less pronounced than surrounding areas, as primarily inferred, indicating vegetation persistence throughout the year. It is possible to relate GDEs with low InSAR coherence with limited seasonal changes in this parameter (stable dense vegetation with associated increased scattering due to moving branches and leaves).

The SARGDE index for GDE detection is defined by using the statistics of normalized datasets that include annual mean and standard deviation of InSAR coherence (σcc’ μcc) and the annual mean of VH intensity data-cube (σvh). It is expressed with the following equation:

Simplified SAR signal scattering mechanisms in vegetation (a) 1, canopy-only direct scattering;
In this case, the VH intensity refers to vertical-horizontal bands generated from “cross-polarized” images from signals returns along perpendicular planes. SARGDEv1 is ~90% similar to the GDE Atlas on a pixel-per-pixel comparison (Fig. 5)
Low InSAR coherence in floodplains
Fig. 4 – Low InSAR coherence in floodplains (Black contours are GDEs identified in GDE atlas, the difference in colour means the difference between coherence time-series) (adapted from Castellazzi et al., 2019)
Comparison between GDE Atlas and generated index map
Fig. 5 – Comparison between GDE Atlas and generated index map (adapted from Castellazzi et al., 2019)

Institutional setting

Water management is a fundamental issue in Australia. While legislative action was taken since the late XIX century to make streams protected by water authorities, only in the 1980s the water management actions began to consider other than those water bodies and started to take into consideration water allocation procedures coupled with environmental (and social) objectives, laying the basis for a water economy. In the 1990s, the incremental cost of water, increasing demand and arise of environmental problems push the public institutions to adopt alternative methods of management and moderate conflicts between stakeholders (Tisdell et al., 2002).

Doody et al. (2017) noted that since water reform began in 1994 the environment has been recognised as a legitimate user of water. Prioritization was given in environmental water uses and, in the case of GDEs, water allocations were to be clearly defined to maintain riparian, terrestrial, wetland, marine and subterranean ecosystems that depend on groundwater. The identification and mapping of GDEs has been pointed as a critical investment to achieve the water reform objectives.

The acquisition of water information is primal to a sustainable water planning, water trading and particularly environmental management (Doolan et al., 2016).

Geographical setting

Australia shows a high variability climate with clear evidence of recurrent droughts followed by extreme flooding events. Highly irregular rainfall and high evapotranspiration create an increased challenge in water resources management, particularly for surface water. The Murray-Darling Basin authority accounted groundwater for around 30% of the water used in the continent and being in many areas the only reliable water source (Dabovic et al., 2019).

The construction of the GDE Atlas concluded that 34% of Australia’s landscape potentially contains GDEs of which 5% are classified with a high GDE potential (Doody et al., 2017). The use of SAR for GDE detection was tested in two contrasting study sites in Australia. one on Victoria and South Australia and one on the Northern Territory (Fig. 2). Site 1 (Mount Gambier area) corresponds to a karstic aquifer system and Site 2 (Wildman-Kakadu area) corresponds to a monsoonal coastal floodplain to inland semi-arid grassland (Castellazzi et al., 2019)

Historical overview

Historical overview
The first space-borne SAR system has been launched in 1978 and has been used in a wide range of applications (Moreira, 2007). Globally acquiring SAR satellites opened the door to potential applications in GDE monitoring. Automated and global SAR acquisition began in 2014 with the launch of Sentinel-1A satellite, followed by its synchronous twin satellite Sentinel-1B in 2016 (Castellazzy et al., 2019).

Concerning GDEs, in 1994 the Council of Australian Governments (COAG) endorsed reforms to move towards a sustainable water industry that included allocations for the environment and greater environmental accountability of water resource developments. In 1996 the National Principles for the Provision of Water for Ecosystems was signed in 1996 to provide a basis for considering Ecological Water Requirements (EWR) as part of water allocation decisions by water resource managers. EWR should be based on the best available scientific information. For GDEs groundwater is be considered in terms of flow, level, pressure and quality required by an ecosystem, and are to be studied through strategic scientific research. The EWR contribute with socioeconomic evaluation and water consumption demands, to the development of Ecological Water Provisions (EWPs) within water management plans. EWPs are a management tool used to achieve ecological objectives often expressed in terms of a target to maintain, restore or rehabilitate (Richarson, et al., 2011).

Evidence of benefits from implementation

The identification of GDEs often requires site-specific information on various indicators such as plant information on water use and groundwater depth which at a regional scale is impractical and cost-prohibitive (Kuginis et al., 2016). Field-based GDEs inventories are not convenient for state-wide, regional, national, or global maps, are labour intensive, and represent one point in time. The use of remote sensing greatly improves data acquisition and is a cost-effective technique (Pérez Hoyos et al., 2016).

Referring specifically to SAR, if compared with the imagery from multispectral platforms, available since the early 1970s, SAR offers the advantage of global day-and-night sensing capability and insensitivity to cloud cover, making SAR data particularly suitable for monitoring changes in Earth’s surface (Castellazzy et al., 2019). The technique presented – SARGDE index – offers improved yearly monitoring, with high resolution, and even under clouded areas. It is also relatively user-independent, as it requires only routine SAR processing.

Such methods coupled with already existing tools (GDE Atlas) ensure that consistent data is available to provide the basis for better-informed decisions.

Replication potential in SUDOE region

The identification of GDE is particularly relevant in the southern European regions, prone to scarcity phenomena and in which these ecosystems can take a significant role in mitigating extreme events. Terrestrial GDE protection is enforced by the European Union Water Framework Directive (WFD) and member-states are required to prevent damages (Rohde et al., 2017). Monitoring networks of surface and groundwater bodies play an important role and ecosystem identification may significantly benefit from the spaceborne acquired information.

Concerning SARGDE possible implementations outside the territories in which it was tested the authors note that some additional research is needed to adapt it to non-Australian vegetation. The index lacked testing for the ability to temporally monitor GDEs impacted by groundwater pumping and to be compared with in situ monitoring networks acquired data (Castellazzi, et al, 2019).

European projects such as SAR2CUBE are looking into defining prototypes to integrate SAR data into everyday processing chains and reduce the entry-level barrier of the InSAR-derived products by providing analysis-ready data (ARD) specifically defined to achieve efficiency and flexibility. This may significantly increase the user base and within the range of SAR applications in European territories.

Future outlook

Moreira (2007) reported that, at the time of publication, increased demand for this technology drove several nations to project and launch more than 20 spaceborne SAR systems. Other authors also noted (Lu et al., 2007) that InSAR promising applications can drive to scientific breakthroughs. These advances are based on longer wavelength SAR images, fully polarized SAR sensors for better characterization of vegetation and ground features, or the lessening of InSAR atmospheric delays which increase the technology’s accuracy. It is also noted that advances in data-mining, with multi-temporal and multi-dimensional techniques, can allow, for example, mapping of time-variant ground surface deformations (natural or caused by human actions) by improving deformation measurements.

Innovation in monitoring procedures, which scientists refer to as a necessary step forward (Carvalho et al., 2019), can imply the integration of technologies such as InSAR in real-time monitoring with greater spatial cover with decreased costs. As an example, the use of satellite data for surveillance of water bodies status within the framework of the EU WFD has encouraged several projects based on European Space Agency’s Copernicus programme, looking into aiding almost real-time machine-learning supported decision making in water resources management.

Advances in the technology may help in its application. The development of a miniaturized version of SAR satellites, by both private and public initiatives, are expected to decrease the cost of the technology and open the way to SAR satellite constellations that can increase global coverage.

Key points of the innovative method

  • SAR imagery can provide a lower cost and be less time consuming if compared to field-based GDEs inventories.
  • The developed index generated good results when compared with existing GDE mapping with ~90% of similarities.
  • The technique is of rather simple use but require the use of commercial software for image manipulation.
  • As it is band-related, testing is still necessary if the index is to be applied in other regions, with a different type of vegetation.
  • Further advancements are expected with the development of more cost-effective SAR systems with increased resolution and acquiring capabilities.

Acknowledgements

This innovative practice was derived from the initial suggestion by Teresa Melo from Civil Engineering Research and Innovation for Sustainability (CERIS) – Instituto Superior Técnico (IST) – Lisbon, and later adjusted resulting from PPA and LNEC discussions.

References

Carvalho, L., Mackay, E.B., Cardoso, A.C., Baattrup-Pedersen, A., Birk, S., Blackstock, K.L., Borics, G., Borja, A., Feld, C.K., Ferreira, M.T., Globevnik, L., Grizzetti, B., Hendry, S., Hering, D., Kelly, M., Langaas, S., Meissner, K., Panagopoulos, Y., Penning, E., Rouillard, J., Sabater, S., Schmedtje, U., Spears, B.M., Venohr, M., van de Bund, W., Solheim, A.L., 2019. Protecting and restoring Europe’s waters: An analysis of the future development needs of the Water Framework Directive. Science of The Total Environment 658, 1228–1238. https://doi.org/10.1016/j.scitotenv.2018.12.255 

Castellazzi, P., Doody, T., Peeters, L., 2019. Towards monitoring groundwaterdependent ecosystems using synthetic aperture radar imagery. Hydrological Processes 33, 3239–3250. https://doi.org/10.1002/hyp.13570 

Dabovic, J., Dobbs, L., Byrne, G., Raine, A., 2019. A new approach to prioritising groundwater-dependent vegetation communities to inform groundwater management in New South Wales, Australia. Aust. J. Bot. 67, 397. https://doi.org/10.1071/BT18213 

Doody, T.M., Barron, O.V., Dowsley, K., Emelyanova, I., Fawcett, J., Overton, I.C., Pritchard, J.L., Van Dijk, A.I.J.M., Warren, G., 2017. Continental mapping of groundwater-dependent ecosystems: A methodological framework to integrate diverse data and expert opinion. Journal of Hydrology: Regional Studies 10, 61–81. https://doi.org/10.1016/j.ejrh.2017.01.003 

Doolan, J., Keary, J., Boully, L., Langford, J., Claydon, G.K., Slatyer, T., Australian Water Partnership, 2016. The Australian water reform journey: an overview of three decades of policy, management and institutional transformation. https://waterpartnership.org.au/wp-content/uploads/2016/08/AWN-Australian-Water-Reform-Journey.pdf 

Eamus, D., Froend, R., 2006. Groundwater-dependent ecosystems: the where, what and why of GDEs. Aust. J. Bot. 54, 91. https://doi.org/10.1071/BT06029 

Kuginis, L., Dabovic, J., Byrne, G., Raine, A., Hemakumara, H., 2016. Methods for the identification of high probability groundwater-dependent vegetation ecosystems. Department of Primary Industries, a Division of NSW Department of Industry, Skills and Regional Development. ISBN 978-1-74256-967-3. https://www.industry.nsw.gov.au/__data/assets/pdf_file/0010/151894/High-Probability-GDE-method-report.pdf 

Lu, Z., Kwoun, O., Rykhus, R., 2007. Interferometric Synthetic Aperture Radar (InSAR): Its Past, Present and Future. Photogrammetric Engineering and Remote Sensing. 73. https://www.researchgate.net/publication/242137872_Interferometric_Synthetic_Aperture_Radar_InSAR_Its_Past_Present_and_Future 

Merz, S.K., 2012. Atlas of Groundwater Dependent Ecosystems (GDE Atlas), Phase 2. Task 5 Report: Identifying and Mapping GDEs. Final Report. Australian Government – National Water Commission. http://www.bom.gov.au/water/groundwater/gde/GDEAtlas_final_17072012.pdf 

Moreira, A., 2007. Spaceborne Radar Technologies for Earth Remote Sensing. Proceedings of International Radar Symposium (IRS) 33-36. https://www.researchgate.net/publication/225001446_Spaceborne_Radar_Technologies_for_Earth_Remote_Sensing 

Murray, B.B.R., Zeppel, M.J.B., Hose, G.C., Eamus, D., 2003. Groundwater-dependent ecosystems in Australia: It’s more than just water for rivers. Ecol Manage Restor 4, 110–113. https://doi.org/10.1046/j.1442-8903.2003.00144.x 

Pérez Hoyos, I., Krakauer, N., Khanbilvardi, R., Armstrong, R., 2016. A Review of Advances in the Identification and Characterization of Groundwater Dependent Ecosystems Using Geospatial Technologies. Geosciences 6, 17. https://doi.org/10.3390/geosciences6020017 

Richardson, S., Irvine, E., Froend, R., Boon, P., Barber, S., Bonneville, B., 2011. Australian groundwater-dependent ecosystem toolbox part 1: assessment framework, Waterlines report, National Water Commission, Canberra. http://www.bom.gov.au/water/groundwater/gde/GDEToolbox_PartOne_Assessment-Framework.pdf 

Rohde, M.M., Froend, R., Howard, J., 2017. A Global Synthesis of Managing Groundwater Dependent Ecosystems Under Sustainable Groundwater Policy. Groundwater 55, 293–301. https://doi.org/10.1111/gwat.12511 

Tisdell, J.G., Ward, J., Grudzinski, T., Cooperative Research Centre for Catchment Hydrology, 2002. The development of water reform in Australia. CRC For Catchment Hydrology, Clayton, Vic. https://www.researchgate.net/profile/John-Ward-29/publication/29458508_The_development_of_water_reform_in_Australia/links/545b50e30cf28779a4dace4f/The-development-of-water-reform-in-Australia.pdf

INTERNET REFERENCES:

Australian Groundwater Dependent Ecosystems Atlas Info Sheet (http://www.bom.gov.au/water/about/publications/document/BOM_GDE_Atlas_info_sheet_WEB.pdf) 

SARSCAPE software description (https://www.sarmap.ch/index.php/software/sarscape/) 

SAR2CUBE project webpage (https://eo4society.esa.int/projects/sar2cube/)

Other sources:

https://spacenews.com/spacety-releases-first-sar-images/ 

Participating entities:

Commonwealth Scientific and Industrial Research Organisation (CSIRO) (https://www.csiro.au/)

Australian Bureau of Meteorology (http://www.bom.gov.au/) 

aquifer
news

Discover more on the Aquifer project news and on aquifer management

aquifer news

Description and objectives of the project

The scientific community recommends a substantial improvement in the knowledge of aquifers, the establishment of reliable monitoring networks and a greater involvement of the administration and users to achieve a sustainable management of aquifers. The main objective...

read more

Information on the project

The Llobregat Delta Water Users' Community has designed recharge basins in Molins de Rei to recharge the Baix Llobregat aquifer. View of one of the reloading basins during the test phase The Llobregat Delta Water Users' Community is one of the nine partners in the...

read more

Success stories in groundwater management

Compilation of groundwater management success stories completed. Throughout April, the 30 cases of innovative practices in groundwater management have already been selected by the clusters participating in the project: PPA, CWP and AV. The task started with the...

read more

PROPOSE AN
INNOVATIVE PRACTICE

You are in charge of an innovative practice regarding aquifer management and you want to referenced it on the Aquifer platform ?

Fulfill the form and propose it to the Aquifer partners.

THE EBOOK

Aquifer offers a range of innovative water management practices. You can download all our fact sheets here.

e-book of innovative practices

DOCUMENTATION

To go further on information related to the management of aquifers