From TERN Landscape Assessment Wiki
Jump to: navigation, search


The EODESM System

The EODESM system provides a basis for national monitoring of the environment through EO but gives consideration to other sources of information, including spatial layers (e.g., relating to elevation and land ownership) and knowledge. EODESM uses EVs retrieved from EO data and combines these to generate land cover and change maps and descriptions according to the Food and Agriculture Organisation (FAO) Land Cover Classification System (LCCS - Version 2). To detect and describe change between any two time-separated periods, the component codes of the LCCS classification (e.g., relating to canopy cover or water depth) and EVs (both internal and external to the FAO LCCS) are compared and evidence is accumulated to detect change according to categories within a pre-defined taxonomy (e.g., dieback, flooding).

The DPSIR Framework and Modifcations

The Drivers-Pressure-State-Impact-Response (DPSIR) is one of many conceptual frameworks used to describe and analyse changes within an ecosystem. Originally developed as the Pressure-State-Response (PSR) framework in 1995 by the European Environment Agency (EEA, 1995), the DPSIR has since been widely adopted and recommended for use by other international institutions (e.g., the European Union (EU) and the Organisation for Economic Co-operation and Development (OECD)). Since its formulation, criticisms of the framework around, for example, inconsistent definitions of the components and its limited ability to demonstrate complex cause-effect relationships (Gari et al., 2016, Patricio et al., 2016), has led to modifications of this framework.

Drivers are defined as a complex phenomenon governing ecosystem change. Those of of anthropogenic origin meet ‘basic human needs’ such as demand for food, energy and recreation and are able to be directly managed. Natural drivers are independent from those that are anthropogenic and include earthquakes, volcanic activity or tectonic drift (Oesterwind et al., 2016). Climate change is partly a natural and anthropogenic driver. Basic human needs require the occurrence of Activities to fulfil these needs. Activities related to anthropogenic drivers can be grouped in sectors, for example agriculture, transport, infrastructure, non-renewable energy or tourism. Agricultural activities include irrigation, cropping, cultivation or fertilisation, with each leading to a range of pressures that influence or dictate wetland states. Depending on the type and extent of activity, these lead to pressures that cause a change within the ecosystem. Pressures are a result of the driver-initiated activity and cause an effect on the ecosystem. They can be managed (e.g. pollution, waste production, altered water regime, vegetation clearing, increase in off-road vehicle use) or otherwise (e.g. fluctuations in climate variables). State indicators are the actual condition of an ecosystem or environment at a specific time and the change in condition at a secondary time step is referred to as the state change. States can be measure quantitatively or qualitatively for a range of physical (e.g., temperature, light), biological (e.g., species, habitat) and chemical (e.g., nitrogen) variables. Accumulation of state change indicators, both positive and negative, provide evidence for the Impacts upon the ecosystem or environment. Any management reactions to these impacts are Responses seeking to monitor, mitigate, manage, protect or adapt through land management or policy decisions from a governance background. A summary of the DPSIR framework is provided in Figure 1, where descriptions are drawn from Lucas et al. (2019a), with these modified from Oesterwind et al., (2016) and Elliott et al., (2017).


Within the DPSIR framework, the state indicators link directly to many of the EVs used within EODESM, a number of which are able to be retrieved from EO data and are used in the classification of land covers. State changes relate to differences in EVs (e.g., water hydroperiod, lifeform) between any two time-separated periods and link directly to the evidence-based change mapping and associated descriptions. The impacts on the ecosystem or environment correspond also with the pre-defined change taxonomy associated with EODESM (Lucas et al., 2019a). Lucas et al. (2019a) also highlight that different economic and natural drivers lead respectively to activities or events/processes that result in different pressures on the environment. For wetland environments, natural and anthropogenic drivers both lead to climate change pressures including sea level fluctuations and changes in precipitation, temperature and storm intensity. With each, changes in different sets of EVs relating to semi-natural/natural and cultivated/managed (primarily vegetated) terrestrial and aquatic environments and both artificial and naturally bare areas and water lead to different impacts that can be associated with a discrete and distinct set of change types. For wetlands, these might include vegetation dieback or increases in hydro-period. Such changes can also be linked with impacts on local to international policy, from which one or more management responses can be considered. On this basis, the overall framework was developed, and its implementation is demonstrated for aquatic ecosystems, namely mangroves, inland riparian zones and coastal wetlands.

Figure: Example of drivers, driver-induced activities, pressures and their consequences on state and state change for a wetland ecosystem. The lower rows show impacts (cumulative evidence from state change). Responses can be tailored to specific impacts.

Implementing a framework for monitoring aquatic environments

Drivers, activities and pressures

For aquatic environments, the main drivers are natural and include fluctuations in sea level, precipitation, temperature and storm intensity. These can also be influenced by human-induced climate change. The main economic drivers are demand for food and water and, to a lesser extent, shelter, energy and recreation. In each activity sector, pressures on the environment are linked to the requirement for land (e.g., to plant crops), transport (to move goods), access (e.g., for tourism) and infrastructure (e.g., for mining or port development). These pressures lead to different impacts.

State indicators and sources of EVs

Within the DPSIR framework, State indicators relate directly to the EVs used as input to the land cover and change classifications undertaken through the EODESM system. Australia has generated a substantive set of EVs, often through Earth Observation. Many are spatially explicit, and several are of a spatial resolution and temporal frequency that are suitable for monitoring as well as being open access. The temporal frequency is important to consider in order to quantify changes in State indicators (EVs) over time. The main sources of relevant and usable data are provided by TERN Landscapes, DEA Wetlands and Riparian and CSIRO Soil and Landscape Grid, noting that focus has been on those that are available at a national level.

Impacts - Environment For aquatic ecosystems, a wide range of impacts from natural events and processes as well as human activities are evident, with these occurring over multiple spatial scales and temporal frequencies. Lucas et al., (2019a) proposed the formation of a globally applicable change taxonomy, which can be mapped and described by comparing evidence accumulated from changes across a range of EVs. For aquatic ecosystems, this taxonomy includes elements resulting from natural and human-induced pressures on the landscape, with these being relevant to vegetation, hydrology, infrastructure development and use, and agricultural activity.

Figure: Change taxonomy relevant to aquatic environments, resulting from natural and human-induced pressures.

Impacts - Policy For Australia, a wide range of international, federal and state/territory policies are impacted upon by the changes outlined in Table 6. Depending on the focus of the policy, impacts could be positive, negative or both. This is a useful distinction if policy makers are seeking to assess, review and evaluate the implications of current policy upon ecosystems within their jurisdiction. For example, regeneration of a wetland in an area that had been cleared would have positive implications for our international agreements.

Figure:Examples of main policies relevant to wetlands.

Responses Where environmental changes (or changes in a specific ecosystem) impact on policy and/or land management relevant to aquatic environments, a range of responses are needed, with these including adjustments to policy, different land management practices and/or societal or economic changes in processes and practices. Such responses will differ between local, state and territory, federal and international governing bodies and other organisations/management units. Five response options are proposed by Lucas et al. (2019a), with these being to monitor, adapt, manage, mitigate or protect. These responses are indicative and open to discussion, and can accommodate local circumstances or perspectives.


The proposed framework generates assessments of environmental change (impacts) based on time-series comparison of EVs retrieved largely from EO data, and land cover and change classes generated from these. Therefore, the accuracy of land cover and change classifications is dependent upon the reliability in the retrieval of the EVs used to generate the output class. As outlined in Lucas et al. (2019a), the EarthTrack mobile app has been developed specifically to support the validation of land cover and change maps currently being generated through DEA, with this mirroring the LCCS classification from EO data but at the ground level. For example, within vegetation categories, EarthTrack allows separate validation of the individual canopy cover, canopy height, leaf type, phenology and lifeform components for the class “tall (> 15 m) closed (70-100 %) broad-leaved evergreen trees” and hence the overall class, noting that this is built entirely from environmental variables. All data relevant to the LCCS classes and changes are made freely available at a global level within minutes of being submitted, thereby creating a continual stream of ground data.

EarthTrack also provides significant capability for developing algorithms for retrieving specific EVs from EO data, as ground-based measures such as LAI, AGB, water chemistry and soil and vegetation moisture can be collected using scientific equipment and techniques and submitted to a central repository. Of note is that EarthTrack also allows collection of plot-based measurements, including the size and species type of individual plants. EarthTrack Mangroves has been developed specifically to record land cover and change within mangrove environments and considers all mangroves species currently present in Australia. The Habitat Assessment Condition Tool (HCAT) has also been developed to support HCAS (Harwood et al., 2016).


A web-based visualisation tool developed to convey the framework and flow of information (Lucas et al., 2019a) was adapted for aquatic environments to highlight the main drivers of change and the resulting pressures on aquatic ecosystems. The main climate-related pressures were associated with fluctuations in ambient conditions (e.g., rainfall, temperature, sea level and storm intensity) but also human demands for basic resources (e.g., food, water, energy). For each pressure, changes in states indicators/EVs (associated with vegetation, water and bare materials) are represented according to whether these are positive, negative, both or neither of these. For example, a positive change is associated with an increase in biophysical units (e.g., canopy cover in % or annual hydroperiod in days) or changes in the rankings of specific FAO LCCS component codes (e.g., 1-3 for leaf type changes corresponding to broadleaved (1), needle-leaved (2) or aphyllous (3)). Pre-defined environmental impacts on aquatic environments (e.g., vegetation dieback, flooding) are then indicated by unique combinations of EVs and their direction of change. Each impact can then be linked to different policies relevant to both the broader or more specific environments (e.g., mangroves) and responses (e.g., monitor, mitigate, manage, adapt and/or protect) are also indicated. This is undertaken such that any changes detected can alert or inform specific policy or land management plans/strategies that these might impact (or otherwise) on, and support decision on the best courses of action.

Kakadu National Park Within Kakadu NP, natural and human-influenced climatic fluctuation is leading to changes in sea level with demonstrated impacts on mangroves (Lucas et al., 2018). With potential to adversely affect the integrity of the expansive freshwater wetlands and their associated vegetation (Bayliss et al., 1997). As illustrated in Figure a, rises in sea level and associated saltwater intrusion will most likely lead to an increase in water depth, salinity, movement and hydroperiod. Which will be manifested in a progressive decrease in vegetative canopy cover and AGB ultimately loss of all describable components of vegetation (i.e., lifeform, leaf type, phenology, height and plant species). These collectively provide accumulated evidence of a loss of vegetation at the expense of increased water in the landscape. This will ultimately impact on policies (Figure b), including the Convention on Biological Diversity and the Ramsar Convention as well as the Park’s status as a World Heritage Site and Key Biodiversity Area. Other policies affected would include state and territory Acts relating to vegetation management, biodiversity conservation, fisheries and national parks. In the example of Figure c, an essential response would be to more closely monitor changes in sea level and vegetation within Kakadu NP, particularly given its cultural and natural significance but also to suggest or implement management strategies such as reinforcement of levees.

Figure:. Visualisation of vegetation loss in Kakadu given the ‘pressure’ sea level change (SLC); a) state change indicators (+ve = green, -ve = red and yellow represents both) are accumulated to provide evidence for a pre-defined impact, b) associated policy impacts and c) the type of response of governing bodies.

Lower Leichhardt River Catchment At the mouth of the Leichhardt River and along the coastal margin, sea level fluctuations and change in both temperature and precipitation have resulted in progressive changes in mangrove extent and condition. Between 1992 and 2002, increasing sea level, including those resulting from the pressures of climate change, have resulted in regrowth of mangroves which has been exacerbated by intense storms bringing flushes of sediment down the river (Asbridge et al., 2016). The EVs that accumulate as evidence for regrowth are changes in lifeform (from trees to shrubs) and increases in canopy cover, canopy height and AGB (Figure a). An increase in elevation may also indicate seaward encroachment of mangroves rather than regrowth resulting from rises in sea level and hence the changes in the extent of the intertidal zone. Increases in all water-related EVs allows the regrowth to be linked back to sea level rise and an associated increase in salinity. Mangrove regrowth has positive impacts on the relevant international policies due to the regeneration of the ecosystem (Figure b). It could have a positive or negative impact on state and territory policies as species composition could be altered. Most policy responses from regrowth would simply involve monitoring of the mangrove succession and expansion (Figure c). States and territories may look to protect the regrowth from potential degradation, thereby aligning with their individual conservation acts.

Figure:. Visualisation of dieback in Leichhardt River given the ‘pressure’ sea level change (SLC); a) state change indicators (+ve = green, -ve = red and yellow represents both) are accumulated to provide evidence for a pre-defined impact, b) associated policy impacts and c) the type of response of governing bodies.

However, a mass dieback of mangroves occurred in 2015/16 following a substantive and sustained reduction in sea level and drier and hotter conditions (Duke et al., 2017). Hydrological changes associated with this event include a decrease in water depth, movement and hydroperiod, although an increase in sediment loads and salinity can be expected (Figure a). The loss of both canopy cover and above ground biomass but no reduction in height provides evidence of defoliation or dieback. Mangrove dieback would be confirmed by the loss of moisture content from the woody components of trees, which could be obtained from L-band Synthetic Aperture Radar (SAR) data. Although the trees affected have no leaves, lifeform, phenology and species remains the same. Over the following months or years, trees will be lost, and the ecosystem change will transition from a LCCS Level 4 change (e.g., in canopy cover) to a more drastic Level 3 change from naturally bare or water category. The mangrove dieback event in the Gulf of Carpentaria negatively impacted on a number of international policies including the UN’s SDGs (particularly in relation to land degradation), CBD and FCCC and most domestic policies (Figure b). With large-scale events that are dictated by changes in climate that are largely beyond control, the main response is to monitor (Figure b, c). This has been achieved, with the framework for national mangrove monitoring put in place using optical satellite sensor data and through DEA (Lymburner et al., 2019). Additional research capacity has been provided to understand the past history of changes and baselines of mangrove extent have been established against which to compare future changes (e.g., the TERN LIDAR capture of the Gulf of Carpentaria and efforts in Kakadu NP).

Figure:. Visualisation of dieback in Leichhardt River given the ‘pressure’ sea level change (SLC); a) state change indicators (+ve = green, -ve = red and yellow represents both) are accumulated to provide evidence for a pre-defined impact, b) associated policy impacts and c) the type of response of governing bodies.

Diamantina Catchment The majority of inland wetlands located within the Diamantina are used under a pastoral lease. Hence the main drivers of change are anthropogenic demands for food (economic). An additional pressure is competition by introduced non-indigenous species and particularly weeds. Climatic pressures include fluctuations in precipitation and temperature, within the local area but also in the upstream reaches of the drainage system. Pre-defined changes include vegetation defoliation, dieback, loss, growth or replacement (by invasive plants), erosion, sedimentation (including siltation), soil and water salinisation and over- or under-grazing, which all impact on the riparian zone. An example scenario is the onset of drought conditions, with this associated with a reduction in the depth and movement of water and annual hydroperiod. Sediment loads in river might decrease, increase or remain stable. The lack of water would reduce canopy cover within the riparian zone vegetation and, to a lesser extent, AGB noting that much of the vegetation along the Diamantina channel network is comparatively productive because of water logging. Accumulating these lines of evidence would lead to a conclusion of vegetation defoliation but not dieback, as the vegetation height would remain similar (Figure a). The opposite effect might be expected when wetter conditions returned. The Diamantina catchment contains several national parks and DIWA sites, which are protected under the Environment Protection and Biodiversity Conservation (EPBC) Act, and KBAs, although is not a World Heritage Site. All would be negatively affected by drought conditions, albeit to varying degrees (Figure b). The majority of policy responses would be towards monitoring in the short and long-term (e.g., within the framework of DEA) and both intra and inter annually, particularly given the environment is highly dynamic (Figure c).

Figure:. Visualisation of defoliation in Diamantina Catchment given the ‘pressure’ precipitation (PC); a) state change indicators (+ve = green, -ve = red and yellow represents both) are accumulated to provide evidence for a pre-defined impact, b) associated policy impacts and c) the type of response of governing bodies.

Murray River The main drivers of change in the Murray River are primarily economic and relate to demand for food; most of the economy is built around agriculture. The main pressures are requirement for land (and hence vegetation clearing) and water (hence altered hydrological regimes), invasion of non-native species (particularly weeds) and pollution. The construction of dams, irrigation, extraction of water and diversion of environmental flows are activities that have led to unseasonal drying and sometimes wetting of wetlands in the Gwydir catchment and have also impacted upon those associated with the Barmah-Millewa forests. Reductions in environmental flows are often manifested in an increase in salinity but also a reduction in water movement and depth as well as annual hydroperiod. The long-term impacts on vegetation includes a loss of vegetation height, particularly if more mature trees are affected, and an associated reduction in canopy cover and AGB (Figure a). These changes will adversely affect a range of policies relating to water usage, fisheries, biodiversity and carbon (Figure b) across NSW/Victoria state and federal policies. There are also implications for the majority of international policies. Whilst monitoring of environmental flows is essential (Figure c), policy and land management responses to drivers wold include mitigation, management (e.g., of fisheries and vegetation) and adaptation (e.g., planting of non-native species to reduce salinity).

Figure:. Visualisation of dieback in Murray River given the ‘pressure’ altered water regime (AWR); a) state change indicators (+ve = green, -ve = red and yellow represents both) are accumulated to provide evidence for a pre-defined impact, b) associated policy impacts and c) the type of response of governing bodies.

Land cover change

Using EODESM and EVs obtained from DEA and TERN, indicative classifications of land cover were generated for the Leichhardt River, Diamantina River and Gwydir (Lucas et al. (2019b) according to the FAO LCCS Level 3 and/or 4 taxonomy (Figure). New classifications were generated for Kakadu NP and the Murray catchment focusing on the Barmah-Millewa forest.

Figure: Example maps of land cover for a) Kakadu National Park, b) the lower Leichhardt River catchment, c) the Diamantina catchment, d) the Gwydir and e) Barmah-Millewa according to the FAO LCCS taxonomy.

For all classifications, the fractional (green) cover and WOfS datasets generated through DEA were most useful as these quantified EVs relevant to the LCCS classification on at least an annual basis. For several time periods, comparisons between these datasets were used to highlight the link between fractional green cover and hydroperiod for the four study areas. All other EVs were generated for one point in time (including ITEM) and hence were not able to be used for change detection or description.

For Kakadu NP, and using freshwater palustrine wetlands (Figure a) as an example, time-series comparison of WOfS and fractional (green) cover between 2015 and 2016 indicated a higher frequency of inundation and a corresponding reduction in fraction (green) cover across most of the wetland. This could be attributed to excessive water in the landscape leading to a less favourable environment for rooted vegetation growth. However, the area affected adjoins a saline creek and hence an alternative conclusion is that saltwater intrusion has occurred, with this leading to expansion of mangroves in the downstream reaches but reductions in vegetation associated with freshwater environments. With continued saltwater intrusion, mangroves might be expected to expand into this area.

A typical time-series of Landsat-derived Normalised Difference Vegetation Index (NDVI) data from DEA for a mangrove area in the Gulf of Carpentaria (Figure 10 b and c) highlighted an increase from around 1992 to 2015, after which a rapid decrease was observed. These time-series are therefore useful in highlighting when environmental changes take place and can direct where to compare EVs as well as land cover maps derived from these through EODESM. When the comparison was made in the between years of local minima (1992) and maximum (2002), a decrease in WOfS annual hydroperiod was observed because of gradual desiccation of land inland of mangroves as these progressively expanded seaward onto accreting sediment banks (arising from large storm events). However, the green vegetation fraction increased as the mangroves expanded seaward and matured. However, between 2014 and 2016, the reduction of hydroperiod was associated with a drop in sea level which led to loss of mangrove canopy cover associated through dieback. The majority of mangroves affected were dominated by Avicennia marina but Rhizophora species occurring towards the central margin were also affected. The reduction in hydro-period was also evident across most of Kakadu NP and was linked also to the reduction in canopy cover along the inland margins of mangroves dominated by Avicennia marina. As with those in the Gulf of Carpentaria, these mangroves experienced dieback in 2015 (Lucas et al., 2018)

Figure: Changes in water hydro-period (red = -ve, blue = +ve) and fractional (green) cover (red = -ve, green = +ve). a) underlying Google Earth imagery for freshwater (palustrine) wetlands, Kakadu NP, b) Kakadu water hydro-period, c) Kakadu fractional cover, d) mangroves, Leichhardt River, 2015-2016, e) Leichhardt River hydro-period 2015-2016, f) Leichhardt River fractional cover 2015-2016, g) mangroves, Leichhardt River, 1992-2015, e) Leichhardt River hydro-period 1992-2015, f) Leichhardt River fractional cover 1992-2015.

In the Diamantina catchment (Figure a), the relationship between annual hydroperiod and the amount of vegetation cover was less clear, with this attributed in part to the limited ability to detect waterlogged areas within the WOfS product. Nevertheless, longer term drying and an associated loss of cover was observed in some areas. The relationship between water inundation frequency and fractional green cover was variable in the Barmah-Millewa wetlands (Figure b), although areas with decreased hydro-period were generally associated with a reduction in canopy cover.

Figure: Changes in water hydro-period (red = -ve, blue = +ve) and fractional (green) cover (red = -ve, green = +ve). a) and d) underlying Google Earth imagery for inland riverine systems, Diamantina River and the Barmah-Millewa forested wetlands, respectively. b) water-hydroperiod, c) fractional cover for Diamantina River. e) water-hydroperiod, f) fractional cover for Barmah-Millewa.

Policy Relevance

The National Voluntary Review (Australian Government, 2018) of Australia Key National Policies and commitments relevant to the SDGs. In this regard, Sections 5.3.2 and 5.4 describe how a EOSDEM-DPSIR approach to a TERN portal with a focus on aquatic ecosystems can: a) provide evidence of State Change and Impacts, b) create spatio-temporal alerts, and c) assist with place-based, tailored decisions related to management and policy.

Metternicht et al (2018) extended the CEOS4SDG analysis of EO for the SDGs, to show the potential of EO-derived products from TERN Landscape for implementation of Global Indicator Framework of the SDGs. Their initial findings point to the direct and indirect contributions of TERN EO-derived products for targets and indicators related to Goals: 15,14,13,11,6,3,2,1, and 9 to a lesser extent. This report adds more evidence to the role of TERN Landscape R&D beyond developing metrics for the GIF. The EODESM-DPSIR as applied in this proof of concept provides information and tools in support of policy development (A in Figure) with other ongoing initiatives. Integrated with other platforms/information systems of government agencies like BOM, DEA, ABS it could support Australia’s SDG Reporting Platform, that envisages data integration across all domains and societal sectors (B in Figure), helping to develop targets (C in Figure) and metrics for policy implementation and validation (D in Figure), and importantly, monitoring of progress towards the SDGs and other multi-lateral environmental agreements, and the quantification of agreed indicators. These areas depend on scientific knowledge, big Earth Data (including EO) and solution-oriented R&D. The below figure shows how, ultimately, TERN Landscape underpins R&D that can contribute to the future NESP, and it also feeds into governance processes and policy and management assessments.

Figure: The role of TERN Landscape and other major national data and information platforms in support of R&D for governance processes and assessments. (A) R&D supports planning of actions and the development of policy and management measures. (B) Development of knowledge and new information is underpinned by collaborative work between stakeholders hosting environmental and socio-economic data. (C ) Data and information integration assist Australia in determining policy targets (e.g. SDG Targets), and in (D) developing metrics for indicators of progress. Modified from: Plag and Jules-Plag (2019).


Asbridge, E., Lucas, R., Ticehurst, C., Bunting, P., 2016. Mangrove response to environmental change in Australia's Gulf of Carpentaria. Ecol. Evol. 6, 3523–3539.

Bayliss, B., Brennan, K., Eliot, I., Finlayson, M., Hall, R., House, T., Pidgeon, B., Walden, D., Waterman, P., 1997. Vulnerability assessment of predicted climate change and sea-level rise in the Alligator Rivers Region, N.T., Australia. Supervising Scientist Report 123, Australian Government, Department of the Environment, Canberra. Retrieved from:

Duke, N.C., Kovacs, J.M., Griffiths, A.D., Preece, L., Hill, D.J.E., van Oosterzee, P., Mackenzie, J., Morning, H.S. & Burrows, D. 2017. Large-scale dieback of mangroves in Australia’s Gulf of Carpentaria: a severe ecosystem response, coincidental with an unusually extreme weather event. Mar. Freshwater Res., 68, 1816-1829.

Elliott, M., Burdon, D., Atkins, J.P., Borja, A., Cormier, R., de Jonge, V.N., Turner, R.K., 2017. “And DPSIR begat DAPSI(W)R(M)!” - A unifying framework for marine environmental management. Marine Poll. Bull. 118, 27-40.

EEA, 1995. Europe's Environment: the Dobris Assessment. European Environmental Agency, Copenhagen.

Gari, S.R., Newton, A. and Icely, J.D., 2015. A review of the application and evolution of the DPSIR framework with an emphasis on coastal social-ecological systems. Ocean Coast. Manag. 103, 63-77.

Harwood, T.D., Donohue, R.J., Williams, K.J., Ferrier, S., McVicar, T.R., Newell, G., White, M., 2016. Habitat Condition Assessment Scheme: A new way to assess the condition of natural habitats for terrestrial biodiversity across whole regions using remote sensing data. Methods Ecol. Evol. 7, 1050–1059.

Lucas, R.M., Finlayson, C.M., Bartolo, R., Rogers, K., Mitchell, A., Woodroffe, C.D., Asbridge, E., Ens, E., 2018. Historical perspectives on the mangroves of Kakadu National Park. Mar. Freshwater Res. 69, 1047-1063.

Lucas, R., et al 2019a. Towards a National Framework for Linking Land Cover Change with Policy.

Lymburner, L., Bunting, P., Lucas, R., Scarth, P., Alama, I., Phillips, C., Ticehurst, C., Held, A., 2019. Mapping the multi-decadal mangrove dynamics of the Australian coastline. Remote Sens. Environ.

Oesterwind, D., Rau, A., Zaiko, A., 2016. Drivers and pressures - Untangling the terms commonly used in marine science and policy. J. Environ. Manage. 181, 8-15.

Patricio, J., Elliott, M., Mazik, K., Papadopoulou, K-N., Smith, C.J., 2016. DPSIR - Two Decades of Trying to Develop a Unifying Framework for Marine Environmental Management? Front. Mar. Sci. 3.

Plag, H.-P., Jules-Plag, S.-A., 2018. A goal-based approach to the identification of essential transformation variables in support of the implementation of the 2030 agenda for sustainable development. Int. J. Digit. Earth. 1-22.