313
canopy volumes. Open source software, including Sen2Cor and ARCSI (Bunting
2017; this book), are increasingly being provided to support refined correction of
satellite sensor data, allowing compensation for differences in topographic illumination and cloud and cloud shadow detection and removal.
Whilst vast amounts of satellite sensor data are being provided, a major issue that
is widely acknowledged by Earth observation scientists is where and how to store
and process these data. It is unfeasible that users individually download these data
to their own computing systems. For this reason, there has been an increasing move
towards the use of data cubes whereby algorithms for pre-processing and analysing
Earth observation data are submitted to centralized storage facilities, with these
allowing targeted access to and analysis of the full time-series of satellite sensor
(e.g., Landsat) data. Using such a facility, the Australian Geoscience Data Cube
(AGDC) Water Observations from Space (WOfS) has generated maps of water
inundation for Australia between 1987 and 2014 using all available Landsat sensor
data (Mueller et al. 2016). Other products that are being generated include intertidal
mudflats and mangrove dynamics over similar time frames. Data from optical sensors can also be interrogated to retrieve data profiles (e.g., of NDVI) extending back
in time and informing on both rapid and gradual changes across the Australian landscape. Similar work has been produced using the Google Earth Engine, including a
decadal survey of global tree cover extent, loss and gain (Hansen et al. 2013). The
capacity to store these data and conduct analysis has been achieved by rapid
advancements in computing storage and parallel processing technologies (e.g.,
within high performance computing environments). Access to these facilitates has
also widened with the provision of web-based services such as the Amazon Cloud
and Microsoft Analytic Platform System (APS). Whilst users are charged for these
systems and some knowledge of data processing is needed, these are providing the
capacity for a greater number of users to access and process data and generate useable products. Many of these products are also being made publicly available, which
presents the opportunity for these to be used in support of nature conservation.
One of the major advances in Earth observation that is likely to significantly
benefit nature conservation is the development of CubeSat technologies (e.g.,
PlanetLabs and Surrey Satellite Technology Ltd.), with these providing small, micro
and nano satellites (defined on the basis of their weight and size) that currently
observe the Earth’s surface on a daily basis in different wavelength combinations
(e.g., red, green, blue (RGB), red edge and/or near infrared). These data can be used
in combination with the more moderate resolution sensors (e.g., Sentinel and
Landsat) to determine, in more detail, the nature and extent of changes. As examples, these allow seasonal variations in vegetation phenology to be captured or the
patterns of inundation to be mapped at high (typically <3 m) spatial resolution.
Perhaps of greatest importance to nature conservation is that these sensors provide
early evidence of changes (e.g., deforestation, ploughing of species-rich pastures,
pollution events), giving time for enforcement and/or management teams to prevent
further damage including loss of habitats and associated species. The increased
future use of targeted (e.g., red edge) channels on these satellites will also allow
better discrimination of, for example, plant species, particularly given the capacity
Expected Advances in a Rapidly Developing Work Area
canopy volumes. Open source software, including Sen2Cor and ARCSI (Bunting
2017; this book), are increasingly being provided to support refined correction of
satellite sensor data, allowing compensation for differences in topographic illumination and cloud and cloud shadow detection and removal.
Whilst vast amounts of satellite sensor data are being provided, a major issue that
is widely acknowledged by Earth observation scientists is where and how to store
and process these data. It is unfeasible that users individually download these data
to their own computing systems. For this reason, there has been an increasing move
towards the use of data cubes whereby algorithms for pre-processing and analysing
Earth observation data are submitted to centralized storage facilities, with these
allowing targeted access to and analysis of the full time-series of satellite sensor
(e.g., Landsat) data. Using such a facility, the Australian Geoscience Data Cube
(AGDC) Water Observations from Space (WOfS) has generated maps of water
inundation for Australia between 1987 and 2014 using all available Landsat sensor
data (Mueller et al. 2016). Other products that are being generated include intertidal
mudflats and mangrove dynamics over similar time frames. Data from optical sensors can also be interrogated to retrieve data profiles (e.g., of NDVI) extending back
in time and informing on both rapid and gradual changes across the Australian landscape. Similar work has been produced using the Google Earth Engine, including a
decadal survey of global tree cover extent, loss and gain (Hansen et al. 2013). The
capacity to store these data and conduct analysis has been achieved by rapid
advancements in computing storage and parallel processing technologies (e.g.,
within high performance computing environments). Access to these facilitates has
also widened with the provision of web-based services such as the Amazon Cloud
and Microsoft Analytic Platform System (APS). Whilst users are charged for these
systems and some knowledge of data processing is needed, these are providing the
capacity for a greater number of users to access and process data and generate useable products. Many of these products are also being made publicly available, which
presents the opportunity for these to be used in support of nature conservation.
One of the major advances in Earth observation that is likely to significantly
benefit nature conservation is the development of CubeSat technologies (e.g.,
PlanetLabs and Surrey Satellite Technology Ltd.), with these providing small, micro
and nano satellites (defined on the basis of their weight and size) that currently
observe the Earth’s surface on a daily basis in different wavelength combinations
(e.g., red, green, blue (RGB), red edge and/or near infrared). These data can be used
in combination with the more moderate resolution sensors (e.g., Sentinel and
Landsat) to determine, in more detail, the nature and extent of changes. As examples, these allow seasonal variations in vegetation phenology to be captured or the
patterns of inundation to be mapped at high (typically <3 m) spatial resolution.
Perhaps of greatest importance to nature conservation is that these sensors provide
early evidence of changes (e.g., deforestation, ploughing of species-rich pastures,
pollution events), giving time for enforcement and/or management teams to prevent
further damage including loss of habitats and associated species. The increased
future use of targeted (e.g., red edge) channels on these satellites will also allow
better discrimination of, for example, plant species, particularly given the capacity
Expected Advances in a Rapidly Developing Work Area