DEA Water Observations (Sentinel 2 MSI)

DEA Water Observations 10m Sentinel-2 Multispectral Instrument

Version
1.1.6 flat-square
Program
Digital Earth Australia
Collection
Geoscience Australia Sentinel-2 Collection 1
Resource type
Derivative
Published Date
17/12/2021

Background

Understanding the amount of surface water in the landscape is important for many applications including water management, wetlands, water connectivity, surface-groundwater relationships and flood mapping. This can lead to more effective emergency management and risk assessment.

This product is the implementation of the DEA Water Observations (WOs) product (previously known as Water Observations from Space, or WOfS) on the Geoscience Australia Sentinel-2 surface reflectance product. The product is available as "Near Real Time", produced within 24 hours of the satellite passing over an area, and also "Definitive", once the satellite imagery can be fully corrected to a high surface reflectance standard.

The Near Real Time version is aimed to provide data to emergency management and decision makers on the extent of water in the landscape within approximately 24 hours of satellite overpass. The definitive version replaces the interim version on a rolling 3-month basis.

What this product offers

The Digital Earth Australia Water Observations (Sentinel-2) product shows where surface water was observed by the Sentinel 2a and Sentinel 2b satellites on any particular day since 2015. The surface water observations are derived from Geoscience Australia Sentinel-2 surface reflectance imagery for all of Australia. The interim, Near Real Time product is available for a rolling window of the most recent three months of data, and is produced within 24 hours of the satellite passing over an area. The definitive product reprocesses the Water Observations once full surface reflectance correction data is available, and is available for any day from 2015 onwards. 

The Water Observations show the extent of water in a corresponding Sentinel 2 scene, along with the degree to which the scene was obscured by data quality issues including; cloud, cloud shadows, and where sensor problems cause parts of a scene to not be observable. The Water observations are based on the WOfS algorithm which underpins all of the Digital Earth Australia surface reflectance water products.

The Near Real Time product is also available as a vector-based Water Observation layer in a 1 to 1 relationship with the input satellite data. Hence there is one vector layer for each satellite dataset available in the Geoscience Australia Sentinel-2 archive.

Applications

The Digital Earth Australia Near Real Time Water Observation product is provided as an interim source of information on water in the landscape to help inform emergency management and decision makers on recent and evolving emergency situations. The product can be used to determine the area of surface water present in the corresponding satellite scene, and can be used for several water monitoring applications. Uses of the individual WOs include:

  • flood extent
  • amount of water in water bodies, major rivers and the coastal zone.

As the WOs are separated from the derived statistics of the associated DEA Water statistical products, the WOs are most useful for performing analyses requiring the investigation of surface water extent for particular times rather than over long term time series.

The Definitive Water Observation product has been reprocessed to provide the best input data, using ancillary data for atmospheric conditions and surface reflection characteristics. The Definitive version is more accurate than the Near Real Time version, but takes much longer to process and become available. Hence it is less useful for emergency response applications, but more useful for long term water monitoring applications.

Related products

Publications

Mueller, N., Lewis, A., Roberts, D., Ring, S., Melrose, R., Sixsmith, J., Lymburner, L., McIntyre, A., Tan, P., Curnow, S., & Ip, A. (2016). Water observations from space: Mapping surface water from 25 years of Landsat imagery across Australia. Remote Sensing of Environment174, 341–352. https://doi.org/10.1016/j.rse.2015.11.003

Data access

Web services Web feature service
CMI RESTful node ID 639
Access constraints

© Commonwealth of Australia (Geoscience Australia) 2021. This product is released under the Creative Commons Attribution 4.0 International Licence.

Security classification Unclassified
Update frequency daily

Technical information

Digital Earth Australia (DEA) Water Observations (Sentinel-2) is a dataset indicating areas where surface water has been observed using the Geoscience Australia Sentinel-2 surface reflectance imagery catalogue. The Near Real Time version of this product includes observations from the most recent three months of data collected by the Sentinel-2 satellites. The Definitive version of this product includes observations from Sentinel-2 satellites from 2015 to the present. Near Real Time products are created with predicted corrections for atmospheric conditions and ground reflectivity, while the Definitive version uses measurements of these correcting factors to provide the best possible Analysis Ready Data for input to the Water Observations process.

Water Observations (WOs) cover all of mainland Australia and Tasmania but excludes off-shore Territories. The dataset is updated automatically as each new Sentinel-2 scene is acquired and processed to interim Analysis Ready Data (ARD) state. 

The Water Observations from Space (WOfS) Water detection algorithm is used to generate a pixel wise classification of Sentinel-2a and Sentinel-2b scenes. Full details of the original algorithms and features of DEA Water Observations can be found in the Water Observations from Space paper by Mueller et al. (2015). The output of this algorithm is a classification with the bit meanings described in table 1, with the exception for bit flag number 2; the sea mask, which is no longer used in the current version of the algorithm. Because of the way the current version (Version 3.1.6) of the WOfS algorithm operates multiple bits can be set simultaneously for each pixel. Hence the value of a pixel in an Observation can be X AND Y AND Z, etc, hence values can range from 0 to 255.

table describing the bit value flagging for WOfS classification

The outputs of the WOfS algorithm are re-grouped into the groups defined in table two. The rules for each group vary on whether the values included are inclusive or exclusive of multiple bit values. Group 1 and group 0 exclude multiple bit values. Group 2 includes multiple bit values. This means that pixels with a flag for only water are included in group 1, pixels with water and any other data flag (ie water AND slope or water AND cloud shadow) are included in group 2.

A table defining the grouping rules for Water observation classes: Group one: water: 128 is exclusive of simultaneous bit values. Group Two: Not Analysed: 4 (solar incident), 8 (terrain shadow), 16 (high slope), 32 (cloud shadow), 64 (cloud)  is inclusive of simultaneous bit values. Group zero: N/A: 1 (no data), 2 (contiguity), is exclusive of simultaneous bit values

Groups Water and Not Analysed are then converted into vector objects. These vector objects are then simplified a buffered by the distance of one water pixel (10 meters) to produce a smoother image and to minimise errors of topology when displayed on a web service, while retaining data integrity. Group 0 is not converted into vectors and as such is displayed as transparent. 

 

Full details of the original algorithms and features of DEA Water Observations can be found in the Water Observations from Space paper by Mueller et al. (2015).

 

Accuracy and limitations

Accuracy

The accuracy of the original WOfS algorithms was assessed using an independent set of 3.4 million validation points. The points were identified based on visual interpretation of Landsat imagery within 20 test areas across Australia. The points were identified in the same locations as the training data, but were selected from different years (i.e. imagery from one set of years was used to generate points to train the algorithm, and imagery from a separate set of years was used to generate the points that were used evaluate the accuracy of the algorithm).

The classification has an overall accuracy of 97%. Areas identified as water within the accuracy assessment data are being correctly identified 93% of the time and are being misclassified as not water 7% of the time. These errors of omission typically occur along rivers, small waterbodies and swamps where the presence of both water and vegetation within the pixel leads a failure to identify water. This means that the DEA Water Observations product is likely to underestimate the extent of water in locations that contain mixed water and vegetation pixels. As a consequence of this the product may not be fit for applications that require information about the inundation characteristics of vegetated wetlands, small farm dams, and rivers less than 50 metres wide. 

Water can be incorrectly detected by the classification algorithm in areas where steep terrain or tall buildings cause frequently shaded pixels. These errors of commission are occurring in 8% of samples used to evaluate the accuracy of the classification. This means that the product may overestimate the amount of water in locations that are adjacent to steep terrain or in dense urban areas. Terrain masks and urban masks were used in the confidence layer to reduce this overestimation, however some residual errors remain. As a consequence of this the product may not be fit for applications that require information about the inundation characteristics of urban areas or locations adjacent to steep terrain.

In addition to the limitations of the classification algorithm, the satellite observation frequency also introduces limitations to the product. The product is likely to be underestimating the extent of inundation for infrequent flood events because the 8 day revisit frequency (best case scenario notwithstanding the possibility of cloud obscuring the floodwaters) will potentially fail to observe the flood peak. This is an intrinsic limitation of the observation strategy. As a consequence of this limitation, the product is not suited to applications that require a. the identification of a ‘maximum extent of inundation’ line, or b. detailed information about the extent of infrequent flood events.

Limitations

Observation of Earth by the satellites used for this service depends on clear skies. Furthermore, the satellites do not observe all places every day. The Sentinel-2 satellites, which are the basis for this service, view a given 320 kilometre wide strip of Australia only once every 10 days. The observations show only what was visible on the day of the satellite pass. As a result, not all historical floods will have been observed by satellite.

The automated surface water detection algorithm, which has been developed by Geoscience Australia, can sometimes mistakenly label large buildings; cloud shadow; large uniform black tarpaulins; or snow as "water". The algorithm is designed to locate large areas of water and as a result may miss small water bodies. 

The satellite archive used for this service is of limited duration (most recent 3 months), and subject to the cloud and repeat coverage restrictions noted above. In addition, Australia is subject to wide variations in weather and climate which can lead to lengthy periods where areas are not observable.

References

Mueller, N., Lewis, A., Roberts, D., Ring, S., Melrose, R., Sixsmith, J., Lymburner, L., McIntyre, A., Tan, P., Curnow, S., & Ip, A. (2016). Water observations from space: Mapping surface water from 25 years of Landsat imagery across Australia. Remote Sensing of Environment174, 341–352. https://doi.org/10.1016/j.rse.2015.11.003

Lineage

The DEA Water Observations dataset is generated using the following steps:

Imagery of the earth’s surface is collected by the Sentinel-2a and Sentinel-2b satellites as a raster based product. Digital Earth Australia receive this data and conduct atmospheric and Bidirectional Reflectance Distribution Function corrections to produce Geoscience Australia Sentinel-2 Analysis Ready Data to either predicted (interim) or measured (definitive) standard.

Once the input data has been processed to ARD the Water Observations algorithm is applied, including cloud and shadow masking, on a per-granule basis and indexed into the Data Cube for further processing. 

Interim (Near Real Time) DEA WO (Sentinel-2) is further converted to vector format for delivery to emergency services via WFS. Interim WO are kept in a most recent 3 months rolling archive. 

Definitive DEA WO (sentinel-2) is computed once final ancillary data has been acquired to allow creation of Definitive ARD, at which point Interim data are reprocessed to the final Definitive WO version.

Data sources

Processing steps

  1. Surface Reflectance Calculation (NBAR + Terrain Illumination Correction)
  2. Water Observations from Space Detection Algorithm 1.2

Major algorithms

Schema / spatial extent

Australia WGS84 Raster Schema

Update frequency asNeeded
Temporal extent
Min. longitude 112.00
Max. longitude 154.00
Min. latitude -44.00
Max. latitude -9.00
Coordinate reference system WGS 84 (EPSG: 4326)

Owner

Commonwealth of Australia (Geoscience Australia)

Principal contributors

Norman Mueller

Subject matter experts

Norman Mueller

License

CC BY Attribution 4.0 International License

Rights statement

© Commonwealth of Australia (Geoscience Australia) 2021. Creative Commons Attribution 4.0 International License.