Information, Resources and Frequently Asked Questions for Allen Coral Atlas Users

Literature written about the methodologies used and products created for the Atlas

• Asner, G.P., N.R. Vaughn, J. Heckler, D.E. Knapp, C. Balzotti, E. Shafron, R.E. Martin, B.J. Neilson, and J.M. Gove. 2020. Large-scale mapping of live corals to guide reef conservation. Proceedings of the National Academy of Sciences https://doi.org/10.1073/pnas.2017628117 
• Emma Kennedy, Chris Roelfsema, Mitchell Lyons, Eva Kovacs, Rodney Borrego-Acevedo, Meredith Roe, Stuart Phinn, Kirk Larsen, Nicholas Murray, Doddy Yuwono, Jeremy Wolff, Paul Tudman “Reef Cover: a coral reef classification system for global habitat mapping from biophysical remote sensing” bioRxiv preprint  https://doi.org/10.1101/2020.09.10.292243
• Li, J., D.E. Knapp, N.S. Fabina, E.V. Kennedy, K. Larsen, M.B. Lyons, N.J. Murray, S.R. Phinn, C.M. Roelfsema, and G.P. Asner. 2020. A global coral reef probability map generated using convolutional neural networks. Coral Reefs doi:10.1007/s00338-020-02055-6
• Li, J.; D.E. Knapp; S.R. Schill; C. Roelfsema; S. Phinn; M. Silman; J. Mascaro; G.P. Asner. 2019. Adaptive bathymetry estimation for shallow coastal waters using Planet Dove satellites. Remote Sensing of Environment 232: 111302 https://doi.org/10.1016/j.rse.2019.111302 
• Li J., N.S. Fabina, D.E. Knapp, G.P. Asner. The Sensitivity of Multi-spectral Satellite Sensors to Benthic Habitat Change. Remote Sensing. 2020, 12, 532 https://doi.org/10.3390/rs12030532
• Li, J.; S.R. Schill; D.E. Knapp; G.P. Asner. 2019. Object-Based Mapping of Coral Reef Habitats Using Planet Dove Satellites. Remote Sensing 11: 1445. https://doi.org/10.3390/rs11121445
• Lyons, Mitchell, Chris M. Roelfsema, Emma V. Kennedy, Eva M. Kovacs, Rodney Borrego-Acevedo, Kathryn Markey, Meredith Roe, Doddy M. Yuwono, Daniel L. Harris, Stuart R. Phinn, Gregory Asner, Jiwei  Li, David E. Knapp, Nicholas S. Fabina, Kirk Larsen, Dimosthenis Traganos, Nicholas J. Murray (2020) Mapping the world's coral reefs using a global multiscale earth observation framework.  Frontiers in Ecology and the Environment. doi.org/10.1002/rse2.157
• Roelfsema, Chris M., Kovacs, Eva M., Ortiz, Juan Carlos, Callaghan, David P, Hock, Karlo, Mongin, Mathieu, Johansen, Kasper, Mumby, Peter J., Wettle, Magnus, Ronan, Mike, Lundgren, Petra. Kennedy, Emma V. Phinn, Stuart R. (2020) Habitat maps to enhance monitoring and management of the Great Barrier Reef. Coral Reefs (2020). https://doi.org/10.1007/s00338-020-01929-3 
• Roelfsema CM, Lyons M, Murray N, Kovacs EM, Kennedy E, Markey K, Borrego-Acevedo R, Ordoñez Alvarez A, Say C, Tudman P, Roe M, Wolff J, Traganos D, Asner GP, Bambic B, Free B, Fox HE, Lieb Z and Phinn SR (2021) Workflow for the Generation of Expert-Derived Training and Validation Data: A View to Global Scale Habitat Mapping. Front. Mar. Sci. 8:643381. https://doi.org/10.3389/fmars.2021.643381
• Xu, Y., N.R. Vaughn, D.E. Knapp, R.E. Martin, C. Balzotti, J. Li, S.A. Foo, and G.P. Asner. 2020. Coral bleaching detection in the Hawaiian Islands using spatio-temporal standardized bottom reflectance and Planet Dove satellites. Remote Sensing 12, 3219 https://doi.org/10.3390/rs12193219

For satellite images obtained from the Allen Coral Atlas website at www.allencoralatlas.org, use the following citation:

The Satellite Coral Reef Mosaic is © 2021 Planet Labs and licensed CC BY-SA-NC 4.0 (https://creativecommons.org/licenses/by-nc-sa/4.0/)

For screenshots of the Allen Coral Atlas, use the following citation:

Image credit: Allen Coral Atlas

• NOAA Coral Reef Watch has specific recommendations for Coral Reef Watch Data and Product Citation, described in more detail in our Science & Methods section.

• Protected Planet’s marine protected area data should be cited as:

  UNEP-WCMC and IUCN (2023), Protected Planet: The World Database on Protected Areas (WDPA), updated monthly, Cambridge, UK: UNEP-WCMC and IUCN. Available at: www.protectedplanet.net. 

• Marine Regions' national maritime boundary data should be cited as:

  Flanders Marine Institute (2022): MarineRegions.org.  Available online at www.marineregions.org.  Consulted February 2022.

• Some other data on the Atlas website is copyright by the data providers (e.g., Open Street Map) and requires proper attribution if reused.

• For more information on citation and attribution visit our GitHub page

Q: What is the depth limitation of the Atlas?

The benthic layer includes areas <10m in depth. The geomorphic layer includes areas <15m in depth.

Q: Can the Atlas detect coral health, or show where certain coral species are?

No, the Atlas does not have the spatial and spectral image resolution nor the mapping method to accurately make these types of classifications or assessments. Other data layers would be needed to make determinations about coral health, and a much higher resolution would be necessary to identify species of corals.
Consider using other web-based platforms such as DataMERMAID that hosts the Atlas’ data layers, as well as other relevant datasets that may address these features in your area of interest.
With the download feature available on the Atlas, you can combine our data layers with other geospatial data in your own GIS platform.

Q: Why isn’t my reef mapped? Why can’t I see my reef?

There are a few possible reasons your area of interest isn’t mapped. It is possible that your area of interest is outside of the scope of the Atlas, above 30degrees North or below 30degrees south latitude, or in deeper water.

If your area is within the latitude and longitude bounds and within the depth abilities of the Atlas but is not mapped, your area of interest may have been too turbid, or otherwise obscured, for the mapping process. Try looking at the satellite imagery with data layers turned off to see if the reef area is turbid. You can also switch the satellite imagery to “Analytic Mosaic” - it can show what may have been obscuring your area of interest, because the analytical mosaic has been used for the actual mapping.

Q: Why did you combine coral/algae as a single mapped class in the benthic community maps?

In the type of satellite imagery currently being used for mapping, coral and the algal forms growing on reefs “look” highly similar and it is difficult to separate them, hence the combined class, especially at a regional scale mapping approach. The microscopic algae that lives inside the living coral has similar concentrations of chlorophyll-A that fleshy algae, growing on top of dead corals, has. This gives them similar spectral reflectance signatures, making them difficult, if not impossible, to differentiate with the technologies we are using for the Atlas project. We’re investigating routes to differentiate these in the future.

Q: How do I download the data (“create a download package”)?

Follow our step-by-step guide for data downloads. 

Q: How do I download just my country data from the Atlas?

The easiest way is to toggle the "Maritime Boundaries" layer on, click inside your country's maritime boundary, then click the download icon.  This gives you immediate access to a pre-built download package with reef habitat polygons and statistics.

Alternatively, you can either draw a polygon around your country manually, or upload your country's maritime boundary. To do this, download the boundary for the country you're interested in (we recommend Marine Ecoregions ) as a KML or GeoJSON. After logging into your account on the Atlas, go to the “My Areas” section and find the “Upload” button. Upload your EEZ boundary file. You will see it form a polygon around the country. Save this area to “My Areas” then follow the steps to download the data from an area of interest.

Q: What is included in the download package?

During the download process you have options of which data layers to include. You can opt for:

1. Visual imagery (2018, 2019, or 2020)
2. Benthic Map layer with file type options
3. Geomorphic Map layer with file type options
4. Annual Turbidity Monitoring data
5. Bathymetry data

Sometimes fewer choices are available.  This can be due to data licensing restrictions or the availability of different data layers within your area of interest.

Q: What is the file size of data downloads?

Download size will depend on the size of the area of interest and the layers you have opted to include.  Raster data (visual imagery, turbidity, and bathymetry) tend to be much larger than vector data (benthic and geomorphic map data).  A visual imagery download for a large AOI, for example, can run into the hundreds of gigabytes, and may not even be possible.  Please always choose the smallest area of interest and fewest data layers that meet your requirements.

Q: How long do I have to retrieve my downloads?

Downloads are available for 30 days after they are created.

Q: I have a specific area I want to view on the Atlas. Can I upload my own polygon boundaries?

Yes. Find the upload button under “My Areas.” Here you can upload a geojson or KML file that will bring up a polygon onto the Atlas platform. It will be saved under your “My Areas” section.

Q: Why can't I download data for an area larger than 100km2?

You can! To save an area this large, or download data of any size, you need to create a free login. Follow the prompt to log in, and then click "Don't have an account? Register for free" if you don't already have one. Large datasets may take some time to generate, therefore with a free account, we will notify you via email when your dataset is ready. Most files take between a few minutes and a few hours to generate.

Q: What are the limits on the areas I can draw or upload?

Very large or highly complex areas can exceed the processing capabilities of the Atlas and degrade overall system performance.  Therefore, we limit areas to not more than 5 million square kilometers and not more than 500 interior rings (holes inside the area).

Q: Is mangrove data available?

The Atlas does not include maps of mangroves. However, Global Mangrove Watch is an excellent resource available for global spatial data of mangroves.

Q: What about seagrass data?

Seagrass is a class included in the benthic map on the Atlas, but is only included when the feature is part of a coral reef.

Q: Is there bathymetry data available?

Yes. Bathymetry data is available for download only.

Q: How do I access the Atlas API?

Our benthic and geomorphic habitat maps are available via a Web Map Service (WMS) and Web Feature Service (WFS) at the links below to facilitate further analysis or exploration of the data, subject to our Terms of Use.  If you have ideas for application integrations with the Atlas, please let us know at support@allencoralatlas.org.

Reproduction or publication of the entire global dataset of habitat maps is prohibited without the prior written consent of the Allen Coral Atlas leadership team at Arizona State University.

https://allencoralatlas.org/geoserver/ows?service=wms&version=2.0.0&request=GetCapabilities

https://allencoralatlas.org/geoserver/ows?service=wfs&version=2.0.0&request=GetCapabilities

Tips for using these services can be found here.

There is also an API endpoint to return metadata about the habitat maps:

https://allencoralatlas.org/mapping/maps

Q: Where do search results come from?

When you use the search feature in the Atlas, we build your results from three sources.

1. The first results are from among your saved areas of interest and our own mapped areas.
2. The next results are searchable reef names provided by ReefBase, a project of WorldFish.
3. The remaining results are powered by OpenStreetMap.

Q: What is the accuracy of the mapping?

The overall map accuracy is between 60-90% and varies per region mapped. Accuracy is influenced by environmental characteristics (e.g. habitat complexity, water depth, turbidity) and the quality of the analytical image mosaic, satellite derived water depth, and reference data set used for training and validation. 

The accuracy is calculated by comparing the mapped classes for many small areas with a validation data set for the same area (see Roelfsema et al 2013). The validation data set is different from the training data as the latter is only used for the classifier to create the map. Validation data and training data are derived from a reference sample set create with the help of expert knowledge, image interpretation existing field data and maps (See Roelfsema et al 2021).

Roelfsema C.M. and S.R. Phinn (2013) Validation Chapter 14, In: Coral Reef Remote Sensing: A Guide for Multi-level Sensing Mapping and Assessment. Goodman, J., Purkis, S. and Phinn, S.R. Springer Publishing pp375-365 ISBN 978-90-481-9291-5

Roelfsema, C. M., M. Lyons, N. Murray, K. E., E. Kennedy, K. Markey, R. Borrego-Acevedo, A. Ordonez Alvarez, C. Say, P. Tudman, M. Roe, J. Wolff, D. Traganos, G. Asner, B. Bambic, B. Free, Z. Lieb, H. E. Fox and R. S. Phinn (2021 soon to be published). "Workflow for the generation of expert-derived training and validation data: A view to global scale habitat mapping." Frontiers in marine Science: Special Issue Remote Sensing for Applied Coral Reef Science and Management.

Q: Why does my area of interest have misclassifications or errors?

Errors may be due to cloud cover or turbidity blocking the image, a masking issue, or because of the limitations of the semi-automated machine learning algorithm.

Imagery issue

First, turn off the benthic or geomorphic map layer and turn on the “Analytic Mosaic” satellite imagery layer to see the imagery used for the actual mapping.  Look for what features may be obscuring your area of interest. Turbidity or cloud cover visible in the imagery may be obscuring the reef area, or another artifact may be blocking the image.

Masking challenges

Additionally, some areas may have an issue with land masking or deep water masking. If you see map classifications appearing on land or in very deep water, this is likely due to an issue with the land or deep water mask.

Quality of analytical low tide image mosaic used for mapping

Other artifacts could be present that are natural such as thin cloud cover or rough seas, or artificially created through in the preprocessing step of the low tide image mosaic. The mosaic used is a composition of many small scenes captured at the lowest tide and cloud free, that are then stitched together almost seamlessly.

Quality of satellite derived depth

The depth is derived from a composite of best available satellite imagery. This is predominantly Sentinel 2 and Landsat 8 imagery due to their high signal-to-noise ratio, with some remote regions of the world filled in with Planet Dove imagery. The depth is best derived for clear waters where bottom feature can be differentiated. Hence, when water is not as clear, or has disturbances (artifacts), errors in depth will occur that can have an effect on the ability to distinguish classes.

Limitations of the semi-automated mapping algorithm

It is important to bear in mind that the maps are created with a semi-automated machine learning algorithm and cleaned up with an object based analysis approach. The algorithm labels certain features based on the availability and quality of the data utilized during the mapping process. The maps are created and validated for whole reef systems across countries and not focused on a small set of reefs. For some areas we only have limited reference data available or the data is dependent on expert interpretation only. A map generated in this way may not be as accurate as a very localized map created with extensive surveys over a small area, but it does allow for mapping on a much larger scale and globally consistent basis. 

If you are finding significant errors, please contact us with a description of the issue to feedback@allencoralatlas.org. Directions for submitting feedback are in the Contact Summary section below.

While many localized errors cannot be immediately corrected on the Atlas, it is always possible to download the data for your local area and alter the data layers in your own GIS platform to reflect additional local data or knowledge you have.

Q: What should I do if I find an error or problem in the map?

We welcome your feedback! 

As this is an evolutionary process, and the Allen Coral Atlas is in a beta stage, the maps and monitoring data may have issues, some of which we are aware of, others we’d love to have you report to us so we can triage what we can and cannot resolve.

Please submit any identified errors with suggested corrections to feedback@allencoralatlas.org including: Reef name(s); coordinates and/or a georeferenced polygon (in kml, shapefile, or JSON format); screen shots from the Atlas of relevant area (annotated with details), explanation of the error and your suggestion for correcting it/them; any relevant field data; your contact details.

We triage the feedback and, if possible, especially when its returning error, determine if it can be solved by adjusting the semi automated mapping process. Some errors can not be resolved as a result of the constraints of the mapping approach and input dataset used.

Q: If I collect new data for the Atlas, will that help improve my area of interest on the Atlas?

With ongoing accessibility of new, more and better satellite sensors next to further developments in mapping and masking routines, theoretically any new data could be integrated in improving and updating existing versions of maps for some mapping regions. However this will depend on available resources for the Atlas project.

If you have georeferenced data for a region on the Atlas, we may be able to utilize it for our ongoing revisions to the maps. Please get in touch with us at submissions@allencoralatlas.org or contact a member of the Atlas team.

Q: When do you update the maps on the Atlas?

Periodically, we will publish updates to the maps you find on the Atlas based on feedback we have received or improvements to the mapping algorithm.  The most recent publication dates of all of our maps can be found here.

Q: What updates have been made to the original habitat maps (completed in 2021) to get to their current state (released in 2022)?

Our mapping updates are a result of adjusting the mapping approach based on feedback from local experts and ongoing method improvements. The impacts of the updates vary between regions. In general, they are minor at a regional scale but could be moderate at a local scale, and differences may be visible in the statistical analysis or the representation of the reefs by the maps. Updates include two types of innovations:

• Global innovations were applied to all 30 mapping regions without rerunning the whole mapping process. This resulted in some geomorphic and benthic categories having been changed and some areas having been deleted, but no new areas were gained.

• Custom regional innovations were applied to only 10 out of the 30 mapping regions. For these 10 regions, we reran the mapping process which included global innovation improvements, but also resulted in new areas gained that were previously not mapped, and more focus was given to adjusting misclassification based on user feedback. Priority was given to regions released early in our project (2018, 2019) where there was the most room for improvement.

See here for more technical details.

Q: How current is the satellite imagery that was used to create the maps?

The dates of the source imagery for all of our maps can be found here.