Predict coral bleaching events

Coral reefs serve as a significant repository of marine biodiversity and are home to about 25 percent of marine species. The health and long-term success of these reefs are critical in maintaining our marine ecosystems. Unfortunately, reefs worldwide face many risks, primarily due to climate change: these risks include ocean temperatures that are too warm for extended periods of time, which endanger coral species and may result in coral bleaching. In this lesson, you'll use ArcGIS Pro and several years of daily sea surface temperature data to create a model that employs time-series forecasting capabilities to explore the impact of sea surface temperature on coral bleaching. Using model results, you'll predict and identify locations showing anomalously high-water temperatures. You'll use these locations to predict areas that are likely to be vulnerable to coral bleaching.

This tutorial was last tested on October 4, 2024, using ArcGIS Pro 3.3. If you are using a different version of ArcGIS Pro, you may encounter different functionality and results.

Requirements

• User, Publisher, or Administrator role in an ArcGIS organization or ArcGIS Enterprise (see configuration details)
• ArcGIS Pro (see options for software access)
• ArcGIS Image Analyst

Add multidimensional data to your map

First, you'll add the data to a map. Then, you'll change the map projection to one appropriate for your analysis.

Add the data

The data you'll use in this tutorial is hosted on ArcGIS Online in a zip file. To add the data to a map, you'll download it, unzip it, and create a new project in ArcGIS Pro.

1. Download the Multidimensional Data to Predict Coral Bleaching Events zip file.

   Note:

   Depending on your web browser, you may be prompted to choose a file location before you begin the download. Most browsers download to your computer's Downloads folder by default.

2. Extract the contents of the MultidimensionalAnalysis_SampleDataset.zip file to a location of your choice.

   The data packaged for this lesson comes from the NCAR Research Data Archive. It contains a Network Common Data Form (netCDF) file included in the Climate Forecast System Reanalysis (CFSR) product, with 35 years of monthly sea surface temperature data at a spatial resolution of 0.5 degrees. The netCDF format is commonly used to store and manage multidimensional scientific data.

3. Open ArcGIS Pro and sign in to your ArcGIS account.

   Note:

   If you don't have access to ArcGIS Pro or an ArcGIS organizational account, see options for software access.

   When ArcGIS Pro opens, it contains a list of blank project templates under the New Project heading. If you have created a project before, it will be included under Recent Projects.

4. Under New Project, click Map.

   

   The Map template can be used to create a project containing a 2D map with the Topographic basemap included.

5. In the Create a New Project window, for Name, type Coral Bleaching Prediction.
6. Ensure that Create a new folder for this project is checked and click OK.

   The project is created with a default map.

   Tip:

   To drag and pan the map, click and hold the left mouse button. To zoom in and out, hold the right mouse button or use the mouse scroll wheel. To rotate the view in the map, press the V key and use the mouse to alter direction.

   Because you'll be working with sea surface temperature data and not with land-based data, you'll update the basemap to one with less topographic detail.

7. On the ribbon, click the Map tab. In the Layer group, click the Basemap button and choose Light Gray Canvas.

   

   The basemap changes to one with less detail. This basemap will emphasize the sea surface temperature data, which you'll add to the map as a multidimensional raster.

8. On the Map tab, in the Layer group, click Add Data and choose Multidimensional Raster Layer.

   

   The Add Multidimensional Raster Layers window appears.

9. If necessary, for Output Configuration, choose Multidimensional Raster.

   

10. For Input File, Mosaic Dataset or Image Service, click the Import Variables button and choose Import Variables From File.

    

    The Import variables from NetCDF, GRIB or HDF files window appears. Using this window, you can browse to the location of a file with the appropriate file type. The data you downloaded at the beginning of the lesson includes a netCDF file, so you'll browse to it.

11. Browse to the location where you extracted the data and select the CFSR_sst.nc file. Click OK.

    The data is added to the Add Multidimensional Raster Layers window. Once you have specified the multidimensional raster for analysis, you can select variables to include. In this case, the source data contains only a sea surface temperature (cfsrsst) variable.

12. For Select Variables, check cfsrsst.

    

13. Click OK.

    A Multidimensional Raster Layer called CFSR_sst.nc_cfsrsst is created using the netCDF file and added to the Contents pane. The data is displayed on the map with a gradient color scheme. Areas in blue have lower temperatures and areas in red have higher temperatures. The temperatures use degrees Kelvin.

    

Change the map projection

When you added the data to the map, the map's projection changed. A map projection is the way that the 3D surface of the planet is converted to a 2D map. It's impossible to make this conversion without distorting some aspect of the world, so there are many different projections that reduce certain distortions while introducing others.

The current projection distorts area, so it does not allow for suitable comparison of size between regions. You'll change the projection to one that better represents the relative size of areas where coral bleaching events may occur.

1. On the ribbon, on the Map tab, in the Navigate group, click the Full Extent button.

   

   The map is zoomed to the full extent of the data.

2. In the Contents pane, double-click Map.

   

   The Map Properties pane appears.

3. Click the Coordinate Systems tab.

   This tab displays the current coordinate system, WGS 1984. You'll change it to the Behrmann projection, which is an equal-area projection.

4. In the search bar, type Behrmann and press Enter.
5. Expand Projected Coordinate System, expand World, and select Behrmann (world).

   

6. Click OK.

   The projection is applied to the map.

   

   The CFSR_sst.nc_cfsrsst layer represents sea surface temperature in degrees Kelvin. You'll update the layer properties to indicate this information.

7. If necessary, in the Contents pane, expand the CFSR_sst.nc_cfsrsst layer.
8. Click Value to select it.
9. Click Value again to edit it. Type Sea Surface Temperature (degrees Kelvin) and press Enter.

   

   The layer properties are updated. You'll also rename the layer. Its current name includes a lot of acronyms and can be difficult to understand.

10. If necessary, click the CSFR_sst.nc_cfsrsst layer name to select it. Click it again to edit it, type Sea Surface Temperature, and press Enter.
11. On the Quick Access Toolbar, click the Save button.

    

    The project is saved.

You've added the data to the map and changed the map projection so you can compare areas globally. You're ready to begin your analysis.

Visualize multidimensional data

Next, you'll investigate your data. Multidimensional data contains several layers of information stacked on top of each other into a cube. The stacked layers can represent data from different depths or height, or as in this example, time.

The sea surface temperature dataset used in this lesson consists of monthly temperature observations from 1980 to 2015, in which each month can effectively be seen as its own layer.

View the data

To view the monthly time slices of your dataset, you'll use the tools in the Multidimensional tab.

1. If necessary, open your Coral Bleaching Prediction project in ArcGIS Pro.
2. In the Contents pane, click the Sea Surface Temperature layer to select it.

   

   When a Multidimensional Raster Layer is selected, the Multidimensional tab becomes available on the ribbon.

3. On the ribbon, click the Multidimensional tab.

   The tab includes exploration, analysis, and data management tools.

4. In the Current Display Slice group, verify that Variable is set to cfsrsst.

   

5. In the Current Display Slice group, click the StdTime drop-down menu.

   The menu contains a list of the monthly time slices included in the data. Time is formatted as year, month, and day.

   

   Time slices represent monthly sea surface temperature observations. You can display individual time slices by choosing them from the menu. Alternatively, you can view them sequentially as an animation.

6. Click anywhere outside of the menu to close it.
7. In the Current Display Slice group, click the Play Slices Along StdTime button.

   

   The map sequentially displays the monthly time slices of the dataset.

8. When finished, click the Play Slices Along StdTime button again to pause the animation.

   

9. Save the project.

You've visualized the multidimensional data and become familiar with some of the multidimensional tools.

Generate trends and forecast sea surface temperature

Next, you'll use the multidimensional geoprocessing tools to investigate your data and run a trend analysis.

Generate a temporal profile

One way to review multidimensional data is with a temporal profile chart. Using the Sea Surface Temperature layer, the temporal profile chart will display a graph with time on the x-axis and the sea surface temperature on the y-axis. This graph will give an overview of how sea surface temperature has changed over time.

1. If necessary, open your Coral Bleaching Prediction project in ArcGIS Pro.
2. If necessary, in the Contents pane, select the Sea Surface Temperature layer.
3. On the ribbon, click the Multidimensional tab. In the Analysis group, click Temporal Profile.

   

   The Sea Surface Temperature - Chart of Sea Surface Temperature pane appears under the map and Chart Properties pane appears to the right of the map.

   The temporal profile chart allows you to define an area of interest using points, lines, or polygons. The chart will then plot values through time for the area of interest.

4. In the Chart Properties pane, click the Point area of interest tool.

   

   You'll set the area of interest to a location with high coral diversity.

5. On the map, with the Point tool active, click a location between Australia's northwestern coast and Indonesia. (If necessary, zoom in to better locate the area.)

   

   Note:

   Your temporal profile will differ based on the exact location you choose.

   The Sea Surface Temperature - Change in cfsrsst over StdTime pane updates to display a temporal profile of monthly sea surface temperatures from 1980 to 2015 for the location you chose.

   Tip:

   You can resize the pane to better see the chart.

   

   The chart indicates a cyclical pattern in sea surface temperatures. This pattern corresponds to the seasonal variation in temperatures at the location. You'll change the binning interval (how the data is grouped together in the chart) to show yearly average temperatures.

6. In the Chart Properties pane, expand Aggregation Options. Set the Time aggregation to Mean.
7. For Time binning options, click the Determine interval size for time aggregation button and change the interval to 1 Years.

   

   The chart updates to display average yearly sea surface temperature variations.

   

   By exploring the temporal profile, you know how the sea surface temperature has changed at the location you chose. The area has experienced higher average temperatures since 2000, which could put this area at risk for coral bleaching.

8. Close the Sea Surface Temperature - Change in mean cfsrsst over Standard Time and Chart Properties panes.

   Note:

   Closing the chart does not remove it from the project. The chart is accessible as an item in the Contents pane.

9. Save the project.

Calculate trends and predict sea surface temperature

Using additional multidimensional tools, you can analyze trends in the data. Understanding existing trends is helpful for predicting which locations will experience warming and be at risk of coral bleaching. Because coral reefs with rich biodiversity are highly concentrated in the region around Australia and Indonesia, you'll limit analysis to this area.

1. On the ribbon, click the Map tab. In the Navigate group, click the Go To XY button.

   

   A navigation pane appears under the map. You can use this pane to navigate to a location using its longitude and latitude.

2. In the navigation pane, for Long, type 127 E, and for Lat, type 10 S.

   

3. Press Enter.

   The map is centered on the coordinates you chose. However, the map scale is still zoomed to a worldwide extent.

4. Under the map, in the scale bar, type 42500000 and press Enter.

   

   The map zooms in. This region represents a critical part of the local and global coral ecosystems and is severely threatened by coral bleaching. You'll use this extent for the rest of your analysis, so be sure not to change it.

   Tip:

   If you change the extent on accident, you can return to it by clicking the Previous Extent button in the Navigate group on the Map tab.

   

   Next, you'll look for trends in the data.

5. On the ribbon, click the Multidimensional tab. In the Analysis group, click Trend.

   

   The Generate Trend Raster tool appears in the Geoprocessing pane. This tool estimates the trend for each pixel in a multidimensional raster layer based on one or more variables.

   Most of the default parameters are appropriate for your analysis. The Sea Surface Temperature layer is the input layer, the dimension is StdTime, and the cfsrsst variable is checked. You'll change the trend line type from linear to harmonic. A harmonic trend line is best used for data that follows a cyclical pattern, such as seasonal temperature.

6. In the Geoprocessing pane, for Trend Type, choose Harmonic.

   

   Note:

   The Generate Trend Raster tool creates an output that uses the cloud raster format (.crf). The .crf format is optimized for cloud-based processing and analytics and can also be displayed and processed in ArcGIS Pro.

   Because your dataset is large and processing may take several minutes, you'll apply a processing extent to limit the analysis to the region around Australia and Indonesia.

7. Click the Environments tab. In the Processing Extent section, for Extent, choose Current Display Extent.

   Note:

   Updating environment settings in a Geoprocessing tool overrides the default analysis and output settings associated with the project.

   

   The analysis processing extent changes to that of the current map extent.

8. Click Run.

   Note:

   Depending on the processing speed of your computer, the tool may take several minutes to run.

   The tool processes the input file and creates the output Sea Surface Temperature_GenerateTrend.crf file, which is added as a layer to the map.

   

   Purple areas are getting warmer, while green areas are getting colder. Most of the area within your map is getting warmer over time. You can use this trend analysis result to predict sea surface temperatures.

9. In the Contents pane, uncheck the Sea Surface Temperature layer to turn it off.
10. On the ribbon, on the Multidimensional tab, in the Analysis group, click Predict.

    

    Tip:

    Ensure that the Sea Surface Temperature_GenerateTrend.crf layer is selected to activate the Predict tool in the Analysis group.

    The Predict Using Trend Raster tool appears in the Geoprocessing pane. This tool uses a trend raster layer to generate a new multidimensional dataset.

    The Input Trend Raster and Variables parameters are already set to your trend raster (Sea Surface Temperature_GenerateTrend.crf) and the sea surface temperature variable (cfsrsst), respectively. The Dimension Definition parameter determines the values or intervals that the tool predicts. You'll set it to a weekly interval from January 1, 2011, to January 1, 2022.

11. In the Geoprocessing pane, change the following parameters:
    • For Output Multidimensional Raster, type Sea_Surface_Temperature_Predict.crf.
    • For Dimension Definition, select By interval.
    • For Start, type 2011-01-01T00:00:00.
    • For End, type 2022-01-01T00:00:00.
    • If necessary, for Value Interval, type 1.
    • For Unit, choose Weeks.

    

12. Click Run.

    The Predict Using Trend Raster tool runs and adds a predicted sea surface temperature trend layer to the map.

    Note:

    When the Sea_Surface_Temperature_Predict.crf layer is added to the map, it may not display initially. If it doesn't display, try refreshing the map and unchecking Layer Cache in the layer properties.

    

    The layer displays weekly predicted sea surface temperatures from January 2011 through December 2021. Like with your original multidimensional raster layer, you can explore the weekly time slices using the tools in the Multidimensional tab.

13. In the Contents pane, uncheck the Sea Surface Temperature_GenerateTrend.crf layer to turn it off.
14. On the ribbon, in the Multidimensional tab, in the Current Display Slice group, click the StdTime drop-down menu.

    

    The time slices start with January 1, 2011, and continue at an interval of seven days until they reach January 1, 2022.

    The predicted sea surface temperature map still does not tell you where coral bleaching events are likely to occur. However, you can use the predicted weekly temperatures and additional multidimensional geoprocessing tools to identify anomalies and pinpoint areas most likely to experience bleaching events.

15. Click anywhere outside of the drop-down menu to close it.
16. Save the project.

You've analyzed trends in your sea surface temperature data and used those trends to predict future sea surface temperatures. Next, you'll use the layers you've created to predict where coral bleaching events will occur.

Predict where coral bleaching events will occur

Coral bleaching events occur when reefs are exposed to elevated water temperatures for long periods of time. Now that you've used trend analysis to predict sea surface temperatures up to January 1, 2022, you'll analyze the prediction data to find locations where water temperatures remain warm for extended periods of time.

Predict coral bleaching

First, you'll calculate the anomalies in your data. In this context, an anomaly is the deviation of an observed value from its average value. Your analysis will show areas that have higher temperatures than average.

1. If necessary, open your Coral Bleaching Prediction project in ArcGIS Pro.
2. On the ribbon, click the Multidimensional tab. In the Analysis group, click Anomaly.

   

   The Generate Multidimensional Anomaly tool opens in the Geoprocessing pane. You'll adjust the tool parameters so that anomalies are identified by comparing the monthly average temperature at each location to the overall average temperature.

3. In the Geoprocessing pane, change the following parameters:
   • For Output Multidimensional Raster, type Sea_Surface_Temperature_GenerateAnom.crf.
   • For Mean Calculation Interval, choose Recurring monthly.

   

4. Click Run.

   The Generate Multidimensional Anomaly tool runs and adds the Sea_Surface_Temperature_GenerateAnom.crf layer to the map.

   

   Areas in blue have temperatures below the average, while areas in yellow and red have temperatures above the average. Much of the area between Australia and Indonesia is yellow at the displayed time slice.

   Like your other multidimensional datasets, this dataset has time slices. Some areas may have above average temperatures one week but below average temperatures the next week. Because coral bleaching events occur when sea surface temperatures are high for long periods of time, you'll calculate statistics for your surface temperature data to determine how often locations experience warm temperatures (between 0.1 and 5 degrees above the mean).

5. In the Contents pane, uncheck the Sea_Surface_Temperature_Predict.crf layer to turn it off.
6. On the Multidimensional tab, in the Analysis group, click Find Argument Statistics.

   

   The Find Argument Statistics tool opens in the Geoprocessing pane. This tool extracts values or bands at which a given statistic is attained. By setting the Statistics Type parameter to Duration, you can find the number of consecutive weeks when sea surface temperature is elevated for each location. When temperature is elevated for a long period of time, coral bleaching occurs.

7. In the Geoprocessing pane, set the following parameters:

   • For Output Raster, type Sea_Surface_Temperature_Statistics.crf.
   • For Statistics Type, choose Duration.
   • For Dimension Definition, choose Interval Keyword.
   • For Keyword Interval, choose Yearly.
   • For Minimum Value, type 0.1.
   • For Maximum Value, type 5.

   

   With these parameters, the tool will find the number of weeks in a year when temperature is between 0.1 and 5 degrees above average for each location.

8. Click Run.

   The tool runs and an anomaly layer is added to the map.

   

   Areas in blue are where temperature is above average for only a few weeks at a time, while areas in yellow and red are where temperature is above average for longer durations. At the current time slice, the area between Australia and Indonesia doesn't have warm temperatures for a long duration. However, this dataset has yearly time slices ranging from 2011 to 2021.

9. In the Contents pane, uncheck the Sea_Surface_Temperature_GenerateAnom.crf layer to turn it off.
10. In the Multidimensional tab, in the Current Display Slice group, click the Play Slices Along StdTime button.

    The map displays the time slices one after another. As the time slices reach 2021, more and more areas experience longer durations of above average temperatures.

    

    Some areas, like the area northeast of New Guinea, encounter extreme durations of elevated temperatures. Other areas have warm temperatures for smaller, but still concerning, durations. The area between Australia and Indonesia does not appear to experience above average temperatures for a prolonged duration.

    From this analysis, you can conclude that your study area is not likely to experience coral bleaching prior to 2022. However, other areas around the world could experience coral bleaching.

11. When the animation ends, save the project.

In this tutorial, you predicted coral bleaching events using historical sea surface temperature data and multidimensional data geoprocessing tools. First, you detected trends in the average water temperatures at each location through time. Using the trends, you predicted sea surface temperatures for the next few years. You identified anomalously warm temperatures and quantified the duration of those anomalous temperatures. By identifying areas that will experience prolonged periods of warm water temperatures, you can identify locations that are likely to experience coral bleaching events, a phenomenon that causes severe environmental and ecological damage.

Tip:

To create an alert map for coral reef stations: Build a map to help identify, prioritize, and manage coral reefs. Use the Living Atlas NOAA Coral Reef Watch (CRW) Virtual Stations feature layer and the Zonal Statistics tool (also located on the Multidimensional tab) to identify the number of weeks each reef location is expected to experience elevated water temperatures. This result can be published and shared with protection and management teams using an ArcGIS Operations Dashboard. For an example, see Coral Reefs at Risk of Bleaching.

Multidimensional data is often used in ocean and climate studies. In addition to identifying locations where coral bleaching is likely to occur, similar tools and processes can be used when investigating trends in temperature, precipitation, and ocean salinity. The multidimensional geoprocessing tools used in this lesson can help provide answers to many important questions in the world today.

You can find more tutorials in the tutorial gallery.

Acknowledgements

• Saha, S., et al. 2010. NCEP Climate Forecast System Reanalysis (CFSR) Monthly Products, January 1979 to December 2010. Research Data Archive at the National Center for Atmospheric Research, Computational and Information Systems Laboratory. Accessed 6/21/19.
• Light Gray Canvas map sources: Esri, TomTom, Garmin, FAO, NOAA, USGS, OpenStreetMap contributors, and the GIS User Community