Get started with change detection for Africa

In this second tutorial of the Explore satellite Imagery for Africa series, you'll continue to learn about satellite imagery and remote sensing, but you now focus on identifying, delineating, and extracting change. You'll look at three real-life examples: a wildfire in Kenya, water inundations in Chad, and a fast-growing city in Ghana. In all three cases, you'll extract change, assess how the landscape was impacted, and understand how your analysis could be used to address resulting challenges.

To perform your analyses, you'll use the Digital Earth Africa Explorer app, the result of a collaboration between Esri and the Digital Earth Africa program. The app enables users to seamlessly explore some of the Digital Earth Africa satellite imagery and other earth observation data for the African continent. You'll also learn how to export the data extracted to a local file or ArcGIS Online for further use.

This tutorial was last tested on August 10, 2023.

Requirements

• Free Africa GeoPortal account or a Publisher or Administrator role in an ArcGIS organization (see options for software access)

Assess wildfire damage

Change detection is one of the fundamental applications in imagery and remote sensing. It typically consists of comparing imagery datasets collected at different times in a single area to determine the type, magnitude, and location of change. Change can occur because of human activity, abrupt natural disturbances, or long-term climatological or environmental trends.

The first example of change detection that you'll work on is in the context of a wildfire. In Tsavo West National Park in Kenya in August 2020, a devastating wildfire raged for several days. The fire caused extensive damage in the park, although thankfully there were no human casualties. As a GIS analyst for the national park, you want to identify the burn scar caused by the fire and extract it into a new data layer. This new burn scar layer will provide a precise outline of the damaged area and will be useful to manage the fire aftermath proactively and mitigate the negative effects of the fire on the park's wildlife and tourism activities.

Set up and explore the imagery

First, you'll open the app, set up your location of interest (or extent), and choose the most relevant imagery.

1. Open the Digital Earth Africa Explorer app.
2. The app opens to the display of the African continent.

   

   You'll first locate the fire location on the map.

3. In the search box, type Taveta, Kenya and press Enter.

   

   The map updates to the new location.

4. Close the Search result pop-up.

   

5. Click the Zoom Out button four times.

   

6. Pan to the east (right) until you see both Taveta and Taita-Taveta, as well as the road A23 that connects the two.

   

   This is the location where the fire occurred in August 2020. To visualize the fire site over time, you'll use the Landsat imagery layer.

   Note:

   Landsat is an earth observation satellite program from the United States. It has been providing multispectral imagery for the entire earth continuously since 1972. You can learn more about Landsat on the program's website.

   You can also learn more about what multispectral imagery is and how to work with it in the first tutorial of this series: Get started with imagery for Africa.

   The Landsat layer included in the app provides a large selection of Landsat imagery from 1972 to present.

7. On the sidebar, click the Explore Imagery button.

   

8. In the Explore Imagery window, for Layer, choose Landsat. For Rendering, choose Agriculture with DRA.

   

   On the map, the imagery updates.

   The agriculture band combination uses the shortwave infrared, near infrared, and blue spectral bands. It highlights healthy vegetation in bright green and bare soil or the absence of vegetation in brown. DRA stands for dynamic range adjustment, and it is a technique to improve the contrast of the image.

   By default, you are seeing the most recent imagery captured, usually no older than a few weeks from today's date. To study the effect of the wildfire, you need to look at imagery that was captured shortly before and after the fire, which occurred from August 8 to August 10, 2020.

9. In the Explore Imagery window, examine the Date section.

   

   The time slider enables you to move through time and see all available individual Landsat images (or scenes) for the current extent. The pointer should currently be on May 20, 2023 or another recent date. You'll start scrolling through time.

10. On the time slider, click the plus and minus buttons to move through time until you reach June 4, 2020.

    

    Tip:

    You can also drag the pointer to the desired date.

    After a few moments, the imagery updates.

    

    This imagery shows how the area looked before the fire. Despite the presence of a few small, white clouds, you can clearly see the undisturbed land covered with vegetation (in bright green) and some patches of bare soil (in reddish brown tones).

11. On the right side of the timeline, click the plus button once to move to June 20, 2020.

    

    This scene shows a similar landscape, although the presence of vegetation is a bit less clear.

    Note:

    Variations in quality are frequent with imagery, so it is important to review several scenes available and focus on the ones with the highest quality. From that point of view, the June 4, 2020, scene is the best image available relatively shortly before the fire.

12. Click the plus button once to arrive to October 10, 2020.

    

    This image was captured after the fire. The burn scar appears clearly in dark brown, contrasting with the unaffected land in light brownish and greenish tones.

13. Click the Plus button twice to arrive to August 26, 2021.

    

    The burn scar is still visible, but it is beginning to fade, as the vegetation is starting to grow back.

Extract the burn scar pixels

Now that you've identified scenes before and after the fire, you'll choose the scene where the burn scar is the most visible to extract the image pixels that correspond to the burnt areas.

1. Use the minus button to go back to October 10, 2020.

   To perform the pixel extraction, you'll use the Create a mask tool. This tool creates a layer in which each pixel corresponding to a specific criteria is masked, or covered with a solid color.

2. In the Explore Imagery window, set the following options:
   • For What do you want to do?, choose Create a mask.
   • For Method, choose Soil Adjusted Veg. Index.
   • For Which values should I mask?, verify that Less than my threshold is selected.

   

   The Soil Adjusted Vegetation Index applies a mathematical formula to compute a ratio between the values of the near infrared and red bands of the imagery. It highlights healthy vegetation and distinguishes it from unhealthy vegetation or areas devoid of vegetation such as bare earth.

   The Soil Adjusted Vegetation Index will output a numeric result for each pixel: a high value means healthy vegetation and a low value means unhealthy or no vegetation. In this example, the lowest values will correspond to the areas burnt by the wildfire. The Less than my threshold option enables you to choose a threshold and mask all the pixels that have a low value and are below that threshold.

   First, you'll draw a polygon to define your area of interest (AOI) and choose a color for your mask.

3. Check the Draw polygon(s) to define extent box.

   

4. On the map, start drawing a polygon to delineate the AOI by clicking each corner (or vertex) of the desired polygon. When you are satisfied with the shape, double-click to complete.

   The result should look similar to the following image:

   

   When you generate the mask, it will only appear within that AOI.

5. Click the Choose a color for your mask button and choose the red color.

   

6. Click Apply.

   After a few moments, the mask appears on the map. In the Explore Imagery window, the Set your threshold slider also appears; you'll use it to choose the mask threshold.

   

   The slider shows that the vegetation index values for your AOI vary from about 0.01 to 0.26. The mask tool chose automatically a threshold of 0.13.

   Note:

   Based on the exact boundaries you chose for your AOI polygon, the values you see might be slightly different.

7. Move the threshold slider back and forth to see how the mask changes based on the threshold value.
   • If you choose 0.01 (or the minimum possible value) as a threshold, no pixel is included in the mask.
   • If you choose 0.26 (or the maximum possible value) as a threshold, all the pixels in the AOI are included in the mask.
   • If you choose a value in between, only some of the pixels in the AOI are included in the mask.
8. Choose a threshold value that captures all the burn scar pixels in the mask but as few other pixels as possible, such as 0.12.

   Your final mask should look similar to the following image:

   

   Note:

   Using the Soil Adjusted Vegetation Index to identify burnt areas is one approach. Another one would be to use the Normalized Burn Ratio, which computes a ratio between the near infrared and shortwave infrared values. In this specific case, the former approach works better than the latter. With imagery, it is always worth trying different methods and choosing the one that gives the best results.

   In the Explore Imagery window, Area Covered indicates that the current mask covers an area of 558.06 kilometers squared (your number may differ slightly). This is the surface area that was burnt in the fire.

   

Export the burn scar layer

Now that you've created the mask, you'll save the resulting layer to ArcGIS Online, enabling you to reuse it for the creation of maps or further analyses. This process requires an ArcGIS Online account.

Note:

If you don't have an ArcGIS Online account, you can get one for free through the Africa GeoPortal (click Sign In, click Create an Africa GeoPortal account, and follow the instructions). Africa GeoPortal is an open mapping community supported by Esri, working together to provide data and insights across Africa.

Alternatively, you can create a free ArcGIS Public account (click Create an ArcGIS public account and follow the instructions).

ArcGIS Online allows you to create dynamic web maps, which you can share with your community.

1. In the Digital Earth Africa Explorer app, on the sidebar, click the Image Export button. If necessary, click it a second time.

   

   The Image Export window appears.

2. In the Image Export window, set the following options:
   • For Export, choose Result Image.
   • For Save location, verify that Save to portal is selected.
   • For Title, type Tsavo burn scar.
   • For Description, type Burn scar from the August 2020 wildfire in Tsavo National Park West, Kenya.
   • For Tags, type wildfire, burn scar, Kenya.

   

3. Click Preview.
4. In the Sign In window, enter your ArcGIS Online or Africa GeoPortal credentials.

   

   In the Image Export window, a preview thumbnail appears.

   

5. Click Save.
6. When the process is complete, close the Image Export window.

   Note:

   Another option is to export the layer to your local computer. For location, choose Save to disk; for TIFF download options, choose As displayed; and for Advanced save options, choose Download as tiff. This option will produce a georeferenced TIFF image that you can open in any GIS software, such as ArcGIS Pro.

Create a web map to showcase the layer

Next, you'll visualize the layer that you saved to ArcGIS Online and create a web map to showcase it.

1. In your web browser, open a new tab and go to ArcGIS Online. Sign in with your ArcGIS Online or Africa GeoPortal credentials.
2. On the ribbon, click Content.

   

3. In the content list, click Tsavo burn scar.

   

4. On the Tsavo burn scar item page, click Open in Map Viewer.

   

   The layer displays in Map Viewer on top of the default topographic basemap.

   

   You'll change the basemap to display more information about the area.

5. On the Contents (dark) toolbar, click Basemap and choose OpenStreetMap.

   

   Note:

   This basemap comes from the OpenStreetMap community mapping project.

   The basemap updates, displaying the boundaries of the national park's conservancies, wildlife sanctuaries, and ranches.

   

   A quick review of the burn scar against the new basemap indicates that the fire, while damaging, did not reach most of the key areas of the park, with the exception of the edge of the Lumo Conservancy.

   Note:

   In real life, the national park's GIS analyst might maintain a more detailed layer of the park's assets, which could be added to this map for higher accuracy and thoroughness.

   Next, you'll save the map and share it with your colleagues in the national park or the community at large.

6. On the Contents toolbar, click Save and open and choose Save as.

   

7. In the Save map window, enter a title (such as August 2020 fire in Tsavo), tags, and a brief description. Click Save.
8. On the Contents toolbar, click Share map.

   

9. In the Share window, for Set sharing level, click Everyone (public).

   

10. Click Save.
11. In the Review sharing window, click Update Sharing.

You extracted the burn scar pixels corresponding to the August 2020 Tsavo National Park fire. You then exported the layer to ArcGIS Online and built a web map to showcase it and share it with your colleagues in the national park and the community at large. You could also use the layer in further analyses. You could add more data layers, for instance, about wildlife's locations and movement patterns, and proceed to further analyses to better understand the impact of the fire.

Delineate inundated areas

Due to heavy rains in July 2022, several areas located in the sous-préfectures of Léré and Guegou (department of Lac Léré, province of Mayo-Kebbi Ouest) suffered severe inundations. As a GIS analyst for the regional government, you've been asked to give a quick assessment of the areas most impacted to help the relief teams best focus their effort.

In this workflow, you'll learn how to extract change by comparing before and after images.

Set up and explore the imagery

First, you'll set up the extent and the imagery.

1. In your web browser, switch to the Digital Earth Africa Explorer app tab.
2. In the search box, type Léré, Chad (or Lere, Chad) and press Enter.

   

   The map updates to the new location.

3. Close the Search result pop-up.
4. Click the Zoom Out button three times.

   On the map, you see two lakes and several towns and villages around them.

   

   You'll compare imagery before and after the July 2022 heavy rains using the Compare Imagery tool.

5. On the sidebar, click the Compare Imagery button.

   

   The Compare Imagery window appears. With this tool, you can specify two images to display on the left and right sides of the map, and swipe between the two.

6. In the Compare Imagery window, set the following options for the left image:
   • Verify that Left Image is selected.
   • For Layer, verify that Landsat is selected.
   • For Rendering, verify that Agriculture with DRA is selected.
   • Check the Select a date check box.
   • Drag the time slider to May 8, 2022.

   

   The imagery updates to show a Landsat scene captured on May 8, 2022, before the heavy rains. The two lakes, Lake Léré and Lake Tréné, have clearly defined contours. You can also see thin blue lines corresponding to the Mayo Kébbi and Bénoué rivers.

   

7. In the Compare Imagery window, set the following options for the right image:
   • Click Right Image to select it.
   • For Layer, select Landsat.
   • For Rendering, select Agriculture with DRA.
   • Check the Select a date check box.
   • Drag the time slider to July 11, 2022.

   

   The imagery updates to show a Landsat scene captured on July 11, 2022, showing the situation on that date. Large areas that were dry land in the previous scene are now covered with water, making the lakes appear overextended and indistinguishable from the rivers.

   

8. On the map, grab the swipe handle and swipe repeatedly from left to right to compare the two images.

   

   Swiping enables you to examine in detail the differences between two images.

   Note:

   In this scenario, you are doing an assessment of the situation in mid-July to help organize the relief effort. However, the heavy rains went on in that region through July, August, and September 2022. Repeated assessments could be made with imagery available for later dates.

Extract the inundation pixels

Now that you have identified scenes before and after the heavy rains, you want to extract the pixels corresponding to inundated areas. These are pixels that changed from dry land in the before-rain scene to water in the after-rain scene.

You'll identify the changed pixels with the Detect change between two different dates functionality.

1. In the Compare Imagery window, choose the following options:
   • Check the Detect change between two different dates check box.
   • For Calculate changes in, choose Water Index.
   • For Visualize changes as a, choose Difference Mask.

   

   The Water Index option applies a mathematical formula to compute a ratio between the values of the green and near infrared bands in the selected scenes. It highlights water-covered areas and distinguishes them from dry land.

   To detect the change between the before- and after-rain scenes, the water index will be applied to each scene, and the difference between the two will be computed. The result will be shown as a mask highlighting the pixels that changed (Difference Mask).

   You'll now draw a polygon to define the area of interest.

2. Check the Draw polygon(s) to define extent check box.

   

3. On the map, draw a polygon to delineate the AOI by clicking each vertex of the desired polygon. When you are satisfied with the shape, double-click to complete.

   The result should look approximately like the image below.

   

4. Click Apply.

   After a few moments, the mask appears in bright green inside the AOI polygon. In the Compare Imagery window, two threshold sliders appear. You will set up the Positive threshold.

   

5. For the Positive slider, move the pointer back and forth to see how it impacts the display of the mask on the map.

   The difference between the water index value of a pixel before or after the heavy rains might be smaller or larger. You need to set the Positive threshold to keep only the change that is large enough to be considered significant, meaning that it represents a true case of inundation.

6. Choose a Positive threshold value where the mask corresponds to the inundation you can observe between the two images, such as 0.45.

   Note:

   You can temporarily increase the Transparency (results) slider to 50% or more to better see the imagery underneath. You can then use the swipe handle to observe the changes and refine your threshold decision.

   A value of about 0.45 seems optimal, which means that the pixels are included in the mask only if the water index difference value is above 0.45. The result should look like the following example image:

   

   The Negative slider enables you to choose a threshold for the pixels that decreased in water content, that is, went from being covered with water to dry land. There are essentially no pixels in this extent that experienced such a change, so you will keep the pointer to its minimum value.

7. Set the Negative slider to -1.

   

8. For Transparency (results), ensure that the pointer is set to 0%, as you don't need the transparency any longer.

   

   In the Compare Imagery window, for Area Decrease / Increase, the number in green indicates that the green mask within the AOI covers an area of about 15 square kilometers (your number may vary). This is the surface area that is inundated.

   

   Note:

   If some areas had gone from being covered with water to dry land, they would be indicated by the pink number.

   The Water Observation from Space (WOfS) Annual Summary dataset, also available through the app, could be useful as a supplement to this analysis, as it offers a summary of general patterns of water cover year by year. For an example of how that layer can be used, see Get started with imagery for Africa (section Monitor a lake's water level in Ghana).

Export the inundation layer

Now that you have created the mask, you'll save the resulting layer to ArcGIS Online, as you did in the previous wildfire burn scar workflow.

1. On the sidebar, click the Image Export button. If necessary, click it a second time.
2. In the Image Export window, set the following options:
   • For Export, choose Result Image.
   • For Save location, verify that Save to portal is selected.
   • For Title, type Inundations in Léré area.
   • For Description, type Inundated areas caused by the July 2022 heavy rains in the Lake Léré region in Chad.
   • For Tags, type inundations, Chad.

   

3. Click Preview.
4. If necessary, in the Sign In window, enter your ArcGIS Online or Africa GeoPortal credentials.

   A preview thumbnail appears.

   

5. Click Save.
6. When the process is complete, close the Image Export window.

   You'll verify that the layer exported successfully to ArcGIS Online.

7. In a separate web browser tab, go to ArcGIS Online. If necessary, sign in.
8. On the ribbon, click Content. In the content list, click Inundations in Léré area.
9. On the Inundations in Léré area item page, click Open in Map Viewer.

   The layer displays in Map Viewer.

   

   You could now save this web map, as you did in the wildfire burn scar workflow, and share it with relief teams to help them prioritize their effort.

Note:

Imagery is used more and more commonly to help with natural disaster management.

However, you should be aware that Landsat imagery will only capture a given location at most once a week, and its spatial resolution is only of 30 meters (per pixel), so it might not provide the most timely and detailed images for all natural disasters. Another option is to use the Sentinel-2 layer also included in the Digital Earth Africa Explorer app, which provides new images a bit more frequently (every 5 days) and offers a higher resolution (10 meters). Significantly more frequent images with much higher resolution are available through commercial vendors.

In this workflow, you extracted the pixels corresponding to the inundations caused by the July 2022 heavy rains in the Lake Léré region and you then exported them as a layer to ArcGIS Online.

Identify urban growth

Urban planners need to evaluate on a regular basis how a city is growing to better plan for urban infrastructure, such as water, electric power, and transport networks, as well as allocating services, such as schools and health care facilities. One way to quickly identify urban growth is by performing change analysis on imagery. As a GIS analyst for the city of Effiduase, Ghana, you'll use this approach to identify how the city has grown between 2017 and 2020.

Note:

There are different ways of visualizing the growth of a city. Demographers can create maps based on detailed census data, when available. Visualizing urban growth with imagery is a different approach that focuses on change in land-cover type—for instance, an area might change from being covered with vegetation to becoming built up with buildings and streets, signifying urban growth. Both approaches can be useful and supplement each other.

Set up and explore the imagery

First, you'll set up the extent and the imagery.

1. In your web browser, switch back to the Digital Earth Africa Explorer app tab.
2. In the search box, type Effiduase, Ghana and press Enter.

   

   The map updates to the new location.

3. Close the Search result pop-up.

   Effiduase (or Efidwase) is a small city northeast of the larger city of Kumasi. It is the capital of Sekyere East, a district in the Ashanti Region of Ghana.

4. Click the Zoom In button two or three times to better see the small city.

   

   To study the city's growth, you'll use the Compare Imagery tool. For the imagery, you'll use the Sentinel-2 Annual GeoMAD dataset.

   Note:

   The Sentinel-2 Annual GeoMAD dataset provides a year-by-year summary of Sentinel-2 multispectral imagery in which clouds and other small issues have been removed. Learn more about how the GeoMAD layer is created in the Digital Earth Africa documentation.

   This dataset is particularly useful to observe a phenomenon such as steady urban growth, since each layer provides a clear overview of how the landscape looked on average for a given year. This is in contrast with the individual Landsat scenes that you used earlier in the tutorial. Landsat scenes provide a snapshot of the landscape at a specific point in time, and are more appropriate to study abrupt phenomena, such as wildfires or inundations.

5. On the sidebar, click the Compare Imagery button.
6. In the Compare Imagery window, set the following options for the left image:
   • Verify that Left Image is selected.
   • For Layer, select Sentinel-2 Annual GeoMAD.
   • For Rendering, verify that Natural Color with DRA is selected.
   • Check the Select a date check box.
   • Drag the time slider to January 1, 2017.

   

   The imagery on the left updates to show the Sentinel-2 Annual GeoMAD layer for 2017.

   

   Natural Color with DRA combines the blue, green, and red bands, and shows colors close to what the human eye would see.

7. In the Compare Imagery window, set the following options for the right image:
   • Click Right Image to select it.
   • For Layer, select Sentinel-2 Annual GeoMAD.
   • For Rendering, verify that Natural Color with DRA is selected.
   • Check the Select a date check box.
   • If necessary, drag the time slider to January 1, 2020.

   

   The imagery on the right updates to show the Sentinel-2 Annual GeoMAD layer for 2020.

8. On the map, zoom out until you can see the entire city. Use the swipe handle to swipe repeatedly from left to right to compare the two images.

   

   While swiping, you can observe that the city has grown between 2017 and 2020. In these areas, vegetation (green tones) was replaced with buildings and streets (white or beige tones). In the next section, you'll extract these urban growth pixels.

Extract the urban growth pixels

You want to extract the pixels corresponding to urban growth. These are pixels where areas that were not built up in 2017—instead being covered with forest, cultivated fields, or other vegetation—have been replaced with urban buildings and streets in 2020.

You'll identify the changed pixels with the Detect change between two different dates functionality.

1. In the Compare Imagery window, choose the following options:
   • Check the Detect change between two different dates check box.
   • For Calculate changes in, choose Soil Adjusted Veg. Index.
   • For Visualize changes as a, choose Threshold Mask.

   

   The Soil Adjusted Vegetation Index, which you learned about in the wildfire burn scar workflow, will yield high values for vegetation and low values for built-up areas. The Threshold Mask method emphasizes the pixels that changed their status between the two images. These could be pixels that went from vegetation to non-vegetation (which will be masked in bright pink), or pixels that went from non-vegetation to vegetation (which will be masked in bright green).

2. Click Apply.
3. For the Threshold and the Difference sliders, move the pointers back and forth to see how they impact the display of the mask on the map.

   

   The Threshold parameter determines what value separates vegetation from non-vegetation. Only the pixels that switch from one status to the other will be masked.

   The Difference parameter determines whether you want to keep all the pixels that switched status or only the ones that also have a high difference of value between the two images.

4. Choose Threshold and Difference values for the mask to cover the urban growth you can observe between the two images, such as a Threshold of about 0.65 and a Difference of 0.

   Note:

   You can temporarily increase the Transparency (results) slider to 50% or more to better see the images underneath, and use the swipe handle to observe the changes and refine your decision.

   

   The result should look like the following example image:

   

   The bright pink color indicates areas that switched from vegetation to non-vegetation and therefore became built up. These are areas of urban growth.

   The bright green color indicates areas that switched from non-vegetation to vegetation. You can see a few isolated such spots on the map. They might, for instance, correspond to buildings that were demolished.

   Note:

   You are currently looking at urban growth in the city of Effiduase and its surroundings. If you were interested only in growth within the city, you could draw a polygon corresponding to the city's administrative boundaries. And the number in pink would indicate the urban growth surface area in the city.

5. For Transparency (results), ensure that the pointer is back to 0%, as you don't need the transparency any longer.

Note:

In this workflow, you identified urban growth using a vegetation index. Another approach would be to use a more specialized index such as the Normalized Difference Built-up Index (NDBI) or the Enhanced Normalized Difference Impervious Surfaces Index (ENDISI), which could be implemented with a more advanced application, such as ArcGIS Pro or ArcGIS Image for ArcGIS Online. You can see ENDISI used in an analysis on the same city of Effiduase proposed by Digital Earth Africa.

Also, using a vegetation index worked well in this case because in the Effiduase area, urban growth corresponds to a loss of vegetation. If you had a city developing in the middle of the desert, a vegetation index would not be effective at detecting urban growth. It is always important to understand your area of studies to choose the most appropriate analysis method.

You extracted the pixels corresponding to urban growth between 2017 and 2020 in the city of Effiduase. As a next step, you could choose to export the new layer, as you did earlier in the tutorial, and create a web map to showcase it or proceed to perform further analyses with it.

In this tutorial, you learned different techniques to perform change detection with imagery and to export the pixels extracted to use them in further visualizations or analyses. In our constantly evolving world, monitoring change is crucial. Besides delineating burn scars, inundated areas, and urban growth, there are countless other applications for change detection—from quickly assessing damage when a disaster strikes to monitoring longer-trend phenomena, such as desertification, deforestation or reforestation, the evolution of water bodies, or coastal erosion, just to name a few. Performing this type of monitoring can enable decision makers to make better-informed decisions, and assess the efficacy of new policies put in place.

Note:

If you are interested in exploring more advanced change detection workflows, see Assess hail damage in cornfields with satellite imagery, Assess burn scars with satellite imagery, Classify land cover to measure shrinking lakes, or Monitor forest change over time. These workflows require a more advanced application, such as ArcGIS Pro.

You can find more tutorials like this in the Explore satellite Imagery for Africa series.

Acknowledgements

• More about the partnership between the Digital Earth Africa program and Esri / Africa GeoPortal: Esri and its Africa GeoPortal initiative entered a collaborative partnership with Digital Earth Africa to extend the impact of the Digital Earth Africa earth observation data. This partnership has led to the development of Digital Earth Africa Explorer, the web app used in this tutorial. Esri is also making imagery data from Digital Earth Africa available as imagery services through the Africa GeoPortal. Users can use these imagery services along with the visualization and analysis tools provided on the portal. The goal is to enable users in Africa to develop their imagery skills and perform imagery analyses to help them address common challenges in Africa, such as food security, water resource management, and nature conservation.
• All imagery layers used in this tutorial are provided by Digital Earth Africa and accessed through the Digital Earth Africa Explorer app.
• The OpenStreetMap basemap is based on the Daylight map distribution of OpenStreetMap data.