Learn about temporal resolution
First, you'll learn about temporal resolution. Temporal resolution refers to the frequency of coverage of a geographic location, usually by a class of satellite sensor platforms or an aerial imagery program. Given a specific location on Earth, how often will you get a new image of that location? This rate is also named the revisit rate. The following are a few examples of satellite and aerial programs and their corresponding revisit rates:
| Program | Revisit rate | Organization |
|---|---|---|
Mostly every 16 days (and every 8 days since the launch of Landsat 9 in September 2021) | ||
Every 5 days | ||
Near daily | Planet (commercial imagery) | |
Once a year | USDA (aerial imagery covering the continental United States only) | |
Every 1 to 2 days |
A high temporal resolution corresponds to a frequent revisit rate, and a low temporal resolution corresponds to an infrequent revisit rate.
Factors determining temporal resolution
Temporal resolution depends on several factors, such as the following:
- Swath width—As a satellite circles around the Earth, its swath width is the horizontal width of the Earth surface it captures in a single pass. Images with a larger swath width can cover the Earth more rapidly, resulting in a higher revisit rate. This is the case of Modis, with a swath width of 2,330 kilometers. This is much larger than, for instance, Landsat 8 with a swath width of 185 kilometers.
- The number of satellites in the fleet (from 1 to hundreds or more)—The satellites of a large fleet can crisscross the Earth and all take pictures in parallel, resulting in a higher revisit rate. This is the case of the PlanetScope program.
- The type of orbital trajectory chosen—For instance, a satellite with a polar orbit circles around the Earth from north to south, usually covering the entire Earth surface in about two weeks. This is the case of Landsat.
In contrast, a geostationary satellite constantly hovers over the same location, as it moves synchronously with the Earth. It can take many images each day of the same spot, but it can’t cover any other locations. This is the case with many weather satellites.
- Frequency of an aerial imagery program—An aerial imagery program requires flying aircrafts regularly. The frequency of these flights might depend on budget constraints and other factors. For instance, NAIP (National Agriculture Imagery Program) gets a revisit rate of once a year only. In recent years, the availability of drones, which are significantly cheaper than aircrafts, has enabled many organizations to produce imagery for their areas of interest with a much higher revisit rate, for instance, once a month.
There are often trade-offs between the different types of resolution. For instance, images with a large swath width have a higher temporal resolution, but they also have a lower spatial resolution, which means that they depict the features on the ground with fewer details. One pixel of a wide-swath Modis image represents 250 x 250 meters on the ground (or more), whereas one pixel of a narrower-swath Landsat 8 image represents a square of 30 x 30 meters on the ground.
Based on your project requirements, you might need imagery with a high temporal resolution, or a lower temporal resolution might be sufficient. It is important to choose your imagery based on your specific goals. In the rest of the tutorial, you will learn about three use cases that each have different temporal resolution requirements.
You learned about temporal resolution, revisit rate, swath width, and a few other essential facts related to the topic. Next, you'll look at the first example of temporal imagery.
Assess the growth of a refugee camp
In 2017, due to intense violence in Myanmar, the Rohingya minority started fleeing toward Bangladesh, taking refuge in the area of Cox’s Bazar. Within the span of a few months, the small Kutupalong refugee camp grew to a massive size.
Humanitarian groups were able to use satellite imagery to monitor this dramatic situation and improve the disaster response.
Observe images captured over time
Sentinel-2 satellite images show how the refugee camp grew. You'll view the images as a video.
Note:
Sentinel-2 is a satellite mission from the European Space Agency. It was launched in 2015 and produces imagery with a spectral resolution of 13 spectral bands, several of which have a spatial resolution of 10 meters. The images cover nearly the entire landmass of Earth, and every place is captured with a temporal resolution of at least every five days. Sentinel-2 images are freely available and can be downloaded through the Copernicus Data Space Ecosystem.
- Play the following video and observe how the refugee camp (in white tones) grows rapidly.
Note:
The video has no sound.
A series of images taken over time in a specific location is referred to as a time series. A time series is a powerful tool to observe change over time. Due to the rapid growth of the Kutupalong camp, a relatively high temporal resolution is needed to monitor the situation effectively.
- Watch the video a second time. Pay attention to how some images contain clouds.
For instance, this is the case for the image captured on 05/09/2017.
A few clouds might not prevent an assessment of the changes on the ground. However, in Bangladesh during the monsoon season (in the summer months), there is frequently an extensive cloud cover that obstructs the view significantly or even completely. The following image is an example of an image of the Kutupalong camp area taken on August 17, 2017, a very cloudy day:
This image is useless to monitor the growth of the Kutupalong refugee camp and was not included in the time series. When deciding on the temporal resolution needed for a specific study, the amount of cloud cover in the area should be taken into account. If too many images are covered with clouds and can't be used, it will cause gaps in the time series.
- In the video, look at the images captured on 05/09/2017 and 09/26/2017.
Because there were several unusable cloudy images in the summer months, the video seems to display a jump in the camp growth between these two dates. As it is, this time series is still very useful for humanitarian organizations. However, if fewer gaps are needed, it might be useful to switch to a satellite imagery program with a higher temporal resolution, such as PlanetScope instead of Sentinel-2. This would increase the chances that enough images were captured on sunny days with a low cloud cover, hence reducing the time gaps.
Note:
Another approach could be to use synthetic aperture radar (SAR) imagery, a type of radar satellite imagery that is not affected by clouds. This approach is out of scope for this current tutorial, but you can learn more about it in the Getting started with SAR satellite imagery series.
You learned that satellite imagery can be used to monitor the growth of a refugee camp to help improve the disaster response. You learned about time series and the relation between temporal resolution and cloud cover. Next, you'll learn how satellite imagery can be used to prevent deforestation.
Prevent illegal deforestation
The Prey Lang Wildlife Sanctuary in Cambodia is the largest lowland evergreen forest remaining in mainland Southeast Asia. Unfortunately, there are many illegal logging attempts in the area. Some nonprofits have been using satellite imagery to try to stop illegal deforestation before it happens. They monitor large areas of the forest and its surroundings regularly to spot new roads being built and potential activity on these roads. If these roads were not planned by the local government, they might be a sign that some logging companies are trying to reach new areas to log trees illegally. Protecting the sanctuary is important not only for wildlife, but also for humans, since local communities rely on the forest for food, water resources, and income, while the forest also helps regulate the global climate by storing large amounts of carbon.
In ArcGIS Pro, you'll look at several images taken in that area to examine the situation.
Download and open the project
First, you'll download the project containing all the data needed for the remainder of the tutorial and you'll open it in ArcGIS Pro.
- Download the Temporal_resolution package.
A file named Temporal_resolution.ppkx is downloaded to your computer.
Note:
A .ppkx file is an ArcGIS Pro project package and may contain maps, data, and other files that you can open in ArcGIS Pro. Learn more about managing .ppkx files in this guide.
- Locate the downloaded file on your computer.
Note:
Most browsers download to your computer's Downloads folder by default.
- Double-click Temporal_resolution.ppkx to open it in ArcGIS Pro. If prompted, sign in with your ArcGIS account.
Note:
If you don't have access to ArcGIS Pro or an ArcGIS organizational account, see options for software access.
The project opens. It contains two maps named Prey Lang and Tangier Island, and the Prey Lang map is currently selected (blue tab).
Examine images in a time series
You'll examine the Prey Lang map, the area of interest (AOI) it focuses on, and the images it contains.
- Observe the Prey Lang map.
The map is focused on an area near the eastern side of the Prey Lang Wildlife Sanctuary. The hatched area on the left represents the Prey Lang Wildlife Sanctuary grounds, which are protected and should not be deforested. For context, in the following overview map, you can see the entire Prey Lang Wildlife Sanctuary (shown in orange) and the AOI on the eastern side of the sanctuary (shown as a small black rectangle).
- In the Contents pane, review the Time series group layer.
The group layer contains four satellite images representing the AOI and taken at different points in time. Together they form a time series. They are PlanetScope (PS) satellite images produced by the Earth-imaging company Planet Labs.
Note:
PlanetScope images are produced by Planet Labs. PlanetScope is a collection of hundreds of satellites that have been deployed from 2014 onward and produces imagery with a 3 meter per pixel resolution. The images cover nearly the entire landmass of Earth, and each location is captured almost daily.
Currently, the PS - Pre-deforestation 01-04-2020.tif image captured on January 4, 2020, is turned on and displays on the map. At that date, you can see that the area is entirely forested and mostly unperturbed. On the eastern side, some newly traced roads are a sign that legal forestation is going to take place soon on unprotected land.
- In the Contents pane, turn on the second image of the series, PS - Legal deforestation starts 04-29-2020.tif, taken on April 29, 2020.
- Turn off the Prey Lang Wildlife Sanctuary layer so that you can better see the image.
You'll compare the two January and April images by swiping between the two layers.
- If necessary, in the Contents pane, click PS - Legal deforestation starts 04-29-2020.tif to select it.
- On the ribbon, click the Raster Layer tab. In the Compare group, click Swipe.
- On the map, drag repeatedly from top to bottom or side to side to peel off the April image and reveal the January one.
As shown in the following example image, you can see that in April, (1) the legal deforestation has started at the lower right corner. (2) A new legal road has also been carved along the sanctuary boundary. Finally, (3) in the middle of the image, some of the small preexisting roads are a bit more visible than before, but not yet a clear cause for concern.
The images collected over the next few months are not provided in this tutorial, to limit the size of the data you had to download. They presented a similar situation on the ground, with legal deforestation continuing, but no clear sign of illegal activities.
You'll examine the next image.
- In the Contents pane, turn on PS - Suspicious road activity 02-02-2021.tif, taken on February 2, 2021, and select it.
- On the map, swipe between the second and third image.
(1) The legal deforestation started earlier continues. (2) Closer to the sanctuary, some deforestation has also started, but it is east of the sanctuary, so it is legal. More concerning, two of the small preexisting roads (3 and 4) seem to have been enlarged and prolonged to reach the sanctuary, and there is now a large clearing (5) where these two roads meet.
Local environmentalists used to detecting signs of illegal deforestation activities are alarmed, as these secondary roads and the clearing could be used to start logging trees illegally in the sanctuary and transport them off site. They decide to obtain a more detailed image—that is, with a higher spatial resolution—for that area to better understand the situation.
Use the tip and cue approach
The tip and cue approach consists of regularly monitoring an AOI over time with one type of satellite imagery, then when abnormal change is detected on the ground (tipping the analyst off), more detailed imagery is brought in (cued) to have a closer look. In this case, the PlanetScope imagery you've seen so far serves as the tip, thanks to its high temporal frequency and 3-meter spatial resolution, while the new SkySat image will act as the cue with its more detailed 0.5-meter resolution.
Next, you'll look at the more detailed SkySat image that was taken a few days after suspicious activity was detected.
- In the Contents pane, turn on the SkySat 02-05-2021.tif image.
Note:
SkySat images are produced by Planet Labs. SkySat is a collection of about 20 satellites that was deployed from 2013 onward and produces imagery with a 0.5-meter resolution and 4 spectral bands. SkySat satellites can be actively maneuvered to capture imagery from any location on Earth.
- On the ribbon, on the Map tab, in the Navigate group, click the Explore button to exit the swipe mode.
- Zoom in and out and pan through the SkySat image to observe the two expanded roads and the clearing.
In the SkySat image, you can see many more details than previously. You can confirm that the two small roads are now clearly traced, starting from the clearing and reaching into the sanctuary. As you zoom in to the clearing, you can see logs that have been temporarily stored there, and two trucks loaded with logs are present.
All these signs indicate that trucks are transporting logs from the preserve. At that point, the local authorities were able to intervene to stop this illegal activity before the nearby sanctuary area had been irreparably damaged by deforestation. This tip and cue approach can continue in the following months and years to ensure that no new illegal activity develops.
- In the Contents pane, turn on PS - Today 03-02-2026.tif and turn off SkySat 02-05-2021.tif.
In 2026, the non-protected part of the AOI has been largely deforested and replaced with agricultural fields, as was legally authorized (1). In contrast, the protected land of the sanctuary remains intact (2). In fact, you can see that the new agricultural fields stop right along the border of the sanctuary (3).
In this type of monitoring, it is essential to have a temporal resolution that is high enough because you want to identify the new roads and other signs of illegal activity as soon as possible so that illegal deforestation can be stopped before it is too late. Furthermore, similarly to the case of the Rohingya refugee camp, some of the images might have a heavy cloud cover, so a high temporal resolution is all the more needed. Hence, PlanetScope, with a revisit rate of about a day, is a good choice.
Tip and cue is a powerful approach for monitoring an area on a regular basis. It is relatively conservative in its data use and cost, since the very detailed imagery, which is more expensive and requires more storage space, is only used when a cause for concern is suspected. In real life, the area monitored could be much larger. For instance, it could cover the entire surface of the Prey Lang Wildlife Sanctuary.
Note:
Learn more about this use of satellite imagery to prevent deforestation in the Prey Lang Wildlife Sanctuary in the article Tackling Deforestation in Cambodia.
Next, you'll learn about using an imagery time series that spans several decades to study an island that has been shrinking over time.
Monitor a shrinking island
You are interested in studying Tangier Island in Virginia, United States, an island that has been shrinking over time due to erosion and sea level rise. Using satellite imagery to visualize this evolution can enable the local community to better understand what areas are most susceptible to shrinkage and how they could best be preserved in the future.
Since your goal is to observe a phenomenon that has been unfolding over decades—as opposed to months or days, you don't need imagery with a high temporal resolution. It is enough for you to examine one image every year, or perhaps even one image every five years. However, you do need a satellite imagery type available for a long time range: it should offer a large archive accumulated over several decades. A good choice is Landsat, since it is a program that has run continuously since 1972. To examine the evolution of Tangier Island, you'll use a Landsat time series that contains roughly one image every five years from 1982 to 2025.
Note:
Landsat is a satellite mission from USGS and NASA launched in 2013. It produces multispectral imagery with 11 spectral bands, most of them with a 30-meter spatial resolution. The images cover nearly the entire landmass of Earth, and every place is captured mostly every 16 days. Landsat is the longest-running satellite imagery acquisition program, providing more than five decades of continuous earth observation data. Landsat images are freely available. Learn how to download your own Landsat imagery.
Explore a multidimensional dataset
Next, you'll switch to the Tangier Island map to explore the Landsat time series.
- Click the Tangier Island tab to switch to the second map.
The map is centered on Tangier Island (circled in yellow in the following example image) and a few small islets in its surroundings.
- Review the Contents pane and locate the Tangier Island time series.crf layer.
In this case, this single dataset contains the entire imagery time series. This dataset type is a multidimensional raster dataset or image cube. It is composed of a stack of regular satellite images organized chronologically—as shown on the following diagram. It is referred to as an image cube, because it is considered three dimensional, with Latitude and Longitude forming the first two dimensions, and Time forming the third one. It is stored in the .crf format.
A multidimensional dataset or image cube. This is a way of organizing a time series that offers more advanced capabilities than just storing each image separately, as you saw in the Prey Lang example. For instance, you can easily browse through the time dimension of the dataset, where every image is called a time slice. You'll learn how to do that now.
- In the Contents pane, click the Tangier Island time series.crf layer to select it.
- On the ribbon, on the Multidimensional tab, locate the Current Display Slice group.
This tool group contains information about the time slices in the dataset, as well as buttons to navigate from slice to slice.
- For StdTime (which stands for Standard Time), expand the drop-down list, and review the list of time slices.
Currently, the first slice, dated from April 17, 1985 (1985-04-17), is selected and displays on the map. The dataset contains 8 time slices, made of 8 Landsat images taken about every 5 years from 1985 to 2025.
- In the drop-down list, select the last time slice dated 2025-10-24.
The image on the map updates. In the span of 40 years, Tangier Island and surrounding islets have significantly shrunk. To better see the difference, you'll turn on a layer delineating the 1985 land boundaries.
- In the Contents pane, turn on the Land boundaries 1985 layer.
Tip:
Ensure you keep the Tangier Island time series.crf layer selected. Otherwise, you'll lose access to the Multidimensional tab.
- On the Multidimensional tab, for StdTime, go back to the 1985-04-17 time slice.
- On the map, confirm that the 1985 boundaries (in red) correspond to the current land boundaries.
In the middle of Tangier Island, the area delineated in green corresponds to an area covered with shallow water with underlying sand banks and a number of buildings on stilts. It might be more or less visible in the different time slices.
- On the Multidimensional tab, for StdTime, click the Step Forward button repeatedly to cycle through all the time slices.
From 1985 to 2025, you can see the coast receding progressively on the main island and the small islets. This is due mostly to erosion and sea level rise.
- Stop at the 2025-10-24 time slice and observe how much has changed compared to the 1985 boundaries.
For instance, the areas (1), (2), and (3) highlighted in yellow in the following example image have lost significant amounts of land. The southern tip of the island (4) has lost land, but it has also moved to the northeast direction, due to shifting sand banks—showing in cream white tones in the imagery.
- Zoom in to various locations to better see the evolution.
Note:
Landsat imagery doesn't have a very high spatial resolution: 1 pixel represents 30 meters on the ground. So you will not see many details on the ground, but you can still see the general land shapes quite clearly.
- Zoom back out to see the entire AOI.
- On the Multidimensional tab, click the Play Slices Along StdTime button.
The time slices display one after the other, forming an animation.
- When you are done observing the animation, click the Pause button to stop it.
- In the drop-down list, choose the 2025-10-24 time slice.
Visualize change over time
Next, you'll perform a short analysis to better understand when specific locations in the Tangier Island area went from being emerged land to being submerged. You'll first generate a layer that shows whether each pixel corresponds to land or water.
You'll do that by applying the Normalized Difference Water Index (NDWI) to your multidimensional dataset, using the Band Arithmetic raster function. The NDWI index computes a ratio between the Near Infrared and Green bands of an image. This is possible because every time slice of your dataset is a Landsat image that contains several bands: Blue, Green, Red, Near Infrared, and Shortwave Infrared.
Note:
While discussing image bands and spectral indexes is out of scope for this tutorial, you can learn more about them in the tutorials Explore imagery: Spectral resolution and Assess hail damage in cornfields with satellite imagery.
- On the ribbon, on the Imagery tab, in the Analysis group, click the Raster Functions button.
- In the Raster Functions pane, search for Band Arithmetic. Click the Band Arithmetic raster function to open it.
- Set the following Band Arithmetic parameters:
- For Raster, choose Tangier Island time series.crf.
- For Method, choose NDWI.
- For Band Indexes, type 4 2.
In the Tangier Island time series.crf dataset, 4 and 2 correspond to the Near Infrared and Green bands, respectively.
- Click the Create new layer button.
The new layer appears in the Contents pane and on the map. It is a raster layer symbolized with a black-to-white color ramp, with the negative values (black or darker tones) representing land and the positive values (light gray or white tones) representing water.
You want to simplify the water representation, so you will create a binary layer where 1 represents water, and 0 represents no water. You'll do that with the Remap raster tool.
- In the Raster Functions pane, click the Back button.
- Search and open the Remap raster function.
- Set the following Remap parameters:
- For Raster, choose Band Arithmetic_Tangier Island time series.crf.
- In the table, edit the first row to read -1 0 0, and the second row to read 0 1 1.
Any pixel with a negative value (-1 to 0) will become 0 (land), and any pixel with a positive value (0 to 1) will become 1 (water).
- Click the Create new layer button.
The new layer appears. For now, it appears entirely black; you'll fix that in a moment. But first, you'll rename it.
- In the Contents pane, click the Remap_Band Arithmetic_Tangier Island time series.crf layer and click it again to enter the edit mode. Change the name to Land or water and press Enter.
Next, you'll change the display settings.
- Confirm that the Land or water layer is selected.
- On the ribbon, on the Raster Layer tab, in the Rendering group, expand the Stretch Type list and choose Minimum Maximum.
Tip:
When you have a raster layer with only two values (0 and 1), the default Percent Clip stretch doesn't display it properly, which is why you are switching to the Minimum Maximum stretch. Learn more about the Stretch option.
The layer updates to show the land in black and the water in white.
When raster functions are applied to a multidimensional layer, they are applied to all the time slices.
- On the ribbon, on the Multidimensional tab, click the Step Forward button a few times to verify that all the time slices appear in black and white.
Create a temporal profile graph
Next, you'll create a temporal profile graph that shows how the pixel values have changed over time at specific locations.
- In the Contents pane, right-click Land or water, point to Create Chart, and choose Temporal Profile.
New panels for the chart appear. You'll zoom in to the north of Tangier Island.
- On the ribbon, on the Map tab, in the Navigate group, click Bookmarks. Under Tangier Island Bookmarks, choose the North area bookmark.
You'll add a first point to the chart.
- In the Chart Properties pane, under Define an area of interest, click the Point button.
- On the map, click at the extreme north within the 1985 boundaries, at the location shown in the following example image:
The spectral profile for this point appears on the chart. It shows for every time slice whether its pixel value was 0 (land) or 1 (water).
Note:
The color for the point and graph line is assigned at random and may vary.
- Point to the first point in the graph that has a value of 1 (water) and read the date in the pop-up.
The first time that this point was recorded as being under water was in 1990.
You'll add two more points to the map.
- On the map, add points 2 and 3, as shown in the following example image:
- Point to the graph to find out the first year that points 2 and 3 were under water.
It was in 1995 and 2005, respectively.
Note:
If you chose slightly different points on the map, you might get different dates.
The temporal profile chart gives a precise understanding of how the coast has receded over the years in that area.
Using imagery time series to visualize the evolution of Tangier Island can enable the local community to better understand what areas are most susceptible to shrinkage and how they could best be preserved in the future. The article The race to save Tangier Island from erosion, sea-level rise describes the efforts under way to develop a resilience plan that will include bolstering the shoreline against erosion and building protective infrastructure to accommodate rising waters.
Multidimensional imagery datasets are a powerful format to perform advanced analysis and identify long-term trends. See, for instance, the Monitor forest change over time tutorial, where you use the LandTrendr algorithm applied to a Landsat multidimensional dataset to study the evolution of a forest over several decades. You can also learn how to create your own multidimensional dataset in the Get started with multidimensional multispectral imagery tutorial.
In this tutorial, you learned about temporal resolution and became familiar with related imagery concepts, such as revisit rate, time series, tip and cue approach, multidimensional imagery datasets, the trade-offs between different resolution types, and the importance of imagery archives spanning a long time range. You examined and manipulated satellite imagery of different temporal resolutions in ArcGIS Pro, looking at three use cases: monitoring the growth of a refugee camp for humanitarian purposes, using imagery to prevent illegal deforestation, and observing how an island has been shrinking over time.
You can learn about other types of imagery resolution, such as spatial and spectral, in the Explore imagery resolution tutorial series.
You can find more tutorials in the tutorial gallery.
