Create a multidimensional mosaic dataset from a set of time series images

This workflow demonstrates how to use imagery in a time series to create a multidimensional mosaic dataset.

Multidimensional mosaic dataset requirements

A multidimensional-aware mosaic dataset contains three fields:

  • A text field named Variable that stores the variable represented in each image.
  • A text field named Dimensions that stores the comma-separated dimension names.
  • One or more fields that define the dimensions.

The Variable and Dimensions fields can be generated using the Build Multidimensional Info tool. To create a mosaic dataset and prepare it for the Build Multidimensional Info tool, use the workflow below.

  1. Create a mosaic dataset using the Create Mosaic Dataset tool.
  2. Use the Add Rasters to Mosaic Dataset tool to add the time series data to the mosaic dataset using the Raster Dataset raster type.
  3. Add the mosaic dataset to the map.
  4. In the Contents pane, right-click the Footprint sublayer, and click Open Attribute Table.
  5. Make sure the table contains the information necessary to create the Variable field. There are two options:
    • If there is only one variable represented in the mosaic dataset, move on to step 7. You specify the variable name in step 8. For example, if the mosaic dataset contains a stack of satellite images collected over time, where all images are scenes from the satellite and have been processed in the same way, you can skip this step.
    • If there are multiple variables represented in the mosaic dataset, you need to specify the variables represented in each image. To do this, create a new text field, or you can choose to populate an existing text field (for example, the Tag field) with the name of the variable that is represented in each image. Select the images in the table that represent the same variable and use the Calculate Field tool to populate the text field with the name of the variable. Repeat this process until all the rasters in the mosaic dataset have a record indicating the variable represented in the raster.
  6. Make sure the table contains the information necessary to create the Dimensions field. There must be a field indicating the date or time that the raster was captured. For Landsat and other satellite sensors, this field is often called AcquisitionDate. If the mosaic dataset contains images collected along a dimension other than time (for example, depth), there must also be a field indicating the dimension measurement for each image. If these fields do not exist, generate new fields and populate them accordingly.
  7. Save all edits and close the attribute table.
  8. Open the Build Multidimensional Info tool.
    1. Set the mosaic dataset as the input for Mosaic Dataset.
    2. Use the Variable Field drop-down menu to select the field that stores the variable information. This is the field you created in step 5 above. If you skipped step 5 because you only have one variable in your dataset, you can type the name of the variable here. For example, type Landsat7 if all of the rasters in the mosaic dataset are Landsat 7 ETM scenes.
    3. Use the Dimension Fields drop-down menu to select the field or fields that store the dimension information. These are the fields you verified or created in step 6.
    4. Click Run.

When the tool is finished running, the Variable and Dimensions fields will be visible in the Footprint attribute table, and the Multidimensional Info tab will be available on the Mosaic Dataset Properties dialog box.

Landsat 8 Example

The following workflow details the steps for creating a multidimensional mosaic dataset containing Landsat 8 Operational Land Imager (OLI) surface reflectance images, collected over multiple dates.

  1. Create a mosaic dataset using the Create Mosaic Dataset tool.
  2. Use the Add Rasters to Mosaic Dataset tool to add the Landsat 8 images to the mosaic dataset.
    1. Set the Raster Type parameter to Landsat 8.
    2. Set the Processing Templates parameter to Surface Reflectance.
    3. Set the Input Data parameter to Folder and click Browse. Navigate to the directory that contains the imagery. Typically, each Landsat image is stored in a main folder containing multiple files and subfolders. Hold the Shift key to select the main folder for each the Landsat image. Click OK.
    4. Under Raster Processing, check the box to Calculate Statistics.
    5. Under Mosaic Post-Processing, check the box to Update Overviews.
    6. Keep all other parameters as defaults and click Run.
  3. Open the Build Multidimensional Info tool.
    1. Set the mosaic dataset as the input for Mosaic Dataset.
    2. For the Variable Field parameter, select the ProductName field. The variable information will populate with Surface Reflectance.
    3. For the Dimension Fields parameter, select the AcquisitionDate field.
    4. Click Run.

When the tool is finished running, the Variable and Dimensions fields will be visible in the Footprint attribute table, and the Multidimensional Info tab will be available on the Mosaic Dataset Properties dialog box.

Related topics