You can publish multidimensional mosaic datasets and multidimensional Cloud Raster Format (CRF) datasets as image services to be shared and analyzed in a web environment.
Considerations when publishing a multidimensional image service
Before publishing a multidimensional image service, it's important to identify the best management structure for the data. The primary data management structures for managing multidimensional raster data are the multidimensional mosaic dataset and the multidimensional CRF. Each data management option has benefits and limitations, and your choice of format will depend on what you want to do with your data. For information about the best data structure for your purposes, see Managing multidimensional raster data.
In addition to the source data structure, you will need to consider the number of services to publish and the compression type to use. For example, it is recommended that you publish separate services for variables that have different temporal resolutions, for variables containing different numbers of bands, or for data displayed with a vector field renderer.
Configure processing templates
When a service contains multiple variables, you can improve data access by configuring the service with predefined processing templates.
A processing template is a raster function or chain of raster functions that is used to process and visualize raster data on the fly. Both CRF datasets and mosaic datasets can store processing templates that can be published with a multidimensional image service.
For example, you may want to store the Multidimensional Filter function as a processing template with your CRF dataset so that users can access one of multiple variables quickly. Or, you may want to create a raster function chain combining the Multidimensional Filter function with the Heat Index function, so that users can generate a heat index using a temperature variable and a relative humidity variable, both stored in the same mosaic dataset.
Publish an image service
When you are ready to publish your multidimensional image service from ArcGIS Pro, you can use the standard workflow for sharing a web layer. No additional steps are required to include the multidimensional information from the dataset.
If you want to update the source multidimensional dataset and have those updates reflected in the image service, use the following settings:
- For both CRF and mosaic datasets—On the General tab of the Share As Web Layer pane, set the Layer and Data Type option to Reference registered data. The web layer will be generated by referencing the source dataset.
- For CRF only—On the Configuration tab of the Share As Web Layer pane, click the Configure Web Layer Properties button for your image layer and check the Has live data check box. The image service will be updated when the source CRF is updated. This is not required for a multidimensional mosaic dataset.
Update an image service
Multidimensional image services published from either a mosaic dataset or a CRF can be updated without stopping the service.
To update a CRF, use the Manage Multidimensional Raster tool to append or replace slices. When appending slices, the date of the new slice must occur after that of the existing slices. If the data has a multidimensional transpose built, check the Update Transposecheck box to update the transpose for the new slices. Alternatively, use the Build Multidimensional Transpose tool to rebuild the transpose.
To update a mosaic dataset, use the Add Rasters To Mosaic Dataset tool to append new slices or overwrite existing slices. Use the Remove Rasters From Mosaic Dataset tool to delete slices. There is no limitation on the order, so you can insert or delete slices as needed.
Once the data is updated, you must refresh the service using a REST request.
When the service is refreshed, web clients consuming the service need to be refreshed, and ArcGIS Pro users need to restart ArcGIS Pro to see the updates.