Data formats supported by Spatial Analyst

The ArcGIS Spatial Analyst extension accepts as input all relevant raster, feature, and table formats that are supported by ArcGIS. While many more input formats are now natively supported in Spatial Analyst, only certain formats are supported for output.

Raster data

Spatial Analyst can operate on raster data that is either file based or in a geodatabase.

Input raster data formats

In ArcGIS 10, all supported raster formats are now natively processed by the Spatial Analyst engine.

Being able to read the supported raster formats directly eliminates the previous conversion step that created intermediate temporary files when executing operations on certain data formats. The benefits obtained from this improvement are reduced processing times as well as reduced disk space consumption.

Output raster data formats

The supported output raster formats are the following:

The location and name you specify for the output raster determines the format it is created in.

ArcGIS supports fewer raster formats for output than it does for input, and Spatial Analyst specifically supports fewer output raster formats than ArcGIS does in general. However, once an output raster has been created, it can easily be converted to another format with the Copy Raster tool.

Feature data

Input feature data

All supported feature classes for ArcGIS 10 are natively processed by the Spatial Analyst engine as input.

Output feature data

For output, the supported feature types are point, multipoint, polyline, and polygon, depending on the particular Spatial Analyst geoprocessing tool being used.

The feature class can be created in a geodatabase (file, personal, or ArcSDE), or if in an output folder, as a shapefile or a coverage.

Creation of temporary (scratch) data

When running Spatial Analyst tools that operate on raster or feature data, there may be some situations that cause the software to create intermediate or scratch files in the process of the operation. After the operation completes and the output has been created, the scratch files are deleted.

The location where these scratch files get created can be controlled with the Scratch workspace environment.

Limitations of data formats

Certain data formats have limits that may be encountered when creating output. Following are some topics that discuss these format limitations:

Multiband raster data

When a multiband raster is used as input, most Spatial Analyst tools operate only on the first band.

The exceptions are certain tools in the Multivariate and Extraction toolsets which do process each of the bands in a multiband input and can create a multiband output. Consult the individual tool references for more specific details.

Mosaic datasets

Any Spatial Analyst tool that supports raster as input can operate on mosaic datasets.

Mosaic datasets can be extremely large (hundreds of thousands of rows and columns of cells), so running any analytical tools on them can consume a lot of system resources and take a lot of execution time. It is important to consider limiting the analysis extent or data resolution in your operations.

Handing multiple resolutions of mosaic datasets

Mosaic datasets can contain multiple resolutions of raster data. It is important to use the appropriate cell size for the type of analysis being done.

Cell size ranges are used to determine what rasters are processed to create the dynamically mosaicked image from the mosaic dataset.

Mosaic dataset raster limits

Mosaic datasets will have a limit on the number of individual rasters that can be analyzed at one time. If this limit is exceeded, areas of NoData may appear in the output.

This limit is controlled by the Maximum Number Of Rasters Per Mosaic property. The default is 20.

Mosaicking rules

Also important to consider are the mosaicking rules, which determine how overlapping individual rasters are handled. More details on how to control the mosaicking is available in the following topic:

To control which layers are used and how they are collated, try creating a mosaic layer.

Other mosaic dataset properties also influence how the individual rasters are handled in the mosaicking process. These include Allowed Mosaic Methods, Default Sorting Order, and Default Mosaic Operator. Some properties, such as Default Resampling Method, will affect the pixel quality.

Geoprocessing analysis environments

In most cases, you will need to define one or both of the following geoprocessing environments to control the analysis:

  • Output Extent

    Use this environment to control the area of the output raster.

  • Cell Size

    Use this environment to control the resolution of the output raster.

Image services

You can perform Spatial Analyst operations only on certain types of services, such as image services or WCS services. Other types of services, such as map services or the layers within them, cannot be used as input to Spatial Analyst tools.

Consult the following documentation resources for more information on using services.

Processing considerations when using image services

Image services are created from raster datasets and mosaic datasets. Consequently, all of the issues in the previous section pertaining to mosaic datasets will also apply when using image services as input.


Pay attention to the Maximum Size Of Requests property, which is used to limit the numbers of rows and columns that can be analyzed at one time. If this limit is exceeded, the operation will not complete.

When working with image services, you should first use the Make Image Server Layer tool to create an image service layer. This allows you to define many of the properties, including the extent (template) and cell size.

Geoprocessing analysis environments

As with mosaic datasets, the Output Extent and Cell Size geoprocessing environments should be defined to properly control the analysis.

Related Topics