multidimensional raster data

[Hamad]

I am working with .tif files for MODIS data. I have created a mosaic from a rasters for 20 years.

The data is considered 1 band for Net Primary Productivity.

Also, I created a Date field for Years by adding new field in the attribute table.

I tried to use "Build Multidimensional Info" tools by setting the name as variable field and Date created as dimension field.

I was not able to create a temporal profile to detect the trend.

[WeiLuo]

I have the same question. have you found a solution? 

[Hamad]

Not yet. I tried to contact the author of the multidimensional exercise but was not able. Let me know if you find a solution. Also, not sure how to deal with gapfilled cells

[AngelaGray]

I think I have the same problem. I'm just using a mosaic dataset of TIF files and trying to use the Build Multidimensional Info tool but I get a 9999 error. 

I'm referencing this blog: https://www.esri.com/arcgis-blog/products/arcgis-pro/imagery/create-multidimensional-raster-data/

And watched this embedded YouTube video where they put 8 Landsat scenes from a different year that cover the same area into a Mosaic Dataset and are defining the variable that can be used in something like the multidimensional raster function. https://youtu.be/HFbTFTnsMWM

Bummer you posted this over a year ago and no one has been able to help.

[SarahOstertag]

Hi all - I have been working through a very similar issue and may have found a possible solution. I am trying to work with 37 years worth of annualized burn mosaics (Monitoring Trends in Burn Severity or MTBS), which are served as GeoTIFF thematic rasters (i.e., single-band rasters with a colormap defined, instead of RBG or other band combo, 8-bit unsigned integer data).

I first used model builder to run a raster iterator through my folder of TIFFs to build raster attribute tables and add fields including the acquisition year (extracted from my filenames), since I need this as the time dimension later. I also converted the colormap that was pre-assigned to my data to values in my attribute table so I could de-code my Raster pixel values.

Then I ran the Create Mosaic Dataset tool with a Custom Product Definition and set it to 1 band, since the "None" option defaults to multi-band. This setting allowed my output mosaic dataset to have only 1 band as required due to the input rasters having a colormap present.

After adding my rasters to the mosaic dataset, I edited the raster attribute table (attached to Footprint) to add a year field and the calculate the "ProductName" column value as "burn severity" (e.g., the appropriate variable name that corresponds to the raster pixel values in my data). This is equivalent to the "Surface Reflectance" variable field  in the video above. I was then able to run the Build Multidimensional Info tool and set Variable as ProductName, which now had "burn severity" as a drop down option, and used my calculated Year field as the Dimension (time). I also went into the mosaic raster Properties from the Catalog and changed the default to enable Time with my new Year field.

Doing all of these steps enabled me to use multidimensional tools on my custom mosaic dataset of images from a non-standard source. I sure hope this helps since it took a lot of trial and error! 🙂

[CarlaAlonso]

Hello Sarah

Could you tell me how did you do this part :

"add fields including the acquisition year (extracted from my filenames) " 

I have added the field  but I don't know how to populate it with the aq year

Thank you for these instructions

[SarahOstertag]

Hi Carla, 

I built a model builder (you could also do this in ArcPy) to clean up the raster data. I've tried to outline the steps below.

Iterator: for XYZ.tif in folder (in my case, I made 1 folder with my GeoTIFF rasters, and there were 38 rasters spanning the study period inside of this folder).  If you set this up in Model Builder or ArcPy, you can then run this same process on each of your rasters (screenshot of model attached):

1. Build Raster Attribute Table

2. Add FILENAME Field (field_name="FILENAME", field_type="TEXT", field_alias="FILENAME")

3. Calculate FILENAME field (Calculate Field): expression FILENAME: "%Name%"

4. Add Year STR Field (Add Field): field_name="Year_STR", field_type="TEXT", field_alias="Year_STR"

5. Calculate Year (Calculate Field): field="Year_STR", expression YEAR_STR = !FILENAME![-4:]

6. Add Year Date Field (Add Field): field_name="Year", field_type="DATE", field_alias="Year")

7. Calculate Field (Calculate Field): field="Year", expression Year = datetime.datetime.strptime(!Year_STR!, '%Y')")

You might need to adjust the python scripts in the Calculate Field steps to match your input raster naming conventions.

Then you can create and add these rasters to a mosaic dataset and go from there.

Good luck!