ArcGIS Pro 3.2, using Presence-only prediction (MaxENT).
I wanted to create a MaxEnt model using species presence points and current climate data, and then use that model to predict occurrence based on future climate data. Both the current and future climate data are WorldClim 2.1 bioclimatic variables at 2.5 arc minutes resolution. There are 19 bioclimatic variable layers in each case.
I was able to create the model with no issues. I then ran the Prediction tool and matched up the 19 training layers to the corresponding 19 future climate layers. I received this error:
File "c:\program files\arcgis\pro\Resources\ArcToolbox\Scripts\SSForest.py", line 6899, in __init__
bands.append(int(order_band[1].upper().replace("BAND_","").replace("LAYER_",""))-1)
ValueError: invalid literal for int() with base 10: ''
I am inexperienced with Python so tried simple troubleshooting. An obvious difference between the two sets of climate data is that the current data comes as 19 separate single-band TIFFs, and the future data comes as one 19-band TIFF. When I exported the individual TIFF bands to geodatabase rasters and used those as inputs in the Prediction tool, it worked with no error.
I could not find anything in the MaxEnt documentation to suggest that multiband rasters are a problem. However, in the process of exporting rasters, which I did at the same time as clipping them to my study area, I found that a batch running of Clip Raster did not work with the individual bands as inputs - only one output was returned. Running Clip Raster (not in batch mode) for one band as input worked fine.
I guess I am just asking if anyone knows if this is expected behaviour? It would have been neat to use the WorldClim data as-is, although at least the workaround is straightforward.
