Spatial Thinning/Localities Removed

[Lukas Bowler]

I am currently working on a niche model for Hot Springs National Park in Arkansas, USA, and I'm running into an issue on the Obtaining Environmental Data component. 

At the moment, I am running user-generated occurrence data, but Hot Springs National Park is relatively small in comparison to a lot of other areas used for niche modeling. Since it's smaller, the program is automatically removing points that are in the same grid cell. This is the message that shows up in the log console: 

Unfortunately, this is happening with all of the species I'm using. It's cutting the points available from 50+ to less than 5 in many cases. For some (especially for specific species found along one stream), 50+ occurrences are being reduced to just 1. 

Is there something I can do to prevent this from happening? I intend to use spatial filtering to minimize accidental sampling biases, but with this happening, I won't have enough data points to even begin to model it. 

Also, sorry if some of my wording doesn't make sense. I am still relatively new to niche models as a whole. I've also read through the vignette, which was extremely helpful.  

Thanks in advance!

Lukas

[Wallace Ecological Modeling App]

Hi Lukas,

Your wording makes perfect sense. This is happening because several of your occurrences are falling into the same environmental grid cell. Wallace removes these points to reduce bias when modeling.
To alleviate this issue, you should increase the resolution of your environmental data by lowering the arcmin. It seems counterintuitive, but the lower the distance value, the higher resolution (so 2.5arcmin is a much higher resolution than 10arcmin). Just a heads up, the highest resolution (30arcsec) will only download one tile, instead of global data, due to the large size and RAM usage. Any points outside the tile will be removed (which may be what is happening, but I don't think so based on the error message). If you are using the 30arcsec option and still too many points are being removed, it may be in your best interest to look externally for high res microclimate data and not WorldClim. These can be uploaded using the User-Specified env data module.

Bethany

[Lukas Bowler]

Thanks for the response! It was very helpful. Since I was already using the 30 arcsec resolution, I have spent some time looking into User-Specified environmental variables. I have found a few that would work for a preliminary model. Unfortunately, when I upload them, my screen turns grey and the log window doesn't give any message to indicate whether it's uploaded, failed, or had any issue. What could this be caused by? I have attached two images showing the before and after of my screen when attempting to upload the environmental data. 

Thanks in advance!

[Wallace Ecological Modeling App]

Is there anything printed in the Rstudio console?

[Lukas Bowler]

This is the error I received. Does it have something to do with the grid/cell size? If so, since I'm getting my data from a third-party organization, is there anything I can do to change that?

[Wallace Ecological Modeling App]

All environmental rasters must have the same extent and resolution (cell size). You'll need to synchronize them to use within Wallace. Please see the Module Guidance for User-specified Environmental Data for more info on the requirements for user data.