Help with GIS probably maps

[EP]

My predictor data were collected at sites (ie points). The GIS operations in HyperNiche require an .asc file with this information, but .asc files from GIS are raster data. Can anyone point me in the right direction for understanding what these files represent, how we create them, and how to use the output? The description in the online manual and other topics posted here are elucidating my problem.

Please tell me what I'm missing. Thanks!

[Peter Nelson]

To make a map, you must have some sort of grid of values in the area
of interest for each predictor in the NPMR model. By grid of values I
mean a raster, .asc, .txt, etc.. This naturally limits the types of
predictors one can use to build a model as many variables aren't
available as GIS layers. For example, you may have measured a bunch of
stuff on your sites, such as soil N, pH, slope and elevation but only
have layers for slope and elevation. You could only make a map with
the models that only have slope and elevation for predictors. You can
try to make maps of your predictors by making models for the missing
predictors, making a map (eg. raster, .asc file, etc.) but that can
get laborious.

Hope that helps.

Peter

> --
> You received this message because you are subscribed to the Google
> Groups "HyperNiche and NPMR" group.
> To unsubscribe from this group and stop receiving emails from it,
> send an email to hyperniche+...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>

[Elizabeth Perotti]

Thanks Peter. Still confused on some points though.

I have all of my data as predictors as point data. Are you saying that I could create a raster for a variable (e.g. pH) (from kriging, etc), export that as a .txt file. This would then have the lat/long for each cell (right?) along with the values. I could then bring that into HyperNiche and run a model (with 1 predictor?) and it would give me probability information? Still confused on what to input, how and what is the output.

--
You received this message because you are subscribed to a topic in the Google Groups "HyperNiche and NPMR" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/hyperniche/i9dXATbeZTY/unsubscribe.
To unsubscribe from this group and all its topics, send an email to hyperniche+...@googlegroups.com.

[Bruce McCune]

Elizabeth, The GIS application is designed for situations where you have grids of predictors, not just your measured predictors associated with the response.

For example, if you have measured pH as a predictor at each point, to create a regression-based map (using NPMR or any other regression method), you would need a grid of pH as a basis for mapping estimates of the response. If you DON'T have such an input grid (and commonly we don't have that for the best predictors), then you are limited to mapping point data with GIS, rather than making a "filled in" map. A point-based map of a response might look like different size symbols depending on the size of the response variable, or a gradation in color of symbols depending on the size of the response variable. To go this route, you don't need to use the GIS facilility in HyperNiche -- just save the estimates of the response variable for each point (i.e. with Response Points or Evaluate Selected Model).

The alternative, as you suggest, is to first model your predictors spatially, but unless you have a lot of data points, it will be difficult to do that convincingly.

Hope this helps.

Bruce McCune