I'm a third-year undergraduate biology major with a focus in ecology. Referencing the methods laid out in A quantitative approach developing regional ecosystem classifications (Host et al 1996) I would like to create ecoregions for the Green Mountain National Forest. Obviously these would be very detailed compared to global ecoregions which distinguish between much more heterogenous climates. The first step is to classify the soil data into polygons. These will then be laid over climate polygons to define the ecoregion polygons.

I'm working with an abridged version of SSURGO data from ArcGIS. In their description on the website they say that they've only included 130 of the most useful variables. These include drainage, particle size, and taxonomic classification, slope, bedrock depth, runoff, erosion, and geomorphic description. I'd like to begin working with the former two (perhaps adding other variables later on) to create the soil polygons. Host et al (1996) used drainage, surface and subsurface texture, and particle size. My professor suggested k-means clustering but the data I have are actually categorical! Drainage is defined as “excessively drained”, “somewhat excessively drained”, …. , and “very poorly drained”. Particle size is defined as “coarse-loamy”, “coarse-silty”, “loamy”, etc.

Can I attempt to make the soil polygons using only those two categorical variables?

If not, what others should I include?

Another possibility, could I just use the USDA taxonomic classifications which are included?

Host et al (1996) don’t use them so I imagine there’s some reason not to, but I wonder if that reason applies to my work.

Also, any suggestions for a specific method to follow for the code in R or method in ArcGIS?

I would like to use comprehensive SSURGO data rather than the abridged version but at this point its not feasible because you can only download 100 thousand acres at a time from the USDA website. I’ve been told that I could write a Python script to do that and I’d like to investigate that in future but for now I’d like to work with the data I already have. I'm really just defining the methods for a first draft of the project, so really I’m looking for a heuristic to improve on in future. Please, do let me know what other information I should provide in order to make it easier to provide a thorough answer.

I'm an early-intermediate R-user and I'm learning ArcGIS as I go along.

Here is link to a csv file of an attribute table of SSURGO data. First 100 rows only.

Host, GE, PH Polzer, DJ Mladenoff, MA White, TR Crow. 1996. A quantitative approach to developing regional ecosystems classifications. Ecological Applications 6: 608-618.

closed as too broad by whyzar, Andre Silva, PolyGeo Mar 19 '17 at 11:43

Please edit the question to limit it to a specific problem with enough detail to identify an adequate answer. Avoid asking multiple distinct questions at once. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • As per the Tour there should be only one question asked per question. Also, please decide whether you wish to ask about R or ArcGIS Desktop in this particular question. – PolyGeo Mar 19 '17 at 11:45