Search About RLL About Mattick About Supplement Add to Supplement PDF file providers Help

Full record view

Brungard, C. B./ J. L. Boettinger 2012: Spatial prediction of biological soil crust classes: Value added DSM from soil survey. - Digital Soil Assessments and Beyond : 57-60. [RLL List # 231 / Rec.# 34652]
Keywords: Band ratios/ Biological soil crusts/ Covariates/ Land managements/ Land managers/ LandSat 7/ Micro-fungi/ Misclassification rates/ Misclassifications/ National parks/ Point data/ Random forests/ Soil map units/ Soil mapping/ Soil surfaces/ Spatial prediction/ Utah , USA/ Validation data/ Value added products/ Decision trees/ Mapping/ Soil surveys/ Soils
Abstract: Digital soil mapping can use data collected during soil survey efforts to produce value added products useful for land management. Using a digital soil mapping approach and point data from a recent soil survey update we modeled biological soil crust classes (communities of cyanobacteria, microfungi, lichens and mosses at the soil surface) in and around Canyonlands National Park, Utah, USA. Modeling was performed with Random Forests. Using a validation data set the misclassification rate was 55%. Although high, the misclassification rate may be acceptable. Overall error results from misclassification between similar classes. Important covariates included soil map unit and Landsat 7 ETM band ratio 5/2. We anticipate that spatial predictions of BSC classes will be useful to land managers when evaluating range health in Canyonlands National Park. Such value added products support the use of digital soil mapping activities in traditional soil survey. © 2012 Taylor & Francis Group.

URL:

[Email correction]

Upload PDF file to the RLL web site

If you have a PDF file of this RLL/Mattic record, and there are no copyright problems involved, you may upload the file to the RLL/Mattick site. The PDF file will be automatically linked to the paper, and available for download by everyone. Only one PDF file can be linked to a paper, any previous link will be lost.

PDF file::
NB! Legal characters: a-z, A-Z, 0-9, hyphen, underscore, dot (i.e. no diacritics, ampersand, space, etc.).

  

Upload URL to PDF file or web site

Alternatively, you can link this RLL/Mattick record to a PDF file or web page placed somewhere else on the web. Again, only a single link can exist for each record; any previous link will be lost.

Copy and paste the URL you wish to link to this record: