Title:

   Supplementary files for "Adaptive Mapping of Design Ground Snow Loads in 
   the Conterminous United States"
   
   
Authors: 

   Jadon Wagstaff
   University of Utah 
   2000 Circle of Hope Dr., 
   Salt Lake City, UT 84112
   jadonw@gmail.com   

   Jesse Wheeler
   University of Michigan 
   323 West Hall 1085 South University, 
   Ann Arbor, MI 48109 
   jeswheel@umich.edu

   Brennan Bean 
   Utah State University
   3900 Old Main Hill
   Logan, UT 84322
   brennan.bean@usu.edu

   Marc Maguire
   University of Nebraska - Lincoln
   1110 S. 67th Street
   Omaha, NE 68182
   marc.maguire@unl.edu

   Yan Sun
   Utah State University 
   3900 Old Main Hill
   Logan, UT 84332
   yan.sun@usu.edu


Sponsoring Agency: 

   American Society of Civil Engineers / Structural Engineering Institute 
     (Award Number: 202827)


Abstract:

    Recent amendments to design ground snow load requirements in ASCE 7-22 have
    reduced the size of case study regions by 91\% from what they were in 
    ASCE 7-16, primarily in western states. This reduction is made possible 
    through the development of highly accurate regional generalized additive 
    regression models (RGAMs), stitched together with a novel smoothing scheme 
    implemented in the R software package remap, to produce the continental-
    scale maps of reliability-targeted design ground snow loads available in 
    ASCE 7-22. This approach allows for better characterizations of the 
    changing relationship between temperature, elevation, and ground snow loads
    across the Conterminous United States. RGAMs are shown to have 10% or 
    better improvement in mean absolute mapping error in two independently 
    created datasets when compared to traditional mapping techniques. 
    Potential implications and limitations of incorporating mapping accuracy 
    into the reliability-targeted load calculation are demonstrated and discussed. 


Details:

   This folder contains a dataset, a shapefile, two R packages, and script 
   that reproduces estimates of cross-validated accuracy in the 
   companion manuscript. The associated data represents a consolidation 
   of information from the following sources.  
   - The National Oceanic and Atmospheric Administration (NOAA) Global 
       Historical Climatological Network - Daily (GHCND) 
       (http://doi.org/10.7289/V5D21VHZ)
       Version: August 25th, 2020.  
   - Gridded climate normals from the PRISM Climate Group 
       (https://prism.oregonstate.edu/normals/) 
   - US EPA's Environmental Ecoregions 
       (https://www.epa.gov/eco-research/ecoregions-north-america)

  Data processing and models were created using R Statistical Software
  with the help of the tidyverse, sf, mgcv, automap, gstat, 
  snowload, and remap packages. 


Files: 

1. rtsl.csv
   This file contains the data that was used to train and test the random
   forest. This data includes the following:
   Variables: 
      - NAME: Name of the measurement station. Note that "stations" may 
              represent multiple stations from the GHCND. The listed name
              is GHCND-derived name of the primary station in the 
              consolidated set
      - STATE: Two-letter state code
      - RC_II: Risk Category II reliability-targeted design ground snow load 
               as calculated in ASCE 7-22 
      - LATITUDE: Decimal degrees latitude
      - LONGITUDE: Decimal degrees longitude
      - ELEVATION: Elevation (meters)
      - PPTWT: 30-year (1981 - 2010) sum of winter precipitation (Dec - Feb) 
               in mm
      - MCMT: 30-year (1981 - 2010) mean coldest month temperature 
              in Celsius

    Note: The reliability-targeted loads in this csv file are nearly identical
          to those used in the dataset used in ASCE 7-22. The slight 
          differences have to do with the inherent randomness in the simulations 
          required for reliability targeted load computation. The difference 
          between this dataset and the original is inconsequential for the 
          results presented in the associated manuscript. 

2. rtsl_cross_validation.R
   Contains the R script necessary to reproduce the tables of cross-validation
   results for the reliability-targeted loads in the associated manuscript. 

3. eco3_simp.Rdata
   An R data object that contains a simplified version of the EPA's level III
   ecoreigions. Used in the rtsl_cross_validation.R script. 

4. remap_0.3.0.tar.gz
   An R software package that performs the regional modeling in the associated 
   manuscript. Used in the rtsl_cross_validation.R script. 

5. remap_0.3.0.tar.gz
   An R software package that contains an adapation of the PRISM mapping 
   approach as described in the associated manuscript. Used in the 
   rtsl_cross_validation.R script. 



Instructions:

To run the example, download all files into the same folder and set 
your working directory in R to be the same as the location of these files.
Note that you may need to update other R packages in order to install
the tar.gz packages described previously. 


Publications: 

For additional details about data processing, please see:
- Wagstaff, J., Bean, B., Wheeler, J., Maguire, M., & Sun, Y. (in press). 
      Regionalized Mapping of Reliability-Targeted Design Snow Loads in the 
      United States. 
- Bean, B., Maguire, M., Sun, Y., Wagstaff, J., Al-Rubaye, S. A., Wheeler, J.,
  	Jarman, S., & Rogers, M. (2021). The 2020 National Snow Load Study.
	Mathematics and Statistics Faculty Publications. Paper 276.
	https://digitalcommons.usu.edu/mathsci_facpub/276.