README.md

In IMMM-SFA/wmpp:

wmpp (Water Management Post-Processing)

wmpp produces Water Scarcity Grid Impact Factor (WSGIF) estimates and PLEXOS inputs from WM simulation results.

Installation

You can install wmpp using devtools:

devtools::install_github("IMMM-SFA/wmpp")

Running on PIC

Install wmpp in your PIC home directory

To run on PIC, you must first ensure R is loaded:

$ module load R/3.4.3

You also need to load some other modules required for the spatial mapping features of wmpp:

$ module load gdal/2.1.2

$ module load proj4/4.9.3

$ module load geos/3.4.2

$ module load gcc/5.2.0

If this is the first time you've used wmpp on PIC, you'll need to install it in your home directory. To do this, start R directly:

$ R

... and install wmpp as noted above:

devtools::install_github("IMMM-SFA/wmpp")

Then quit R, without saving the workspace image:

q("no")

Important: Even though it's possible to begin running wmpp from within the R environment on PIC, this should be avoided. Instead, you should set up some bash scripts to run wmpp on a computing node.

1. Extracting flow data from WM outputs

Computation of both Water Scarcity Grid Impact Factor (WSGIF) and plexos inputs requires flow time series for relevant grids where hydropower dams (for hydropower WSGIF) and catchment outlets (for thermal WSGIF and thermal/hydro plexos input) are located. The prep_flow function will search inside a WM results directory and extract these data from all available gridded flow outputs. The function has three arguments:

1. wm_output_dir is used to specify the WM results directory.

2. hyd_output_dir set the output directory for the extracted time series.

3.- in_parallel (default = FALSE) tells the functions whether or not run the regional results folders across separate cores.

An example WM output directory is located at /pic/projects/wat_nrg_nex/hydromet/Tian/CMIP5_BCSD_runoffs/output/access1-0/

To extract the relevant flow information you need to set up a simple .R script. Create a new folder in your $HOME directory (e.g. /people/my_username/wmpp_analysis/) and add the following R script (named get_flows.R or similar):

#################
# .R script for extracting flow time series for hydropower and thermal WSGIF computation

# load the wmpp library
library(wmpp)

# exectute prep_flow
prep_flow("/pic/projects/wat_nrg_nex/hydromet/Tian/CMIP5_BCSD_runoffs/output/ccsm4/",
          hyd_output_dir = "/people/my_username/wmpp_analysis/flows_hyd_huc4_ccsm4/",
          in_parallel = TRUE)
#################

Then create a bash (named something like batch_get_flows.sh) in the same directory:

#!/bin/bash

#SBATCH -p shared
#SBATCH -n 4
#SBATCH -A your_allocation
#SBATCH -t 120
#SBATCH --output=/dev/null
#SBATCH --error=/dev/null

R CMD BATCH get_flows.R

Note: WM gridded flow files are in the order of 1Gb each, causing slow load and computation within the R environment. Currently, the above script can take up perhaps up to 30 minutes (depending on how many flow files are in the results directory), even with the in_parallel option set to TRUE.

Then send the job for computation:

$sbatch batch_get_flows.sh

Once hydropower dam and huc4 flows have been prepared, you can compute both WSGIF and plexos inputs for hydro, fixed hydro, and thermal units.

2. WSGIF computation

The function get_wsgif can be run with the following .R script:

#################
# .R script for computing WSGIF (hydro and thermal)...
# ... for all regions, basins, balancing authorities ...
# ...and WECC as a whole for a given WM run.

# load the wmpp library
library(wmpp)

# execute get_wsgif for desired scenario
get_wsgif("HIST_MOSART_2010_WM.csv",        ## the historical part of the scenario
          "RCP85_MOSART_2010_WM.csv",       ## the future part of the scenario
          "flows_hyd_huc4_ccsm4")           ## flow file directory (created in prior step)
#################

Check the documentation for additional optional function arguments simply entering ?get_wsgif in the R console.

3. Plexos input computation

Similarly to WSGIF, Plexos inputs are computed from prepared WM huc4 flows.

Set up the following .R script:

#################
# .R script for computing plexos inputs (hydro, fixed hydro, and thermal).

# load the wmpp library
library(wmpp)

# execute get_plexos_inputs
get_plexos_inputs(
  "HIST_MOSART_2010_WM.csv",      ## historical part of simulation (base_yr contained within)
  "RCP85_MOSART_2010_WM.csv",     ## future part of simulation (plexos input computed for all these yrs)
  "flows_hyd_huc4_ccsm4/",        ## flow file directory (created in (1) above, using prep_flows)
  huc4_shape_dir = "./",          ## location of huc4 shapefile (will be automatically downloaded if missing)
  base_yr = 1963,                 ## year from which flow deviations are computed for derating
  "ccsm4_plexos_outputs/")        ## output directory
#################

Check the documentation for additional optional function arguments simply entering ?get_wsgif in the R console.

If you experience any problems running wmpp then please email Sean Turner or raise an issue on this reposity.

IMMM-SFA/wmpp documentation built on March 21, 2020, 4:26 p.m.