Allow for 'climate ambient' for future downscaling to be a subset of a greater historical time period

[CaitlinA]

I want to be able to create a weatherDB, and then simulate SOILWAT based on this weatherDB, that runs from 1916 - Current (historical TP) and 2020 - 2059 & 2060 - 2099 ( 2future TPS with different RCP/GCMs).

I envision the that sc1 would be 1916 - 2010, and sc2 -> scX would be the typical TP.RCP.GCM format.
The creation of the future scenarios would rely on a subset of historical TP, that is the same length in years, as the future TP (i.e. 1970 - 2010, for future scenarios 2020 - 2059, 2060 - 2099).

If I set up the time frames of simulation like this:

#------ Time frames of simulation (may be modified by treatments)
sim_time <- list(
  # current simulation years is calculated as \code{simstartyr:endyr}
  # spinup_N is calculated as \code{startyr - simstartyr}
  # years used for results is determined as \code{startyr:endyr}
  simstartyr = 1915,
  startyr = startyr <- 1970,
  endyr = endyr <- 2010,

  #Future time period(s):
  # Each list element of 'future_yrs' will be applied to every
  #   climate.conditions
  # Each list element of 'future_yrs' is a vector with three elements
  #   \code{c(delta, DSfut_startyr, DSfut_endyr)}
  # future simulation years = delta + simstartyr:endyr
  # future simulation years downscaled based on
  #   - current conditions = DScur_startyr:DScur_endyr
  #   - future conditions = DSfut_startyr:DSfut_endyr
  # NOTE: Multiple time periods doesn't work with external type
  #   'ClimateWizardEnsembles'
  DScur_startyr = startyr,
  DScur_endyr = endyr,

  future_yrs = list(
    c(d <- 50, startyr + d, endyr + d),
    c(d <- 90, startyr + d, endyr + d - 1) # most GCMs don't have data for 2100
  )

Or should I be looking towards further change options and functionality?

[dschlaep]

• This will create a project with 3 climate conditions (marked text copied from comments in description file)
  • "current" conditions:
    • simulated years = 1915-2010 current simulation years is calculated as \code{simstartyr:endyr}
    • aggregated output based on years 1970-2010 years used for results is determined as \code{startyr:endyr}
  • future scenario 1:
    • simulated years = 1965-2060 future simulation years = delta + simstartyr:endyr
    • aggregated output based on years 2020-2060
    • downscaled based on future GCM data of years 2020-2060 future conditions = DSfut_startyr:DSfut_endyr and corrected based on GCM data of years 1970-2020 current conditions = DScur_startyr:DScur_endyr
  • future scenario 2: ...

--> The current code version does not offer the capability to use different (and "real calendar") years for different climate conditions for the simulation runs (it does that for the bias-correction of future climate conditions though).
--> The features associated with PRs #268 and #316 will allow to aggregate over several time windows; thus, with those features you could then run one simulation from 1916-2099 and aggregate for 1916-2018, 2020-2059, and 2060-2099.
--> In the absence of that, what you could do currently is to setup the weather database and treat every climate condition "as if it is current", i.e., specified as rows in the input files and use the columns YearStart and YearEnd of the treatment-design file to tell which years to simulate

[CaitlinA]

Thanks for the advice! I knew we had discussed doing this before, but wasn't sure of the status. I think getting the branch(es), associated with PRs above, in shape to merge will be next priority.

…

On Thu, Jul 19, 2018 at 6:03 AM, daniel ***@***.***> wrote: - This will create a project with 3 climate conditions (marked text copied from comments in description file) - "current" conditions: - simulated years = 1915-2010 current simulation years is calculated as \code{simstartyr:endyr} - aggregated output based on years 1970-2010 years used for results is determined as \code{startyr:endyr} - future scenario 1: - simulated years = 1965-2060 future simulation years = delta + simstartyr:endyr - aggregated output based on years 2020-2060 - downscaled based on future GCM data of years 2020-2060 future conditions = DSfut_startyr:DSfut_endyr and corrected based on GCM data of years 1970-2020 current conditions = DScur_startyr:DScur_endyr - future scenario 2: ... --> The current code version does not offer the capability to use different (and "real calendar") years for different climate conditions for the simulation runs (it does that for the bias-correction of future climate conditions though). --> The features associated with PRs #268 <#268> and #316 <#316> will allow to aggregate over several time windows; thus, with those features you could then run one simulation from 1916-2099 and aggregate for 1916-2018, 2020-2059, and 2060-2099. --> In the absence of that, what you could do currently is to setup the weather database and treat every climate condition "as if it is current", i.e., specified as rows in the input files and use the columns YearStart and YearEnd of the treatment-design file to tell which years to simulate — You are receiving this because you were assigned. Reply to this email directly, view it on GitHub <#322 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AM0e_3vHYah1x_3AUiGCsX4K9zR3PFNVks5uIIOqgaJpZM4VVW9u> .

-- Caitlin Andrews Ecologist Southwest Biological Science Center U.S. Geological Survey Mobile: 802-922-3494 Office: 928-556-7036