changes.md

In PrincetonUniversity/agroEcoTradeoff: Landuse Tradeoff analysis model

title: "Changes" author: "Lyndon Estes" output: html_document: highlight: tango theme: spacelab toc: yes

v2.0.1

Improvements

• Added (back) ability to select a subset of crops, based on target values passed into targets, via tradeoff_mod, tradeoff_batch, and pareto. Switch is made in input_handler.
• Reordered arguments in tradeoff_mod.
• Added save for batch parameters to tradeoff_batch.
• Added more plot flexibility (N crops plotted, colors) to imp_plot.
• Added heat to create raster showing frequency of conversions.
• tradeoff_batch passes out the bcode now
• pareto uses bcode passed from tradeoff_batch, and saves optitab to batch directory.
• tradeoff_batch impact tables now come out as data.frames from both todisk TRUE and FALSE cases. Changes to batch_param made to accomodate (table rather than list). From the todisk=FALSE case, the output conversion tables are served as a list that is the second element of the output list.
• Added imp_sorter function for selecting subsets of conversion maps and ordering them to correctly match names of selected impact tables.
• Added new measures to impact for calculating total travel hours (impact_mod2), and total biodiversity cost (impact_mod3).

Bugs

• Removed hard-coded input_key ("ZA") in tradeoff_batch call to tradeoff_mod.
• Fixed .tif append to impacts_tab.rda output from tradeoff_batch.
• Fixed tradeoff_batch so that impacts_tab.rda is a data.frame not a list.
• Conversion tables from tradeoff_batch were being re-ordered because of random numbers/time-stamp structure in run_code. The run/batch/name is simplified and fixed, with corresponding changes to tradeoff_batch, tradeoff_mod, pareto.

v2.0

Summary of improvements

1. Current production becomes non-spatial, adjusted with coefficient

   • Notes:

     • currprod.rda is main file to read

   • Status: Done

2. Fraction available cropland needs to be incorporated

   • Status:
     • Done in convert
     • Done in impacts

3. New biodiversity impact metric incorporated

   • Status: Mostly - still need to change the biodiversity metric so that impact can be weighted evenly on protected weighted rarity and intactness

Changes

• fetch_inputs

  • including altering carbon-names.rda, dumping cropnames.rda entirely
  • Updated fetch_inputs to read in convertible fraction data.table, and to find spatial metadata.
  • Put in logic to check for equality of row number in input data.tables, and to call a halt if there are NAs in any of the inputs.
  • removed datasets in R data folder, because they should be installed as part of library, and therefore not called with a specific path by this function. To allow users to generate their own data.
    • All inputs now live under external/data/[ID], and are read in directly from there. These changes are designed to separate model from R package structure.

• input_handler

  • yield_mod
    • Name changed from yield_mod_dt
    • fixed to remove pp_curr from list of modified rasters
    • noticed potential standardization issue
    • renamed to just yield_mod
    • moved yield standardization to input_handler
  • added yield standardization here
  • switched off one potential yield modification option
  • all data.tables, including mask, reduced down to location with >0 farmable areas.
  • key set on index of reduced mask
  • cost/yield is performed on those areas only
  • rearranged order of inputs to properly deal with recyled input list
  • disabled yield modification function for now, on the assumption that this becomes an offline process for the time being. If ybeta_update is set to 1, model will halt.

• targets

  • name changed from targets_dt
  • replaced scheme for modifying current production with simple modification vector, which draws on vector of current production, with argument named currprodmod.
  • yield_mod argument removed.

• constraints

  • name changed from contraints_dt
  • allocation logic checked
  • j loop in inner allocation loop replaced with data.table-centric code

• convert

  • Name changed from convert_dt
  • Replaced j for loop with data.table syntax

• impact

  • Name changed from impact_dt
  • Modularized - impact metrics now calculated in sub-modules, except for total area converted.
    • Revised metrics for cost - mean cost and cost/yield calculated also - these will be more appropriate than total cost
    • New metrics for biodiversity, based on rarity. Several flavors to choose from - mean vegetation type rarity of converted pixels (0-1); rarity + protectedness (priority, 0 - 1); average intactness (0 - 1); mean of intactness and priority (the measure currently used by constraints); ha of forest reserves lost
    • Added additional biodiversityChanged biodiversity impact to be based on both intactness and rarity - i.e. run it on the cons prioritization layer.
  • Reduced number of input arguments to 2 (il and conv).

• tradeoff_mod

  • Updated to incorporate changes to 4 primary sub-modules.

• tradeoff_batch

  • Simplified (yield_mod disabled) and parallelized with foreach and doMC.
  • Output saving options altered. Conversion tables written out without xy coordinates, to save disk space and write times.
  • Outputs now write into external/output

• pareto

  • Modularized
  • pareto_step function created, removing this portion of code from inside main function
  • batch_params combines parameter permutations into matrix used by pareto and tradeoff_batch
  • batch_stat calculates summarizes impact metrics of interest from a batch run, which are fed to non-dominator.
  • non-dominator Separate function created for code that removes non-dominated solutions from output table of impacts
  • Function engine is tradeoff_batch.
  • parameter returned with optimization table in output list

PrincetonUniversity/agroEcoTradeoff documentation built on May 8, 2019, 3:12 p.m.