In PrincetonUniversity/agroEcoTradeoff: Landuse Tradeoff analysis model

Spatial Tradeoff Model (trademod)

• Evaluates tradeoffs between:
  • Meeting production targets for key crops
  • Satisfying other land use constraints, including biodiversity and carbon protection

• Four primary components:

  1. targets - determines amount of production increases needed for a given scenario

  2. constraints - determines probability of pixels being converted to meet targets

     • Can be altered by change in yield potential, or application of 3 constraints
       1. Carbon density of landscape
       2. Biodiversity value (existing PAs, degree of fragmentation, mammal richness)
       3. Transport cost (in development)

  3. convert - determines which pixels are converted to meet the crop production target

  4. impacts - assesses impacts in terms of land converted, carbon lost, protected areas transformed, etc. (outstanding)

Some examples

A few of the input datasets

library(trademod)
path <- full_path(proj_root("trademod"), "trademod")
load(full_path(path, "data/cropnames.rda"))
y <- brick(full_path(path, "external/ext_data/ZA-potential-yields-std.tif"))
clist <- list("C" = brick(full_path(path, "external/ext_data/ZA-carbon-priorities.tif")), 
              "bd" = brick(full_path(path, "external/ext_data/ZA-cons-priorities.tif")), 
              "cost" = raster(full_path(path, "external/ext_data/ZA-cost.tif")))
crop_frac <- brick(full_path(path, "external/ext_data/ZA-crop-convert-fractions-base.tif"))
pot_yield <- brick(full_path(path, "external/ext_data/ZA-potential-yields.tif"))
na1 <- calc(y, sum, na.rm = FALSE)
na2 <- calc(pot_yield, sum, na.rm = FALSE)
na3 <- calc(crop_frac, sum, na.rm = FALSE)
valind <- which(!is.na(values(na1 + na2 + na3)))

par(mfrow = c(2, 2), mar = c(1, 2, 1, 3), oma = c(0, 0, 0, 1))
plot(pot_yield[[1]], main = "potential maize yield", axes = FALSE, box = FALSE)
plot(pot_yield[[5]], main = "potential soy yield", axes = FALSE, box = FALSE)
plot(clist$C, main = "carbon stocks", axes = FALSE, box = FALSE)
plot(clist$bd, main = "biodiversity value", axes = FALSE, box = FALSE)

And a few analyses

Let's establish production targets

A doubling in the need for 9 crops (over current production), except for maize (4X) and soybean (8X):

setwd(path)
currprod <- brick(full_path(path, "external/ext_data/ZA-current-production.tif"))
prod_targ <- c("maize" = 4, "cassava" = 2, "ground" = 2, "cotton" = 2, "soy" = 8, "pulse" = 2, 
               "sunflower" = 2, "sugarcane" = 2, "wheat" = 2)
potprod <- brick(full_path(path, "external/ext_data/ZA-potential-production-current-area.tif"))
target <- targets(prod_targ = prod_targ, currprod = currprod, potprod = potprod, ybeta = NULL)
target

And then run constraints to see which areas are most likely to be converted

We'll first assume no constraint other than crop yield potential is applied

setwd(path)
con1 <- constraints(y = y, ybeta = NULL, cbetas = c(0, 0, 0), clist = clist)
names(con1) <- cropnames
plot(con1,axes = FALSE, box = FALSE) 

Next we'll assume that we care about protecting carbon and biodiversity also

setwd(path)
con2 <- constraints(y = y, ybeta = NULL, cbetas = c(1, 1, 0), clist = clist)
names(con2) <- cropnames
plot(con2,axes = FALSE, box = FALSE) 

Lastly, let's see what get's converted given the crop production targets we need to meet

setwd(path)
converted1 <- convert2(con1, target, crop_frac = crop_frac, pot_yield = pot_yield, valind = valind, 
                       cropnames = cropnames)
converted2 <- convert2(con2, target, crop_frac = crop_frac, pot_yield = pot_yield, valind = valind,
                       cropnames = cropnames)

If crop yield is the only constraint

c1r <- dt_to_raster(dt = converted1, CRSobj = projection(con1))
plot(c1r, axes = FALSE, box = FALSE) 

If we also care about carbon and biodiversity

c2r <- dt_to_raster(dt = converted2, CRSobj = projection(con2))
plot(c2r, axes = FALSE, box = FALSE) 

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