Introduction

This document puts together the information needed for the Deforestation Metrics paper that we expect to submit in 2022. In the paper, we want to use Brazilian soy to show differences in information from the backward- and forward-looking measures which we calculate as:

## Reading layer `BRA_adm0' from data source 
##   `C:\Users\michael.lathuilliere\Documents\GIS DataBase\Admin\BR-IBGE\BRA_adm0.shp' 
##   using driver `ESRI Shapefile'
## Simple feature collection with 1 feature and 67 fields
## Geometry type: MULTIPOLYGON
## Dimension:     XY
## Bounding box:  xmin: -73.98971 ymin: -33.74708 xmax: -28.84694 ymax: 5.264878
## Geodetic CRS:  WGS 84

Soy deforestation risk

We start by looking at Brazilian soy deforestation exposure for imports into the EU27 + UK in 2017.

Soy deforestation exposure (in ha) in the EU27 + UK supply chain refers to the amount of deforestation associated with the production of soybean imported in 2017. Soy deforestation exposure is calculated by:

  1. Overlapping soybean crop maps from Song et al (2021) in 2017 with deforestation maps (INPE-PRODES) in 2012-16.
  2. This quantity of deforestation is then summed within each Brazilian municipality and across years and assigned to soybean production and export in 2017
  3. This quantity of “past” deforestation is then use to calculate a deforestation per soybean tonnes produced which is used to assign deforestation to soybean imported in 2017 (according to the Trase supply chains - Brazil SEI-PCS v.2.6.0)

In 2017, the EU27 + UK imported a total of 12.933054 Mtonnes of soy products divided into 40.2 % beans, 59.8 % cake and 0 % oil.

These imports were associated with 0.045475 Mha of deforestation.

Deforestation in the EU27 + UK soy supply chain was identified in key Brazilian biomes:

## Adding missing grouping variables: `COUNTRY_OF_PRODUCTION`
## # A tibble: 7 × 4
## # Groups:   COUNTRY_OF_PRODUCTION [1]
##   COUNTRY_OF_PRODUCTION BIOME          SOY_DEF_EXP   PCT
##   <chr>                 <chr>                <dbl> <dbl>
## 1 BRAZIL                AMAZONIA           2591.   0.057
## 2 BRAZIL                CAATINGA           1914.   0.042
## 3 BRAZIL                CERRADO           29651.   0.652
## 4 BRAZIL                MATA ATLANTICA      912.   0.02 
## 5 BRAZIL                PAMPA             10404.   0.229
## 6 BRAZIL                PANTANAL              3.63 0    
## 7 BRAZIL                UNKNOWN BIOME         0    0

A first attempt at the map that can be improved

Soy-driven deforestation

The soy-driven deforestation relates to the deforestation in 2017 that is expected to become soybean considering past trends. This indicator is calculated by:

  1. Obtaining the amount of deforestation in 2017 (INPE-PRODES)
  2. Allocating this deforestation to future soybean production based on past trends in the municipality (2012-2016)

We assign the totality of this soy-driven deforestation to the EU27+UK in absolute terms for 2017.

We expect this method to change for the paper (work in progress).

In 2017, the EU27 + UK were associated with 108501.35 ha of soy-driven deforestation with the following break-down per biome:

## Adding missing grouping variables: `COUNTRY_OF_PRODUCTION`
## # A tibble: 4 × 4
## # Groups:   COUNTRY_OF_PRODUCTION [1]
##   COUNTRY_OF_PRODUCTION BIOME          SOY_DRIVEN_DEF   PCT
##   <chr>                 <chr>                   <dbl> <dbl>
## 1 BRAZIL                AMAZONIA               9751.  0.09 
## 2 BRAZIL                CERRADO               98608.  0.909
## 3 BRAZIL                MATA ATLANTICA          113.  0.001
## 4 BRAZIL                PANTANAL                 29.0 0

We can then associate these values with a deforestation intensity indicator that we can calculate as the ratio of deforestation in 2017 (per municipality) to existing “forest availability” (to be defined). This indicator could be used as a weighting factor to highlight the areas in Brazil which might require attention for future soy production in deforested areas.

The deforestation intensity I is:

\(I = SD*(TD/F)\)

where:

We calculate this metric first here using F = forest cover available by Mapbiomas:

Table of territorial deforestation in 2017:

## # A tibble: 6 × 2
##   BIOME          TERR_DEF
##   <chr>             <dbl>
## 1 AMAZONIA        676249.
## 2 CAATINGA         15006.
## 3 CERRADO         712968.
## 4 MATA ATLANTICA   37483.
## 5 PAMPA            67276.
## 6 PANTANAL         31263.

Then derive Figure 2 for the publication:

and then zoom in on Bahia (Matopiba region)

## # A tibble: 76 × 8
##    TRASE_ID   MUN_NAME_C           STATE SOY_D…¹ FORES…² TERR_…³ INTEN…⁴ SOY_D…⁵
##    <chr>      <chr>                <chr>   <dbl>   <dbl>   <dbl>   <dbl>   <dbl>
##  1 BR-1718907 SANTA ROSA DO TOCAN… TOCA…   1553.  1.16e6   3046.   0.003   0.002
##  2 BR-1721257 TUPIRAMA             TOCA…    439.  4.83e5    612.   0.001   0.001
##  3 BR-1703073 BARRA DO OURO        TOCA…    491.  8.96e5   1692.   0.002   0.001
##  4 BR-1711951 LAGOA DO TOCANTINS   TOCA…    483.  7.40e5   1676.   0.002   0.001
##  5 BR-1703842 CAMPOS LINDOS        TOCA…   1566.  2.34e6   2332.   0.001   0.001
##  6 BR-1712702 MATEIROS             TOCA…   1868.  1.65e6   3958.   0.002   0.001
##  7 BR-1701101 APARECIDA DO RIO NE… TOCA…    716.  8.08e5    992.   0.001   0.001
##  8 BR-1718881 SANTA MARIA DO TOCA… TOCA…    746.  1.28e6   2626.   0.002   0.001
##  9 BR-1718451 PUGMIL               TOCA…    117.  1.89e5    340.   0.002   0.001
## 10 BR-1720655 SILVANOPOLIS         TOCA…    579.  8.50e5   1461.   0.002   0.001
## # … with 66 more rows, and abbreviated variable names ¹​SOY_DRIVEN_DEF,
## #   ²​FOREST_HA, ³​TERR_DEF, ⁴​INTENSITY, ⁵​SOY_DEF_INT
## # ℹ Use `print(n = ...)` to see more rows