Cattle production sensitivity analysis

Introduction

This document explores the difference in municipal level beef production (in carcass weights CW) following the estimates of cattle herd population for the TraseH2O project. Two cattle population estimates were calculated for the purposes of calculating cattle water use and GhG emissions (enteric methane and manure). The steps used to derive cattle population follow those now used by MCTI; we follow this method and use:

MCTI/Census estimate: The IBGE 2006 agricultural census to derive cattle age categories; this approach is followed by keeping and removing dairy cows from the cattle population.
Anualpec estimate: The Anualpec data was used to derive the cattle age categories; this approach includes the dairy cows by default.

We compared these estimates at the country level in 2016 with those published by MCTI:

MCTI without dairy cows = 198,383,961 heads
MCTI with dairy cows = 217,925,318 heads
Anualpec with dairy cows = 218,190,739 heads
reported by MCTI = 217,492,377 heads

We calculate beef production per municipality for the 2010-2019 period and check differences based on the cattle population by considering the dairy cow and bull populations.

First we calculate a new slaughter rate based on the revised live cattle population. Considering all available years (2010-2019).

There are differences in the slaughter rate based on whether the dairy cows and bulls are included in the cattle population. Previous work in zu Ermgassen et al (2020) state that the offtake rate for beef was about 20% meaning that the herd is renewed every 5 years. We check the average annual state slaughter rate estimates.

The annual means of the slaughter rates are still close to 20% so we can maintain the 5-y assumption. Ermgassen et al (2020) describe that in some states the offtake rate numbers are too high (PE, AP, DF, and BA), or too low (SP):

in PE, AP, DF, and BA, the national median should be use (and calculated based on the make-up of the cattle population)
in SP a value of 0.4072 is used

In our case, the median slaughter rates are as follows:

With median Brazil-side slaughter rates:

## # A tibble: 1 × 5
##   SR_BEEF_DAIRY SR_BEEF SR_BEEF_DAIRY_NO_TOUROS_CENSUS SR_BEEF_NO_TOUROS_CENSUS
##           <dbl>   <dbl>                          <dbl>                    <dbl>
## 1         0.176     0.2                          0.181                    0.206
## # ℹ 1 more variable: SR_BEEF_DAIRY_NO_TOUROS_ANUALPEC <dbl>

We then check how these numbers impact the beef production per municipality and then aggregate to the state and country level production.

Beef production is obtained as follows:

\[Beef~production = H_{m,y}*s_{s,y}*(cw_{s,y} + 0.0063LW_s)\] where H_m,y is the cattle herd in municipality m and year y, s_s,y is the slaughter rate in state s and year y (corrected for cattle movement), and cw_s,y is the carcass weight per state s and year y (from IBGE). We also include offalls at 6.3% of liveweight at slaughter (which depends on the state, see table below on slaughter liveweights used by MCTI).

We take the average liveweight of male and females at slaughter in the different regions of Brazil:

and calculate beef production per state for the years 2010-2017.

We derive the total CW per year and considering the different estimates

Our estimates are generally lower than what Erasmus proposed in his earlier work. We also see a range of roughly 0.2 Mtonnes of CW depending on the assumptions made when calculating the cattle population chosen for slaughter.

Our values are close to those determined by Erasmus, the difference being more apparent when zooming into a single state as seen below for Sao Paulo Erasmus doesn’t think that excluding bulls is quite relevant for this type of calculation, do we can probably just consider estimates with and without dairy cows only (with Census and Anualpec calculations).

In the figure, this means that the carcass weight is somewhere between the blue and red points.

These slight differences seen at the state level will then create smaller differences in the “per tonne” indicators for beef production:

deforestation per tonne of CW
Greenhouse gas emissions per tonne of CW
Enteric methane, manure and water use per tonne of CW

We then calculate the carcass weight per municipality and year for each estimate of cattle population between 2010 and 2017, and then provide 2 sets of results:

beef production per municipality for every year between 2010 and 2017
cumulative beef production per municipality over a 5-year period between 2015 and 2017

## [1] "saved at C:/Users/MICHAE~1/AppData/Local/Temp/RtmpwX3Ice/file63a01f2720b0/beef_production_cw.csv"

## [1] "saved at C:/Users/MICHAE~1/AppData/Local/Temp/RtmpwX3Ice/file63a01f2720b0/beef_production_cw_5y.csv"

Additional ideas:

We could further break down the slaughter per animal age and whether the animal was confined or not to give a better estimate of the carcass weight. At this tage we keep as is for the purposes of calculating the performace in each municipality.

Cattle production sensitivity analysis

Michael Lathuilliere

Nov 2 2024

Introduction

References