This writeup shows the primary analyses I think are worth including th in the report (marked with “Figure X” in their title), secondary analyses that may be appropritate for the appendix, and other analyses that are likely too detailed or messy to include in the report, but that are useful for understanding the data and trends.

I have included headings that specify the hypothesis being tested, subheadings to help with the structure, and titles for graphs.

Hypothesis: Global pasture area has historically increased, but recently began decreasing

Hypothesis: Global ruminant meat and milk production is decoupling from pasture area

Hypothesis: The drop in global pasture area is explained by changes in ruminant productivity

To Do: Analysis, from James’ draft report, on how much pasture area would be needed if productivity was at 1961 levels, holding all else constant

  • This analysis to illustrate the importance of productivity will be easy to include from James’ draft chapter or adapt as necessary.

Hypothesis: Global ruminant productivity has increased

Global Cattle Meat and Milk Yields

  • FAO reports animal yields as carcass weight, which is the product of slaughter weight and dressing percentage.

  • Cattle meat and milk yields have increased dramatically, with meat and milk yields per animal nearly 33.8% and 32.9 %higher, respectively in 2013 than in 1961. As the report describes, this improvement is emblematic of the “livestock revolution” that has improved land productivity of animal production and reduced pasture expansion pressure.

Global “Stocking Rate” Proxy

  • It is not just yields that have increased and contributed to growing production. While FAO does not report data on time taken to reach slaughter weigh, it does report “stocks” (or number of live animals), which can be used to calculate stocks per pasture area. This is different than stocking rate since the numerator includes both grazing animals and landless animals (those in stalls, feedlots, backyards etc.).
  • However, we can use this “stocks per pasture area”" metric it to assess intensification that arose from everything besides yield increases. The stocks per pasture area value accounts for changes in actual stocking rate and transitions from grazing to landless production irrespective of its effect on yields. However the stocks per pasture area value is also affected by changes in yields. As yields go up, stock numbers may decline, as they have in the U.S. And if yields go down, stock numbers may increase in response.

  • FAO’s data indicates that stocks per pasture area have increased as animal yields have too. Globally, stocks per pasture area can increase for various reasons. Producers within countries can increase stocking rates or shift animals from pasture to landless production (e.g. milk sheds). Producers can shift between raising animals primarily for meat or milk. For instance, the number of cattle per hectare in a country may remain constant, but if producers shift from meat to milk, more of the cattle will be in stalls and the stocks per pasture area value will increase. And production can shift across countries—when production moves from a lower stocking rate country to a higher stocking rate country, for instance, global rates increase.

  • There is a clear hiatus in grwoth from 1980 to 2000, and then a continuation of the earlier growth starting in 2000.

  • We can examine this further, looking at regional trends.

  • Regionally, we see that stocks per pasture area have followed different trends in different areas. They have increased in many developing regions such as Africa and South America, and have been relatively steady or declined in higher yield regions such as much of Europe and North America. We also see the unreliability of comparing post-USSR data with pre-dissolution data here.

  • We also see that the rapid global rise in the 2000s is due to a rapid rise in Southern Asia. This S. Asia rise is due both to an increase in cattle stocks and a decline in pasture area. While the rise in stocks per pasture hectare seems anomalous compared to other regions, neither the increase in stocks nor the drop in pasture themselves are particularly large or rapid compared to other regions. However, the combination of the two changes makes for a large change in stocks per pasture area. Also, since India has one of the largest cattle populations in the world (it recently surpassed Brazil), changes here strongly influence global trends. Since India’s cattle are predominantly used for milk rather than slaughter, and dairy cattle are often raised and fed in stalls for much or all of their lives, this gives India a particularly high value for cattle per pasture area.

Hypothesis: Cropland area for feed crops has grown, but this growth is small enough that the trend in pasture + feed cropland is similar to the pasture trend

Hypothesis: Pasture area has recently been declining or flat in most regions

Regional Changes in Pasture Area since 1961

  • Since 1961, pasture area has declined more than 20% in Oceania and Europe, remained relatively constant in Africa, increased about 10% in the Americas and increased more than 20% in Asia and the Former Soviet Union.

  • However, the Soviet Union data is inherently innacurate and unreliable. When the USSR dissolved, pasture went from 326.5 Mha for the USSR to 360 Mha for all the new countries, a one-year 10% increase.

  • The biggest changes, at least in % terms, have been in W. Asia, E. Asia, S. America and Former Soviet Union. The Former Sovet Union data is less reliable.
## 168 codes from your data successfully matched countries in the map
## 5 codes from your data failed to match with a country code in the map
## 75 codes from the map weren't represented in your data

  • You can see this further by looking at individual countries. China, Brazil and Saudi Arabia dominated pasture expansion since 1961, accounting for 40%, 84%, and 46% of the total pasture area change, respectively. This adds up to more than 100% since during this time periodt there was also pasture contraction in some countries.

  • Note: Countries shaded in grey are are missing or did not exist both in 1961 and 2016.

Regional Changes in Pasture Area since 2000

  • However, since 2000, pasture area has declined in every region except the Former Soviet Union where it has remained relatively steady, and in the Americas where pasture area has continued modestly increasing.

  • The increase in the Americas is largely driven by an increase in North, rather than South America, as can be seen in the more fine-grained breakdown above.
## 198 codes from your data successfully matched countries in the map
## 5 codes from your data failed to match with a country code in the map
## 45 codes from the map weren't represented in your data

  • You can see this further by looking at individual countries. The US dominated pasture expansion in the Americas since 2000, accounting for 73% of the total pasture area change.

  • Note: Countries shaded in grey are are missing. Australia is excluded from above map because its change in area is orders of magnitude larger than any other country. Countries in white had negligible (less than 5 thousand hectares) change. Sudan and S. Sudan are excluded since they separated during this period and their data is unreliable.

Peak Pasture Year by Region and Country

Europe & Oceania Pasture Peaked Decades Before Other Regions
year major_group peak
2009 Africa 787084720
2013 Americas 827841970
2002 Asia 856620000
1969 Europe 88719300
2010 Former Soviet Union 367296800
1973 Oceania 460225000
Pasture Area in Most Sub-Regions Peaked Before the 2000s
year minor_group major_group peak
1988 Caribbean Americas 5381000
1993 Central America Americas 91875000
1992 Eastern Africa Africa 247051020
2000 Eastern Asia Asia 522661000
1995 Eastern Europe Europe 14159000
2010 Former Soviet Union Former Soviet Union 367296800
1961 Former Yugoslavia Europe 6570000
1961 Middle Africa Africa 137255000
1980 Northern Africa Africa 79220000
1962 Northern America Americas 282850000
1961 Northern Europe Europe 19868500
1973 Oceania Oceania 460225000
2013 South America Americas 466110600
1989 South-Eastern Asia Asia 17878000
1961 Southern Africa Africa 156312000
1966 Southern Asia Asia 96717000
1961 Southern Europe Europe 24401000
2005 Western Africa Africa 183831000
2006 Western Asia Asia 222231000
1967 Western Europe Europe 25774200

  • Overall, every major region has seen its pasture area decline since it peaked. Every minor region has also either peaked or flattened off. This in itself is not particularly interesting to me since during any time frame there will be inherently be a peak year unless the most recent year had the highest level.

  • An interesting trend, however, is that pasture area peaked in every minor region by 2010 except for South America, and it peaked in most regions by the early 2000s.

  • We can see that over time, as different countries reached their peak pasture area (some as early as 1961, others as late as 2013), the portion of global pasture area inside countries taht have peaked has risen recently. It has gone up in fits and spurts, perhaps because of how countries report data or the FAO estimates missing data, but it clearly rose faster starting around 1990 than it did before.

  • By 2000, over 3/4 of pasture area was in countries where pasture area peaked. Of course, if pasture area goes up in any countries in the future, these values will fall. In other words, this finding may not hold up over time.

Hypothesis: Developed regions have seen greatest pasture productivity improvements and least pasture area growth

Pasture Area by Income Group

  • Pasture area and productivity trends are inherently tied to broader economic trends in countries.

  • Wealthier countries have typically seen a greater decline in pasture area since 2000, but also saw greater increases historically. This is particularly notable with middle income countries

  • Note: Countries are grouped according to their World Bank socioeconomic category as of 2016. Since not all countries included in the data exist today, this required assigning today’s categories for Russia to the USSR, for Ethiopia PDR to Ethiopia, for Belgium to Belgium-Luxemburg, for Serbia to Serbia & Montenegro, and assigning a value of Lower-Middle to Western Sahara, which is recognized as a discrete entity in FAO data but not in the World Bank’s.

  • The breakup of the USSR inevitably creates problems here since its constituent countries (e.g. Russia and Kazhakstan) have different income groups. The only ways to resolve this is to A) look at trends after the USSR dissolution, as Godde et al. did, B) omit the former USSR countries (including Russia) entirely; or C) change the income categorizations of the former USSR countries (which defeats the purpose of looking at trends by income group)

There are other ways of visualizing this data too or of grouping regions. We could choose one of these to display in the report. 1) The first graph is looking at pasture area since 1961 omitting the USSR and its countries 2) The second graph is looking at CHANGE in pasture area since 1961 omitting the USSR and its countries 3) The third graph is looking at change in pasture area since 1993 after the USSR breakup, including former USSR countries

  • This graph clearly indicates how pasture area has only increased substantially in middle income countries (particularly upper-middle income countries; not shown). Below you can see that this was primarily driven by dramatic increases in reported pasture area in China & Brazil, and that the 1990s incrase in high-income area was driven by Saudi Arabia and the US. In short, the rise in pasture area since the 1960s was largely driven by a rise in pasture area in a handful of countries.

  • After the break-up of the USSR, pasture rose for all income groups, but then declined below early 1990 levels except for low income countries. This trend holds whether you compare current pasture area today to area in 1992, 1993, 1994 or any other year in the 1990s.In other words, the trend is not an artefact of our decision to choose a particular post-USSR year.

Productivity by Income Group (Aggregated)

  • These graphs indicate that yields increased most (whether measured in absolute or % terms) in the high income group of countries, which also saw a drop in total pasture area over time. Including the USSR does not alter the general trends, though it does show a spike just following the USSR breakup.

Productivity by Income Group (Disaggregated at Country Level)

  • The following plots and tables show that beef yields have increased most in high and upper-middle income countries (looking at individual countries rather than at the aggregate level like we did above), and have recently decreased in low income countries. These analyses are all for the period 1993 to 2013. This is because we are interested in looking at differences by income group and the dissolution of the USSR makes it difficult to do so before this time period. An alternative would be to exclude the former USSR entirely.

  • Nevertheless, the recent trend is important and holds whether we look at change since 1993, since 2000, or most other years. Yields have increased most (in absolute terms) in higher income countries. THis holds whether I calculate a simple average for each income group or an average weighted by each country’s number of animals or area of pasture (not shown).

  • Note: Size of the dots represents number of cattle in 1993. Ignore the y axis on the graph. It is easy to remove when I make this plot not be interactive. I just wanted it interactive so you can identify individual countries.

Average beef yield change 1993-2013, Unweighted
Income Group Change in Yield
H 117.72917
L -19.30769
M 163.74026
Average beef yield change 1993-2013, Weighted by Number of Cattle
Income Group Change in Yield
H 5.3680995
L -0.3701941
M 2.6422932

Hypothesis (New): Pasture Area Decreases Are Explained Partially by Yield CHanges & GDP

In Progress: Regression of Country level Pasture on GDP, Yields, Consumption and Stocks.

  • I am still trying to identify the most appropriate regression specification for our data. There is a lot of temporal correlation and it may make more sense to look at changes in variables or the log of variables since our data for most variables has a highly skewed distribution. This may not matter since statistical significance tests arguably dont apply to population level data like this (we don’t have sampled data, so sample statistics don’t apply in the same way). But poeple prefer to see tests of significance anyway.

To Consider Doing: Plot country-level GDP vs. meat yields for 1 year, illustrating relationship

  • I could plot yields vs. GDP too on a scatter plot with one point for each country. However this graph illustrates the same underlying message as the ones above regarding yields increasing most in high income countries, and so we would only want one of them.

Hypothesis: Production and pasture area is shifting towards developing and tropical countries with low productivity and high biodiversity and carbon stocks

- The above graph plots the portion of global beef production each year that occured in countries with beef yields in different quantiles (low, medium and high) for that year. You can see that there has been a slight shift in production. In 1961, <13% of global production came from coutnries with the lowest yields. In the 2010s, >18% came from these countries. Production shifted primarily away from countries with high yields - the highest yielding countries went from providing >75% of production to supplying about 70%. The portion from countries with yields in the middle hasn’t shifted as substantially.

Pasture Area Has Increased Most in Humid Regions and Fell in Arid Ones

- Pasture Area by Dominant Climate

- Pasture Area by Dominant Climate

  • Changes in pasture area by climatic zone differ too. I have the data from Goddde now to replicate her graph (shown above).

  • The climate categories for countries were, according to Cecile Godde, " defined by Robinson et al. (2011)… Each country was associated with the climate type of that country for which the bovine number from grazing only and mixed crop-livestock systems was the highest after Herrero et al. (2013)“”

Hypothesis: Many projections for increasing pasture area do not align with recent trends. Nevertheless they indicate that if productivity slows or stalls, pasture area will dramatically grow with large environmental impacts.

- To Do: Comparison of Pasture Area Projections.

Misc. / Additional Appendices