Moving Matters

The Changing Location of U.S., Brazilian and African Agriculture

Location, Location, Location!

Where crops grow has a big effect on the soil, climate and market access realties that influence crop growth—a classic Genomics by Environment by Management by Socioeconomic effect! But crops also move over time, thus changing these G x E x M x S realities.

G.E.M.S is using innovative analytical methods applied to both remote-sensed data and spatially explicit statistical survey data, often at different spatial and temporal scales, to track the shifting location of global agriculture.

Measuring the location and spatial movement of production makes it possible to assess the effects of crop movement on crop production and productivity. This crop-shifting information, in conjunction with other G.E.M.S data and analytical tools, is used to inform future breeding and crop management efforts. Agricultural location data are also indispensable to generating new insights into how agriculture might adjust to future changes in technology, climate and other (e.g., pest and disease) risks.

Measurement Matters

G.E.M.S Workflow—From Statistical Tables to Geo-referenced Agricultural Maps

G.E.M.S has created a suite of replicable and flexible workflows to process and geo-code published agricultural statistics in ways that facilitate analyses of the spatial location and historical movement of crop production in countries around the world.

We begin by locating and ingesting long-run, historical, sub-national crop production data and statistical boundary files for the countries and crops of interest. GEMSTools includes sophisticated, informatically-enabled, data processing algorithms that we deploy to clean and harmonize the raw and semi-processed statistical and boundary file data. These algorithms are integrated into a novel suite of modularized Spatio-Temporal Agricultural Analysis Tools (S-TAAT) as part of GEMSTools, and deployed via a Jupyter notebook. One S-TAAT module implements semi-automatic procedures that, among other things, match administrative-level production data with corresponding shapefile boundaries. Matching geographical boundaries to statistical data is tricky, not least because regional boundaries split and shift over time. For example, from 1920 to 2015, the number of municipalities in Brazil more than quadrupled.

To standardize the spatial representation of the data, S-TAAT also includes a suite of replicable analytical methods to map crop production data to a baseline set of administrative boundaries. Which standardizing approach to use depends on the nature of the underlying data and the purpose to which the spatially standardized data are to be put.

Making Spatial Sense of Shifting Statistical (Brazilian) Boundaries, 1920-2015

Spatially Standardizing (South African) Maize Production Boundaries, 1918-2015

Past Findings and On-Going Research

  • During the period 1879 to 2007 the average U.S. corn plant moved 320 kilometers in a northwesterly direction. That movement fundamentally changed the crop’s G x E x M x S relationship. We estimate this crop movement accounted for upwards of 20 percent of the 7-fold increase in U.S. corn production during the 20th century (Beddow and Pardey 2014).
  • InSTePP personnel at the University of Minnesota are working closely with Embrapa colleagues in Brazil, to further expand the S-TAAT in G.E.M.S for our joint study of the considerable shifts in space and time (single to multi-cropping) of Brazilian maize over the past 100 years. Ongoing work shows:
    • The movement of maize in Brazil is upwards of 603 kilometers (from 1920-2015)—even more dramatic than the significant U.S. movement—but once again in a northwesterly direction
  • Using procedures in GEMSTools to overlay “S” (geo-coded crop production) data on “E” (geo-coded climate) data. The average Brazilian maize plant is now grown in much warmer climes than was the case a century ago. This starkly contrasts with the situation in the United States where it appears that the average U.S. corn plant is now grown in a colder climate than was the case a century ago.
  • In other on-going work:
    • A colleague at Stellenbosch University, South Africa is working closely with University of Minnesota InSTePP personnel to deploy the S-TAAT on the G.E.M.S platform to analyze the movement of maize in South Africa. Here, it seems, significant policy distortions introduced during the first half of the 20th century incentivized the geographical expansion of maize production into areas where it could not be sustained once the policy distortions were removed in the post-Apartheid period.
    • The InSTePP team is also using S-TAAT to wrap up work mapping and measuring the long-run movement of maize in other African countries (Ethiopia, Kenya, Nigeria and Tanzania). To do this they spatialized tabular agricultural data drawn from historical statistical (including colonial government) records of agricultural production painstakingly compiled by InSTePP over the past decade.