Reproducibility package for Shifting Landscapes: The Impact Of Agricultural Innovation On Deforestation

Indigenous Lands and Conservation Units (MapBiomas territory layers)

Files: data/raw/mapbiomas_col6/{f,p,a,s}_20km_{year}_{1–13}.csv (806 files); data/raw/administrative_boundaries/il_gee_20km.csv, cu_gee_20km.csv. MapBiomas Collection 6.0 data for the Brazilian Amazon, accessed through the MapBiomas public asset in Google Earth Engine (GEE). Accessed June 2022. Land use and land cover: share of each pixel covered by forest (f) and pasture (p), 1995–2010, and by soy (s) and other crops (a), 1996–2010. The original 30 × 30 m classes were grouped as follows: forest = classes 1, 9, 10; pasture = 15, 21; soy = 39, 41; other crops = 20, 36, 40, 46, 47, 48. They were then aggregated in GEE to the share of each pixel on a 1/6° grid (about 18.5 × 18.5 km, the paper's unit of analysis). Each file covers one class, year and tile (13 tiles per year); column remapped is the share and .geo is the pixel centroid. Indigenous lands and conservation units: pixels whose centroid lies within an indigenous land (il) or a federal or state conservation unit (fcu_ip, fcu_su, scu_ip, scu_su), assigned in GEE. The package contains these aggregated exports. The GEE scripts are not included; the steps are described in the Data Appendix of the paper (E.2, E.3.1).

Data is publicly available and included in the reproducibility package.

Creative Commons Attribution-Share Alike 4.0 International License

https://gee-community-catalog.org/projects/mapbiomas/#license

MapBiomas. 2022. Indigenous Lands and Conservation Units (MapBiomas territory layers) [dataset]. Accessed 2022-06.

Brazilian Agricultural Census 1995–1996, 2006 and 2017

Files: data/raw/agricultural_census/ (39 .xlsx files). Aggregated municipality-level results from the IBGE Agricultural Census, downloaded from SIDRA. See the paper's Data Appendix for the list of tables and variables. Accessed 2020-01. A full list of the files is included in data_hash_report.csv.

Data is publicly available and included in the reproducibility package.

Attribution-NoDerivs 3.0 Unported

https://creativecommons.org/licenses/by-nd/3.0/deed.en

https://sidra.ibge.gov.br/pesquisa/censo-agropecuario/censo-agropecuario-2017/resultados-definitivos

IBGE. 1996, 2006, 2017. Censo Agropecuário 1995–1996, 2006 and 2017. Rio de Janeiro: IBGE [dataset]. Accessed 2020-01.

Brazilian Demographic Census 1991

Files: data/raw/census_1991/literacy_91.xlsx, population_91.xlsx, rural_urban_91.xlsx, pop_census_income_earners.dta. Aggregated results from the 1991 IBGE Population Census and its Sample Supplement. The Sample Supplement data are pre-processed because of their size; see the paper's Data Appendix. Accessed 2022-03. Processing note: SIDRA exports with column headers renamed by the authors; aggregated to AMCs in clean_census1991.R.

Data is publicly available and included in the reproducibility package.

Attribution-NoDerivs 3.0 Unported

https://creativecommons.org/licenses/by-nd/3.0/deed.en

https://www.ibge.gov.br/estatisticas/sociais/populacao/25089-censo-1991-6.html

IBGE. 1991. Censo Demográfico 1991. Rio de Janeiro: IBGE [dataset]. Accessed 2022-03.

FAO-Global Agro-Ecological Zones v3.0 Soy Suitability

Files: data/raw/fao_gaez/high_soy.tif, high_soy.asc, low_soy.tif, low_soy.asc, ZS_high.csv, ZS_low.csv, dA_gee_20km.csv. Potential soy yields under high and low input levels for each pixel and municipality in Brazil, aggregated by the authors to 20 x 20 km pixels and municipalities. Accessed 2020-01. Processing note: Rasters clipped to Brazil; averaged to the 1/6° grid in GEE (dA = high − low) and to AMCs with QGIS Zonal Statistics (mean).

Data is publicly available and included in the reproducibility package.

FAO Terms and Conditions

https://www.fao.org/contact-us/terms/en/

https://gaez.fao.org/

FAO. 2015. Global Agro-ecological Zones (GAEZ v3.0). FAO, Rome, Italy [dataset]. Accessed 2020-01.

Suitability of Global Land Area for Pasture (FGGD)

Files: data/raw/fao_gaez/pasture_suitability/A_p_20km_*.csv (13 files). Pasture suitability index for each pixel in Brazil, accessed through the FAO Map Catalog and aggregated by the authors to a 20 x 20 km resolution. Accessed 2023-04. Processing note: Raster averaged to the 1/6° grid in GEE; exported in 13 tiles.

Data is publicly available and included in the reproducibility package.

FAO Terms and Conditions

https://www.fao.org/contact-us/terms/en/

https://gaez.fao.org/pages/data-access-download

FAO and IIASA. 2007. Suitability of Global Land Area for Pasture (FGGD). FAO Map Catalog [dataset]. Accessed 2023-04.

Environmental Fines Data

File: data/raw/ibama_fines/all_non_classified_fines.parquet. Description: Records of federal environmental fines (autos de infração) issued by IBAMA across Brazil, from the IBAMA open data portal (Fiscalização – Auto de Infração). The file covers fines issued from 1996 to May 2025. The analysis uses fines issued between 1996 and 2010. Accessed September 2025; the IBAMA extract is dated 22 May 2025 (field ULTIMA_ATUALIZACAO_RELATORIO). Processing note: The portal distributes one CSV file per year (auto_infracao_YYYY.csv, semicolon-separated, UTF-8). The authors combined the yearly files for 1996–2025 into a single parquet file. The files for 1977–1995 were not included. Within the included years, no rows were filtered and all columns published at the time of access were kept (68 columns). Values were not cleaned; all filtering, deduplication, and aggregation to the municipality-year level are done in the package code (code/DataWork/cleaning/). Access: The data contain personal information (infractor names and CPF/CNPJ numbers), so the file is not included in the reproducibility package, in line with the World Bank Data Policy. To run the code, users must: Download the yearly CSV files from the following link. Combine the files from 1996 onwards into one parquet file named all_non_classified_fines.parquet. Place it in data/raw/ibama_fines/. Note on versions: IBAMA keeps revising past records, and older extracts are not published. A new download will therefore not match the authors' 22 May 2025 extract exactly. For 1996–2010, today's files have only 1–4 additional fines per year, and fewer than 0.1% of records have changed status or municipality. Results may differ when using the published ones.

Data is publicly available, but since it contains personal information it is not included in the reproducibility package.

Open Data

https://dadosabertos.ibama.gov.br/about

https://dadosabertos.ibama.gov.br/dataset/fiscalizacao-auto-de-infracao/resource/b2aba344-95df-43c0-b2ba-f4353cfd9a00

The Brazilian Institute of Environment and Renewable Natural Resources (IBAMA). 2023. Dados Abertos do Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renováveis: Fiscalização – Auto de Infração [dataset]. Accessed 2025-09.

National Registry of Legal Entities (CNPJ) – Establishments

Files: data/raw/slaughterhouses/. Establishment-level records from the Receita Federal's Cadastro Nacional da Pessoa Jurídica (CNPJ), monthly release of June 2025 (reference date 14 June 2025). They are used to identify cattle slaughterhouses (CNAE 1011201/1011205) and their openings and closings in 1996–2010. Processing note: The establishment files (K3241.K03200Y[0-9].D50614.ESTABELE.csv, about 15 GB) are not included in the package because of their size. To get them, run bash data/raw/slaughterhouses/download_cnpj.sh (needs curl and unzip, about 4.8 GB download). The script downloads the original, unchanged files from the June 2025 release (folder 2025-06) at the link below, unzips them into data/raw/slaughterhouses/ and adds the .csv extension the code expects. The lookup tables (CNAE.csv, motivo.csv, municipios.csv) are included in the package. The verification was run with the files provided by the authors; if Receita Federal revises the files in this release, results may differ slightly.

Data is publicly available, but due to size constraints, they are not included in the reproducibility package.

https://dados.gov.br/dados/conjuntos-dados/cadastro-nacional-da-pessoa-juridica---cnpj

Receita Federal do Brasil. 2025. Cadastro Nacional da Pessoa Jurídica – CNPJ [dataset]. Accessed 2026-04.

Transport Logistics: Federal Roads 2005

Files: data/raw/administrative_boundaries/roads_geometries/rodovia_2005.* (shapefile). Federal road network (paved and unpaved) in 2005, based on the 2002 network. Accessed 2020-08.

Data is publicly available and included in the reproducibility package.

Attribution-NoDerivs 3.0 Unported

https://creativecommons.org/licenses/by-nd/3.0/deed.en

https://www.ibge.gov.br/geociencias/cartas-e-mapas/redes-geograficas/15793-logistica-dos-transportes.html?edicao=24589&t=acesso-ao-produto

IBGE. 2005. Logística dos Transportes [dataset]. Accessed 2020-08.

Regional Divisions of Brazil

Files: data/raw/administrative_boundaries/BR_Municipios_2024/, BR_UF_2024/, BR_Pais_2024/ (shapefiles), plus auxiliary municipality files in data/raw/administrative_boundaries/. Geometries of municipalities, states and national borders. Accessed 2025-05.

Data is publicly available and included in the reproducibility package.

Attribution-NoDerivs 3.0 Unported

https://creativecommons.org/licenses/by-nd/3.0/deed.en

https://www.ibge.gov.br/en/geosciences/territorial-organization/regional-division/21536-regional-divisions-of-brazil.html

IBGE. 2023. Regional Divisions of Brazil. Rio de Janeiro: IBGE [dataset]. Accessed 2025-05.

Biomes of Brazil

Files: data/raw/administrative_boundaries/Biomas_250mil/lm_bioma_250.* (shapefile). Geometries of Brazilian biomes at 1:250,000. Accessed 2025-05.

Data is publicly available and included in the reproducibility package.

Attribution-NoDerivs 3.0 Unported

https://creativecommons.org/licenses/by-nd/3.0/deed.en

https://www.ibge.gov.br/geociencias/informacoes-ambientais/vegetacao/15842-biomas.html?=&t=downloads

IBGE. 2023. Biomes of Brazil. Rio de Janeiro: IBGE [dataset]. Accessed 2025-05.

Minimum Comparable Areas (AMCs) 1872–2010

File: data/raw/administrative_boundaries/municipio_de_para.csv. Mapping of municipalities into minimum comparable areas, spatial units that are consistent over time, 1991–2010. Accessed 2020-01.

Data is publicly available and included in the reproducibility package.

https://revistas.usp.br/ee/article/view/106909/124729

Ehrl, P. 2017. Minimum comparable areas for the period 1872–2010: an aggregation of Brazilian municipalities. Estudos Econômicos (São Paulo) 47: 215–229.