The Effect of Best Crop Practices in the Pig and Poultry Production on Water Productivity in a Southern Brazilian Watershed
Abstract
:1. Introduction
2. Materials and Methods
2.1. General Approach
2.2. Study Area and Animal Production System Data
2.3. Calculation of Water Productivity
- WPindirect+direct,broiler,Mass,Farm is water productivity of chicken meat produced on mass base (kg Carcass Weight m−3);
- WPindirect+direct,pig,Mass,Farm is water productivity of pork meat produced on mass base (kg Carcass Weight m−3);
- WPindirect+direct,broiler,Energy,Farm is water productivity of chicken meat produced on food energy base (MJ m−3);
- WPindirect+direct,pig,Energy,Farm is water productivity of pork meat produced on food energy base (MJ m−3);
- WPindirect+direct,broiler,Mon,Farm is water productivity of chicken meat produced on a monetary base (R$ m−3);
- WPindirect+direct,pig,Mon,Farm is water productivity of pig meat produced on a monetary base (R$ m−3);
- Qdirect+indirect,broiler,Farm is water consumption for broiler in fattening stage production (m3 year−1);
- Qdirect+indirect,pig,Farm is water consumption for pig production in fattening and pre-chain stages (m3 year−1).
- Qindirect_direct,broiler,Farm is the total water consumed for broiler purchased feed production + water consumed for broiler production (m3 year−1) considering broiler fattening stage;
- Qindirect_direct,pig,Farm is the total water consumed for pig purchased feed production + water consumed for pig production (m3 year−1) considering pig fattening and pre-chain stages;
- Qindirect,broiler,Feed is the total water consumed (evapotranspiration (ET); fresh matter (FM)) for purchased broiler feed production (m3 year−1). It was based on the ratio of the yield of the field (cropland) for producing broiler feed and the ET from the field (from harvest of the previous crop through to harvest of the crop) [1];
- Qindirect,pig,Feed is the total water consumed (ET; FM) for purchased pig (fattening and pre-chain stages) feed production (m3 year−1). It was based on the ratio of the yield of the field (cropland) for producing pig feed and the ET from the field (from harvest of the previous crop through to harvest of the crop) [1];
- Qdirect,broiler,Animal is the total water consumed for broiler drinking (m3 year−1);
- Qindirect+direct,pig,Animal is the total water consumed for pig drinking in pre-chain and fattening stages (m3 year−1);
- Qdirect,broiler,Housing is the total water consumed for services (cooling, cleaning) (m3 year−1) for broiler production;
- Qindirect+direct,pig,Housing is the total water consumed for services (cleaning) (m3 year−1) for a pig in pre-chain and fattening stages production.
- WPindirect,broiler,Feed is water productivity of broiler feed consumption (kgFM m−3);
- WPindirect,pig,Feed is water productivity of pig feed consumption (fattening and pre-chain stages) (kgFM m−3).
- (a)
- With regional climate data ET0, a grass reference surface was calculated using the FAO Penman-Monteith equation.
- (b)
- To model Tc, the ET0 was adjusted for the individual crop with plant-specific parameters (e.g., the plant-specific basal crop coefficient (Kcb)). Plant-specific parameters are provided in Table S2.
- (c)
- The calculation of Tact incorporates the effect of daily water stress due to water-limited conditions by linking the datasets on plants, soil, and climate on Tc. A water stress coefficient (Ks) incorporated water stress and reduced Tc to Tact. To determine the water stress coefficient (Ks), a simple tipping bucket approach was combined with regional soil and precipitation data. The equation for Tact (mm) applied here was:
Production Region | CR—Main Crop | Sowing Date 1 | Harvest Date | Vegetation Period (days) 2 | Mean CY (t FM ha−1) 3 ± SE | Mean HY (t FM ha−1) 3 ± SE |
---|---|---|---|---|---|---|
Mato Grosso State City: Primavera do Leste | fm—safra | 21 September | 24 January | 125 | 7.6 ± 0.71 | 9.3 ± 1.07 |
fs—soy | 22 October | 15 January | 85 | 3.1 ± 0.77 | 3.8 ± 0.34 | |
smf—soy | ||||||
smf—safrinha | 16 January | 4 June | 140 4 | 5.7 ± 0.07 | 7.3 ± 1.14 | |
Paraná State City: Cascavel | fm—safra | 24 September | 24 January | 125 | 10.1 ± 1.03 | 11.7 ± 1.02 |
rm—safra | ||||||
fs—soy | 30 October | 23 January | 85 | 3.4 ± 0.36 | 4.1 ± 0.53 | |
rs—soy | ||||||
smf—soy | ||||||
smf—safrinha | 24 January | 12 June | 1404 | 5.1 ± 1.14 | 7.6 ± 1.20 | |
Rio Grande do Sul State City: Vacaria | fm—safra | 15 October | 17 February | 125 | 7.1 ± 1.33 | 10.3 ± 1.83 |
rm—safra | ||||||
fs—soy | 15 November | 8 February | 85 | 2.9 ± 0.6 | 4.0 ± 0.53 | |
rs—soy |
2.4. Scenarios Analyzed
- 2nd scenario (M50): lower soil evaporation of 50% was applied considering the same mean annual crop yield analyzed in the CY scenario;
- 4th scenario (HY50%): a restriction of 50% in the soil evaporation was applied for the same highest state mean annual crop yield analyzed in the HY scenario.
2.5. Sources of Uncertainty
3. Results and Discussion
3.1. Feed Crop Water Productivity Scenarios and Potential Water Saving
3.2. Water Input and Water Productivity of the Chicken Meat and Pig Pork
3.3. Uncertainty Data Analyses
3.4. Applications of the LEAP-Guidelines
- To analyze the temporal variation in the water productivity indicators and their range, depending on the farming system.
- To investigate the effectiveness of single and combined measures of farmers to improve water productivity and the water scarcity footprint and
- To analyze the uncertainty considering evapotranspiration combined with the uncertainty of farm-basic data, environmental farm conditions, animal production features, and complementary data, such as maximum animal housing and size of the barns.
4. Conclusions
Supplementary Materials
Author Contributions
Funding
Acknowledgments
Conflicts of Interest
References
- FAO. Water Use in Livestock Production Systems and Supply Chains—Guidelines for Assessment (Version 1); Livestock Environmental Assessment and Performance (LEAP) Partnership; FAO: Rome, Italy, 2019; p. 130. [Google Scholar]
- OECD. Sustainable Management of Water Resources in Agriculture; Organisation for Economic Co-operation and Development (OECD): Paris, France, 2010; p. 120. [Google Scholar] [CrossRef]
- Mekonnen, M.M.; Hoekstra, A.Y. Four billion people facing severe water scarcity. Sci. Adv. 2016, 2, e1500323. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- ISO. Environmental Management—Water Footprint—Principles, Requirements and Guidelines; ISO 14:046; International Organization for Standardization: Geneva, Switzerland, 2014. [Google Scholar]
- ISO. ISO/TR 14073: Environmental Management—Water Footprint—Illustrative Examples on How to Apply ISO 14046; International Organization for Standardization: Geneva, Switzerland, 2016; p. 62. [Google Scholar]
- ANA. Brazilian Water Resources Report—2017: Full Report; National Water Agency—ANA: Brasilia, Brazil, 2018; p. 169. [Google Scholar]
- Cunha, A.P.M.A.; Zeri, M.; Deusdará Leal, K.; Costa, L.; Cuartas, L.A.; Marengo, J.A.; Tomasella, J.; Vieira, R.M.; Barbosa, A.A.; Cunningham, C.; et al. Extreme Drought Events over Brazil from 2011 to 2019. Atmosphere 2019, 10, 642. [Google Scholar] [CrossRef] [Green Version]
- Gutiérrez, A.P.A.; Engle, N.L.; De Nys, E.; Molejón, C.; Martins, E.S. Drought preparedness in Brazil. Weather Clim. Extrem. 2014, 3, 95–106. [Google Scholar] [CrossRef] [Green Version]
- Coelho, C.A.S.; Cardoso, D.H.F.; Firpo, M.A.F. Precipitation diagnostics of an exceptionally dry event in São Paulo, Brazil. Theor. Appl. Climatol. 2016, 125, 769–784. [Google Scholar] [CrossRef]
- Marengo, J.A.; Alves, L.M.; Alvala, R.C.S.; Cunha, A.P.; Brito, S.; Moraes, O.L.L. Climatic characteristics of the 2010-2016 drought in the semiarid Northeast Brazil region. An. Acad. Bras. Ciências 2018, 90, 1973–1985. [Google Scholar] [CrossRef] [PubMed]
- FAO. Climate Change, Water and Food Security. In Water Report; Food and Agriculture Organization of the United Nations—FAO: Rome, Italy, 2011; p. 200. [Google Scholar]
- Collins, M.; Knutti, R.; Arblaster, J.; Dufresne, J.L.; Fichefet, T.; Friedlingstein, P.; Gao, X.; Gutowski, W.J.; Johns, T.; Krinner, G.; et al. Long-term Climate Change: Projections, Commitments and Irreversibility. In Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change; Cambridge University Press: Cambridge, UK; New York, NY, USA, 2013. [Google Scholar]
- de Fraiture, C.; Wichelns, D. Satisfying future water demands for agriculture. Agric. Water Manag. 2010, 97, 502–511. [Google Scholar] [CrossRef]
- Rockström, J.; Karlberg, L.; Wani, S.P.; Barron, J.; Hatibu, N.; Oweis, T.; Bruggeman, A.; Farahani, J.; Qiang, Z. Managing water in rainfed agriculture—The need for a paradigm shift. Agric. Water Manag. 2010, 97, 543–550. [Google Scholar] [CrossRef] [Green Version]
- Hoff, H.; Falkenmark, M.; Gerten, D.; Gordon, L.; Karlberg, L.; Rockström, J. Greening the global water system. J. Hydrol. 2010, 384, 177–186. [Google Scholar] [CrossRef]
- Bhattacharya, A. Chapter 3—Water-Use Efficiency Under Changing Climatic Conditions. In Changing Climate and Resource Use Efficiency in Plants; Bhattacharya, A., Ed.; Academic Press: London, UK, 2019; pp. 111–180. [Google Scholar] [CrossRef]
- ANA. Uso da Água na Agricultura de Sequeiro no Brasil (2013–2017); National Water Agency—ANA: Brasilia, Brazil, 2020; p. 63. [Google Scholar]
- Empinotti, V.L. Water Issues and the Brazilian Agricultural Agenda. In The Oxford Handbook of Food, Water and Society; Allan, T., Bromwich, B., Keulertz, M., Colman, A., Eds.; Oxford University Press: Oxford, UK, 2019. [Google Scholar]
- Drastig, K. World Food Supply and Water Resources: An Agricultural-Hydrological Perspective (AgroHyd)—Final Report; Leibniz-Gemainschaft: Berlin, Germany, 2016. [Google Scholar]
- Drastig, K.; Prochnow, A.; Kraatz, S.; Libra, J.; Krauß, M.; Döring, K.; Müller, D.; Hunstock, U. Modeling the water demand on farms. Adv. Geosci. 2012, 32, 9–13. [Google Scholar] [CrossRef] [Green Version]
- Bastiaanssen, W.G.M.; Steduto, P. The water productivity score (WPS) at global and regional level: Methodology and first results from remote sensing measurements of wheat, rice and maize. Sci. Total Environ. 2017, 575, 595–611. [Google Scholar] [CrossRef]
- Brauman, K.A.; Siebert, S.; Foley, J.A. Improvements in crop water productivity increase water sustainability and food security—A global analysis. Environ. Res. Lett. 2013, 8, 024030. [Google Scholar] [CrossRef]
- Mekonnen, M.M.; Hoekstra, A.Y.; Neale, C.M.U.; Ray, C.; Yang, H.S. Water productivity benchmarks: The case of maize and soybean in Nebraska. Agric. Water Manag. 2020, 234, 106122. [Google Scholar] [CrossRef]
- Rudnick, D.; Irmak, S.; Ferguson, R.; Shaver, T.; Djaman, K.; Slater, G.; Bereuter, A.; Ward, N.; Francis, D.; Schmer, M.; et al. Economic Return versus Crop Water Productivity of Maize for Various Nitrogen Rates under Full Irrigation, Limited Irrigation, and Rainfed Settings in South Central Nebraska. J. Irrig. Drain. Eng. 2016, 142, 04016017. [Google Scholar] [CrossRef]
- Adeboye, O.B.; Schultz, B.; Adekalu, K.O.; Prasad, K.C. Performance evaluation of AquaCrop in simulating soil water storage, yield, and water productivity of rainfed soybeans (Glycine max L. merr) in Ile-Ife, Nigeria. Agric. Water Manag. 2019, 213, 1130–1146. [Google Scholar] [CrossRef]
- Flach, R.; Skalsky, R.; Folberth, C.; Balkovic, J.; Jantke, K.; Schneider, U.A. Water productivity and footprint of major Brazilian rainfed crops—A spatially explicit analysis of crop management scenarios. Agric. Water Manag. 2020, 233. [Google Scholar] [CrossRef]
- Sul, R.G.d. Resolução CONSEMA 372/2018; Government of Rio Grande do Sul State: Porto Alegre, Brazil, 2018; Volume 372. [Google Scholar]
- Martins, E.S.; Filho, J.I.S.; Sandi, A.J.; Miele, M.; de Lima, G.J.M.M.; Bertol, T.M.; Amaral, A.L.; Morés, N.; Kich, J.D.; Costa, O.A.D. Comunicado Técnico 506—Coeficientes Técnicos Para o Cálculo do Custo de Produção de Suínos, 2012; Empresa Brasileira de Pesquisa Agropecuária (EMBRAPA)—Embrapa Suínos e Aves: Concórdia, Brazil, 2012; p. 10. [Google Scholar]
- Miele, M.; de Abreu, P.G.; Abreu, V.M.N.; Jaenisch, F.R.F.; Martins, F.M.; Mazzuco, H.; Sandi, A.J.; Filho, J.I.S.; Trevisol, I.M. Comunicado Técnico 483—Coeficientes Técnicos Para o Cálculo do Custo de Produção de Frango de Corte, 2010; Empresa Brasileira de Pesquisa Agropecuária (EMBRAPA)—Embrapa Suínos e Aves: Concórdia, Brazil, 2010; p. 14. [Google Scholar]
- EMBRAPA. Tabela de Drawback para Frango de Corte—Memorial Técnico Descritivo Para Índices de Equivalência Entre Insumos Produtos Exportados; Empresa Brasileira de Pesquisa Agropecuária (EMBRAPA)—Embrapa Suínos e Aves: Concórdia, Brazil, 2019; p. 24. [Google Scholar]
- Cobb-Vantress. Suplemento de nutrição e desempenho do frango de corte—COBB 500; Cobb-Vantress: Rod Assis Chateaubriand, Brazil, 2015; p. 14. [Google Scholar]
- EMBRAPA. Tabela de Drawback Para SUÍNOS—Memrial Técnico Descritivo Para Índices de Equivalência Entre Ingredientes de Ração de Suínos e os Seguintes Produtos Industrializados; Empresa Brasileira de Pesquisa Agropecuária (EMBRAPA)—Embrapa Suínos e Aves: Concórdia, Brazil, 2019; p. 18. [Google Scholar]
- Agroindustry. Pig Ratio Formulation; Agroindustry from Southern Brazil: Casca, Brazil, 2019. [Google Scholar]
- USDA. USDA National Nutrient Database for Standard Reference, Release 27; US Department of Agriculture, Agricultural Research Service: Beltsville, MD, USA, 2014; p. 153. [Google Scholar]
- EMBRAPA. Embrapa Swine and Poultry: Prices 2019. Empresa Brasileira de Pesquisa Agropecuária (EMBRAPA), 2019. Available online: https://www.embrapa.br/en/suinos-e-aves/cias/precos (accessed on 4 February 2020).
- IBGE. Produção Agrícola Municipal (PAM). Instituto Brasileiro de Geografia e Estatística (IBGE), 2019. Available online: https://sidra.ibge.gov.br/pesquisa/pam/tabelas (accessed on 4 February 2020).
- INMET. Brazilian Climate Data (2008–2018). Instituto Nacional de Meteorologia (INMET), 2019. Available online: https://portal.inmet.gov.br/ (accessed on 12 February 2019).
- Alvares, C.A.; Stape, J.; Sentelhas, P.; Gonçalves, J.; Sparovek, G. Köppen’s climate classification map for Brazil. Meteorol. Z. 2013, 22. [Google Scholar] [CrossRef]
- Streck, E.; KÄMpf, N.; Dalmolin, R.; Klamt, E.; Nascimento, P.; Schneider, P.; Giasson, E.; Pinto, L.F.S. Solos do Rio Grande do Sul; UFRGS: Porto Alegre, Brazil, 2008. [Google Scholar]
- IBGE. Folha SH. 22 Porto Alegre e Parte das Folhas SH. 21 Uruguaiana e SI. 22 Lagoa Mirim: Geologia, Geomorfologia, Pedologia, Vegetação, uso Potencial da Terra/Fundação Instituto Brasileiro de Geografia e Estatistica [v. 33]; Instituto Brasileiro de Geografia e Estatística (IBGE): Rio de Janeiro, Brazil, 1986. [Google Scholar]
- Palhares, J.C.P. Comunicado Técnico 102—Consumo de Água na Produção Animal; Empresa Brasileira de Pesquisa Agropecuária (EMBRAPA)—Embrapa Pecuária Sudeste: São Carlos, Brazil, 2013. [Google Scholar]
- Drastig, K.; Palhares, J.C.P.; Karbach, K.; Prochnow, A. Farm water productivity in broiler production: Case studies in Brazil. J. Clean. Prod. 2016, 135, 9–19. [Google Scholar] [CrossRef] [Green Version]
- FEPAM. Critérios Técnicos para o Licenciamento Ambiental de Novos Empreendimentos Destinados à Suinocultura; Fundação Estadual de Proteção Ambiental Henrique Luiz Roessler (FEPAM): Porto Alegre, Brazil, 2014. [Google Scholar]
- Allen, R.; Pereira, L.; Raes, D.; Smith, M. FAO Irrigation and drainage paper No. 56. Rome Food Agric. Organ. U. N. 1998, 56, 26–40. [Google Scholar]
- Bodner, G.; Loiskandl, W.; Kaul, H.P. Cover crop evapotranspiration under semi-arid conditions using FAO dual crop coefficient method with water stress compensation. Agric. Water Manag. 2007, 93, 85–98. [Google Scholar] [CrossRef]
- von Hoyningen-Huene, J.F. Einfluss der Landnutzung auf den Gebietswasserhaushait—Die Interzeption des Niederschlags in Landwirtschaftlichen Pflanzenbeständen; DVWK-Schrift Hamburg: Berlin, Germany, 1983. [Google Scholar]
- Braden, H. Ein Energiehaushalts- und Verdunstungsmodell für Wasser und Stoffhaushaltsuntersuchungen landwirtschaftlich genutzter Einzugsgebiete. Mitt Dtsch. Bodenkdl Ges. 1985, 42, 294–299. [Google Scholar]
- Kroes, J.G.; Dam, J.C.v. Reference Manual SWAP; Version 3.0.3; Alterra: Wageningen, The Netherlands, 2003. [Google Scholar]
- Brazil. Sistema de Zoneamento Agrícola de Risco Climático; Ministério da Agricultura, Pecuária e Abastecimento (MAPA): Brazil, 2019. Available online: http://indicadores.agricultura.gov.br/zarc/index.htm (accessed on 12 January 2019).
- Staff, U.-S.S.D. Soil Survey Manual; USDA: Washington, DC, USA, 1993. [Google Scholar]
- Rezende, M.K.A. Evapotranspiração, Coeficientes de Cultivo Simples e Dual do Milho Safrinha Para a Região de Dourados-MS. Bachelor’s Thesis, Universidade Estadual de Maringá, Maringá, Brazil, 2016. [Google Scholar]
- Krauß, M.; Keßler, J.; Prochnow, A.; Kraatz, S.; Drastig, K. Water productivity of poultry production: The influence of different broiler fattening systems. Food Energy Secur. 2015, 4, 76–85. [Google Scholar] [CrossRef]
- CONAB. Série Histórica de Safras (safra e safrinha). Companhia Nacional de Abastecimento (CONAB). Available online: https://www.conab.gov.br/info-agro/safras/serie-historica-das-safras?start=20,2019 (accessed on 12 January 2020).
- Nascimento, F.M.; Bicudo, S.J.; Rodrigues, J.G.L.; Furtado, M.B.; Campos, S. Produtividade de genótipos de milho em resposta à época de semeadura. Rev. Ceres 2011, 58, 193–201. [Google Scholar] [CrossRef]
- CONAB. Acompanhamenot da Safra Brasileira de Grãos—Safra 2019/20—Terceiro Levantamento; Companhia Nacionl de Abastecimento (CONAB): Brasília, Brazil, 2019; p. 106. [Google Scholar]
- Sanches, A.C.; Souza, D.P.d.; Jesus, F.L.F.d.; Mendonça, F.C.; Gomes, E.P. Crop coefficients of tropical forage crops, single cropped and overseeded with black oat and ryegrass. Sci. Agric. 2019, 76, 448–458. [Google Scholar] [CrossRef] [Green Version]
- Ridoutt, B.G.; Page, G.; Opie, K.; Huang, J.; Bellotti, W. Carbon, water and land use footprints of beef cattle production systems in southern Australia. J. Clean. Prod. 2014, 73, 24–30. [Google Scholar] [CrossRef]
- Ibidhi, R.; Hoekstra, A.Y.; Gerbens-Leenes, P.W.; Chouchane, H. Water, land and carbon footprints of sheep and chicken meat produced in Tunisia under different farming systems. Ecol. Indic. 2017, 77, 304–313. [Google Scholar] [CrossRef]
- Nakamura, K.; Itsubo, N. Carbon and water footprints of pig feed in France: Environmental contributions of pig feed with industrial amino acid supplements. Water Resour. Ind. 2019, 21, 100108. [Google Scholar] [CrossRef]
- Bai, X.; Ren, X.; Khanna, N.Z.; Zhang, G.; Zhou, N.; Bai, Y.; Hu, M. A comparative study of a full value-chain water footprint assessment using two international standards at a large-scale hog farm in China. J. Clean. Prod. 2018, 176, 557–565. [Google Scholar] [CrossRef]
- Palhares, J.C.P. Pegada hídrica das aves abatidas no Brasil na década 2000–2010. In 3° Seminário de Gestão Ambiental na Agropecuária; FIEMA: Bento Gonçalves, Brazil, 2012; p. 7. [Google Scholar]
- Palhares, J.C.P. Pegada hídrica dos suínos abatidos nos Estados da Região Centro-Sul do Brasil. Acta Sci. Anim. Sci. 2011, 33, 309–314. [Google Scholar] [CrossRef] [Green Version]
- ANA. Manual de Usos Consuntivos da Água no Brasil; National Water Agency (ANA): Brasilia, Brazil, 2019; p. 75. [Google Scholar]
- Palhares, J.C.P.; Afonso, E.R.; Gameiro, A.H. Reducing the water cost in livestock with adoption of best practices. Environ. Dev. and Sustain. 2019, 21, 2013–2023. [Google Scholar] [CrossRef]
- de Brito, P.L.C.; de Azevedo, J.P.S. Charging for Water Use in Brazil: State of the Art and Challenges. Water Resour. Manag. 2020, 34, 1213–1229. [Google Scholar] [CrossRef]
- Mekonnen, M.M.; Neale, C.M.U.; Ray, C.; Erickson, G.E.; Hoekstra, A.Y. Water productivity in meat and milk production in the US from 1960 to 2016. Environ. Int. 2019, 132, 105084. [Google Scholar] [CrossRef]
- Renault, D.; Wallender, W.W. Nutritional water productivity and diets. Agric. Water Manag. 2000, 45, 275–296. [Google Scholar] [CrossRef]
- Cohn, A.S.; VanWey, L.K.; Spera, S.A.; Mustard, J.F. Cropping frequency and area response to climate variability can exceed yield response. Nat. Clim. Chang. 2016, 6, 601–604. [Google Scholar] [CrossRef]
- Andrea, M.C.d.S.; Dallacort, R.; Tieppo, R.C.; Barbieri, J.D. Assessment of climate change impact on double-cropping systems. SN Appl. Sci. 2020, 2, 544. [Google Scholar] [CrossRef] [Green Version]
- Zhuo, L.; Liu, Y.; Yang, H.; Hoekstra, A.Y.; Liu, W.; Cao, X.; Wang, M.; Wu, P. Water for maize for pigs for pork: An analysis of inter-provincial trade in China. Water Res. 2019, 166, 115074. [Google Scholar] [CrossRef]
- Solaymani, S. CO2 emissions patterns in 7 top carbon emitter economies: The case of transport sector. Energy 2019, 168, 989–1001. [Google Scholar] [CrossRef]
- Vellenga, L.; Qualitz, G.; Drastig, K. Farm Water Productivity in Conventional and Organic Farming: Case Studies of Cow-Calf Farming Systems in North Germany. Water 2018, 10, 1294. [Google Scholar] [CrossRef] [Green Version]
- Palhares, J.C.P.; Morelli, M.; Junior, C.C. Impact of roughage-concentrate ratio on the water footprints of beef feedlots. Agric. Syst. 2017, 155, 126–135. [Google Scholar] [CrossRef] [Green Version]
Term | Data | Source |
---|---|---|
Broiler Production | ||
Input | ||
Number of Heads (head year−1) | 6,108,600 | Farm Survey |
Initial weight (kg) | 0.042 | Cobb [31] |
Production cycle (days) | 42 | Cobb [31] |
Cycles (year) | 6 | Farm Survey |
Output | ||
Finishing weight (kg) | 2.85 | Cobb [31] |
Mass output (kg CW head−1) | 2.15 | Cobb [31]; Embrapa [30] |
Energy output (MJ kg−1) | 4.98 | USDA [34] |
Revenues (R$ kg−1) | 1.61 | Embrapa [35] |
Pig Production | ||
Input | ||
Number of Heads (head year−1) | 55,071 | Farm Survey |
Initial weight (kg) | 28.7 | Agroindustry [33] |
Production cycle (days) | 105 | Agroindustry [33] |
Cycles (year) | 3 | Farm Survey |
Output | ||
Finishing weight (kg) | 134.94 | Agroindustry [33] |
Mass output (kg CW head−1) | 97.16 | Embrapa [32] |
Energy output (MJ kg−1) | 15.72 | USDA [34] |
Revenues (R$ kg−1) | 3.39 | Embrapa [35] |
Indirect Water | ||
Input | ||
Crop mean yield | t ha−1 | IBGE [36] |
Climate data | Parameters 1 | INMET [37] |
Climate classification | RS: Cfb; PR: Cfa; MT: Aw | Alvares et al., [38] |
Type of soil | Clay | Streck et al., [39]; IBGE [40] |
Direct Water | ||
Input | ||
Animal drinking | l head−1 day−1 | Palhares [41] |
Service—cleaning | l head−1 day−1 | Drastig et al. [42]; FEPAM [43] |
Service—cooling | l head−1 day−1 | Drastig et al. [42] |
Crop Rotation | Broiler (WPindirect,broiler,Feed) | Pig (WPindirect,pig,Feed) | ||||
---|---|---|---|---|---|---|
RS | PR | MT | RS | PR | MT | |
Safra (fm-safra) | ||||||
WPindirect,Feed (kg FM m−3) | 0.64 | 0.99 | 0.83 | 0.69 | 1.07 | 0.89 |
Soy (fs-soy) | ||||||
WPindirect,Feed (kg FM m−3) | 0.28 | 0.34 | 0.38 | 0.31 | 0.38 | 0.31 |
Safra (rm-safra) | ||||||
WPindirect,Feed (kg FM m−3) | 0.63 | 0.90 | - | 0.68 | 0.97 | - |
Soy (rs-soy) | ||||||
WPindirect,Feed (kg FM m−3) | 0.28 | 0.31 | - | 0.30 | 0.35 | - |
Safrinha (smf–safrinha) | ||||||
WPindirect,Feed (kg FM m−3) | - | 1.20 | 1.37 | - | 1.30 | 1.48 |
Soy (smf–soy) | ||||||
WPindirect,Feed (kg FM m−3) | - | 0.53 | 0.70 | - | 0.57 | 0.75 |
Broiler (WPindirect,broiler,Feed) | Pig (WPindirect,broiler,Feed) | |||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Crop Rotations | RS | PR | MT | RS | PR | MT | ||||||||||||
M50% | HY | HY50% | M50% | HY | HY50% | M50% | HY | HY50% | M50% | HY | HY50% | M50% | HY | HY50% | M50% | HY | HY50% | |
Safra (fm-safra) | ||||||||||||||||||
WPindirect,Feed (kg FM m−3) | 0.81 | 0.93 | 1.19 | 1.18 | 1.16 | 1.38 | 1.02 | 1.01 | 1.24 | 0.88 | 1.02 | 1.29 | 1.28 | 1.25 | 1.49 | 1.10 | 1.09 | 1.34 |
Soy (fs-soy) | ||||||||||||||||||
WPindirect,Feed (kg FM m−3) | 0.37 | 0.39 | 0.51 | 0.42 | 0.41 | 0.50 | 0.47 | 0.45 | 0.56 | 0.40 | 0.43 | 0.56 | 0.46 | 0.45 | 0.54 | 0.41 | 0.48 | 0.61 |
Safra (rm-safra) | ||||||||||||||||||
WPindirect,Feed (kg FM m−3) | 0.69 | 0.91 | 1.00 | 0.96 | 1.05 | 1.12 | - | - | - | 0.74 | 0.98 | 1.08 | 1.04 | 1.14 | 1.21 | - | - | - |
Soy (rs-soy) | ||||||||||||||||||
WPindirect,Feed (kg FM m−3) | 0.30 | 0.38 | 0.41 | 0.33 | 0.37 | 0.39 | - | - | - | 0.32 | 0.41 | 0.44 | 0.37 | 0.42 | 0.44 | - | - | - |
Safrinha (smf–safrinha) | ||||||||||||||||||
WPindirect,Feed (kg FM m−3) | - | - | - | 1.38 | 1.93 | 2.21 | 1.62 | 1.77 | 2.10 | - | - | - | 1.49 | 2.08 | 2.38 | 1.75 | 1.92 | 2.26 |
Soy (smf–soy) | ||||||||||||||||||
WPindirect,Feed (kg FM m−3) | - | - | - | 0.62 | 0.65 | 0.76 | 0.79 | 0.85 | 0.95 | - | - | - | 0.66 | 0.70 | 0.82 | 0.85 | 0.91 | 1.03 |
Region | Crop Rotation | Broiler | Pig | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
M50% | HY | HY50% | M50% | HY | HY50% | ||||||||
km3 | % | km3 | % | km3 | % | km3 | % | km3 | % | km3 | % | ||
RS | Safra (fm-safra); Soy (fs-soy) | 0.0126 | 22 | 0.017 | 29 | 0.0257 | 45 | 0.0076 | 22 | 0.011 | 31 | 0.016 | 46 |
Safra (rm-safra); Soy (rs-soy) | 0.0047 | 8 | 0.017 | 29 | 0.0205 | 35 | 0.0028 | 8 | 0.010 | 29 | 0.012 | 35 | |
PR | Safra (fm-safra); Soy (fs-soy) | 0.0071 | 17 | 0.006 | 15 | 0.0124 | 30 | 0.0042 | 17 | 0.004 | 15 | 0.007 | 29 |
Safra (rm-safra); Soy (rs-soy) | 0.0026 | 6 | 0.007 | 15 | 0.0093 | 20 | 0.0015 | 6 | 0.004 | 15 | 0.006 | 20 | |
Safrinha (smf–safrinha) Soy (smf-soy) | 0.0040 | 13 | 0.009 | 28 | 0.0115 | 38 | 0.0024 | 13 | 0.005 | 29 | 0.007 | 38 | |
MT | Safra (fm-safra); Soy (fs-soy) | 0.0045 | 19 | 0.007 | 17 | 0.0146 | 33 | 0.0062 | 21 | 0.008 | 26 | 0.012 | 41 |
Safrinha (smf–safrinha) Soy (smf-soy) | 0.0029 | 13 | 0.005 | 20 | 0.0077 | 31 | 0.0021 | 14 | 0.003 | 20 | 0.005 | 31 |
Indicators | Safra (fm-safra); Soy (fs-soy) | Safra (rm-safra); Soy (rs-soy) | Safrinha (smf-safrinha) Soy (smf-soy) | ||||
---|---|---|---|---|---|---|---|
RS | PR | MT | RS | PR | PR | MT | |
Broiler Production | |||||||
Mass | |||||||
WPindirect+direct,broiler,Mass,Farm (kgCW m−3) | 0.23 | 0.31 | 0.30 | 0.22 | 0.28 | 0.43 | 0.53 |
%Blue water/%green water | 0.2%/99.8% | 0.2%/99.8% | 0.2%/99.8% | 0.2%/99.8% | 0.2%/99.8% | 0.3%/99.7% | 0.4%/99.6% |
Energy | |||||||
WPindirect+direct,broiler,Energy,Farm (MJ m−3) | 1.15 | 1.57 | 1.50 | 1.12 | 1.42 | 2.15 | 2.62 |
Revenues | |||||||
WPindirect+direct,broiler,Mon,Farm (R$ m−3) | 0.49 | 0.67 | 0.64 | 0.48 | 0.61 | 0.92 | 1.12 |
Pig Production | |||||||
Mass | |||||||
WPindirect+direct,pig,Mass,Farm (kg CW m−3) | 0.16 | 0.21 | 0.18 | 0.15 | 0.19 | 0.29 | 0.35 |
%Blue water/%green water | 0.2%/99.8% | 0.3%/99.7% | 0.3%/99.7% | 0.2%/99.8% | 0.3%/99.7% | 0.4%/99.6% | 0.5%/99.5% |
Energy | |||||||
WPindirect+direct,pig,Energy,Farm (MJ m−3) | 2.44 | 3.36 | 2.83 | 2.37 | 3.05 | 4.56 | 5.51 |
Revenues | |||||||
WPindirect+direct,pig,Mon,Farm (R$ m−3) | 0.73 | 1.01 | 0.85 | 0.71 | 0.91 | 1.36 | 1.65 |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Share and Cite
Carra, S.H.Z.; Palhares, J.C.P.; Drastig, K.; Schneider, V.E. The Effect of Best Crop Practices in the Pig and Poultry Production on Water Productivity in a Southern Brazilian Watershed. Water 2020, 12, 3014. https://doi.org/10.3390/w12113014
Carra SHZ, Palhares JCP, Drastig K, Schneider VE. The Effect of Best Crop Practices in the Pig and Poultry Production on Water Productivity in a Southern Brazilian Watershed. Water. 2020; 12(11):3014. https://doi.org/10.3390/w12113014
Chicago/Turabian StyleCarra, Sofia Helena Zanella, Julio Cesar Pascale Palhares, Katrin Drastig, and Vania Elisabete Schneider. 2020. "The Effect of Best Crop Practices in the Pig and Poultry Production on Water Productivity in a Southern Brazilian Watershed" Water 12, no. 11: 3014. https://doi.org/10.3390/w12113014
APA StyleCarra, S. H. Z., Palhares, J. C. P., Drastig, K., & Schneider, V. E. (2020). The Effect of Best Crop Practices in the Pig and Poultry Production on Water Productivity in a Southern Brazilian Watershed. Water, 12(11), 3014. https://doi.org/10.3390/w12113014