mnqoliveira / future-sugar

Source code for the analysis submitted to the Brazilian Bioenergy Science and Technology Conference (BBEST 2017). Abstract in folder 'publications'.

Future sugar yield | Sugarcane

• Abstract citation: OLIVEIRA, M. P. G. de; BOCCA, F. F. ; RODRIGUES, L. H. A. . Classification trees for delving into sugarcane production in Brazil under climate change. In: BBEST 2017 - Brazilian Bioenergy Science and Technology Conference, 2017, Campos do Jordão, SP. Proceedings of the Brazilian Bioenergy Science and Technology Conference 2017 (BBest2017), 2017. p. 59-60.

• Poster here: Oliveira, Monique; Rodrigues, Luiz Henrique; Bocca, Felipe Ferreira (2017): Classification trees for delving into sugarcane production in Brazil under climate change. figshare. Poster. https://doi.org/10.6084/m9.figshare.14695527.v2

Parts of this work received contributions from @waroce and from @boccaff. I am now sharing it as it may be useful for someone else. I would only like an acnowledgement in case it is used. :)

Methods

Variables

As much as possible, the recommended procedures stated by White et al. (2011)¹, regarding assessments of climate change in agroecosystems, were followed. The variables that were modified in the simulations are summarized in Table 1.

Table 1. Variables that were changed in the simulations.

Category	Configuration
Models	DSSAT-Canegro (Inman-Bamber, 1991; Jones et al., 2003; Singels et al., 2008)2 3 4 and APSIM-Sugar (Keating et al., 1999)5
Locations	Jataí (GO), Piracicaba (SP) and Presidente Prudente (SP)
Climate scenarios	GCMs: ACCESS1-0, bcc-csm1-1, BNU-ESM, CanESM2, CCSM4, CESM1-BGC, CSIRO-Mk3-6-0, GFDL-ESM2G, GFDL-ESM2M, HadGEM2-CC, HadGEM2-ES, inmcm4, IPSL-CM5A-LR, IPSL-CM5A-MR, MIROC5, MIROC-ESM, MPI-ESM-LR, MPI-ESM-MR, MRI-CGCM3, NorESM1-M, FGOALS-g2, CMCC-CM, CMCC-CMS, CNRM-CM5, HadGEM2-AO, IPSL-CM5B-LR
	CMIP5 Representative Concentration Pathways: RCP 4.5 and RCP 8.5
	Decades: From 2010 to 2039 and from 2040 to 2069.
	Generation Methods: Delta and Mean and variability
Soil	Dataset 2 (Marin et al., 2011a)6: yellow with predominance of clay, higher water holding capacity, smaller hydraulic conductivity, runoff potential moderately high, 3% slope and well drained.
	Dataset 3 (Marin et al., 2011a)6: red with predominance of sand, lower water holding capacity, higher hydraulic conductivity, runoff potential moderately low, 3% slope and well drained.
Fertilizers	Nutrient limitation was not simulated
Irrigation	Rainfed and irrigated. Irrigation was set to be applied when available soil water - calculated up to 1 m of depth - was lower than 40%. One rule was included for no irrigation if it should happen in the last 98 days before harvesting.
Dates	Planting dates in May 15th, August 15th and November 15th and one-year cycles.
Cultivar	RB867515 (Marin et al., 2015)7

Climate data

Daily maximum and minimum temperature and relative humidity data from 1980 to 2009 were retrieved from AgMerra (Ruane et al., 2015)⁸ for three locations: Piracicaba, where sugarcane has been traditionally grown, and Jatai and Presidente Prudente, both in expansion regions (Caldarelli e Gillio 2018; Spera et al 2017)^{9 10}. Climate change scenarios were generated using the AgMIP Climate Scenario Generation Tools with R, made available by the Agricultural Model Intercomparison and Improvement Project (AgMIP) (Ruane et al., 2015)⁸, for the CMIP5 IPCC Scenarios. Both the delta method and the mean and variability method were used with a baseline from 1980 to 2009. No additional downscaling was performed. Carbon dioxide values were defined as 360 ppm for the baseline. Other scenarios follow the values of Table 2.

Table 2. Concentrations of carbon dioxide in the atmosphere used in simulations.

	Near-term (2010 to 2039)	Mid-century (2040 to 2069)
RCP 4.5	423 ppm	499 ppm
RCP 8.5	432 ppm	571 ppm

Simulations

In order to represent, for each location, one condition, we combined the outputs for each type of soil, irrigation scenario and planting dates using an weighted average obtained by the following proportions:

• The grouping by planting dates followed (Marin 2013)¹¹: the final value accounted for 28 % of the output obtained for the simulations in May, 44% for the simulation in August and 28%, for the simulations in November. Even though sugarcane cycle in the first year may be of 12 or 18 months, only 12 months were simulated, following Marin 2013¹¹.
• The two soil types, representing higher or lower water holding capacity, were combined based on the assumption that the soil proportion in the mill area would follow the proportion in the cities (IBGE, 2006)¹². This led to 67.2% of clayey in Jatai, 29.8 % in Piracicaba and 100% in Presidente Prudente. The remainder accounted for sandy soils.
• As for irrigation, the current situation was represented by 27% of irrigated sugarcane area in Jataí, 4% in Piracicaba and 18% in Presidente Prudente (ANA, 2017)¹³.
• To account for the decline in yield that happens each ratoon, since only planted sugarcane was simulated, biomass outputs were reduced in 20% up to the 5th growth for rainfed areas and 15% for irrigated areas.
• We only used the calibration from the RB867515 cultivar Marin et al. (2015)⁷ because other cultivars had not been calibrated for both DSSAT and APSIM. The last comprehensive survey (IAC, 2017)¹⁴ pointed this cultivar as the one that is often most planted in all Brazilian states.

A few points are worth highlighting. First of all, both DSSAT and APSIM take into account the effects of increased carbon dioxide, as described in Marin et al. (2015)⁷. According to Marin et al. (2015)⁷, these are the most widely used models for sugarcane growth. Furthermore, even though Brazilian production is mostly rainfed (Walter et al., 2014)¹⁵, because we simulated these scenarios for areas known as expansion areas (Scarpare et al., 2016)¹⁶, the effect on irrigated sugarcane was also evaluated. We modified both codes to set irrigation rules in accordance to our method. Since we were evaluating sugar content and water deficit is important for the storage of sugar content, irrigation closer to harvesting was not allowed. The limit was set to 98 days before harvest.

Finally, Marin et al. (2015)⁷ suggest more targeted experiments to understand the physiology Brazilian cultivars so that they can be better represented in sugarcane models for application in Brazil. This is even more relevant given the development of new varieties adapted to new environments. Given these studies have not yet been published, the values used still depend on the authors’ calibration.

References

Footnotes

1. White, J.W., Hoogenboom, G., Kimball, B.A., Wall, G.W., 2011. Methodologies for simulating impacts of climate change on crop production. F. Crop. Res. 124, 357–368. doi:10.1016/j.fcr.2011.07.001 ↩

2. Inman-Bamber, N., 1991. A growth model for sugar-cane based on a simple carbon balance and the CERES-Maize water balance. South African J. Plant Soil 37–41. doi:10.1080/02571862.1991.10634587 ↩

3. Jones, J.W., Hoogenboom, G., Porter, C.H., Boote, K.J., Batchelor, W.D., Hunt, L. a., Wilkens, P.W., Singh, U., Gijsman, a. J., Ritchie, J.T., 2003. The DSSAT cropping system model. Eur. J. Agron. 18, 235–265. doi:10.1016/S1161-0301(02)00107-7 ↩

4. Singels, A., Jones, M., Berg, M. Van Den, 2008. DSSAT v4. 5 Canegro sugarcane plant module: scientific documentation. South African Sugarcane Res. Inst. Mt. … 1–34. ↩

5. Keating, B.. a., Robertson, M.. J., Muchow, R.. C., Huth, N.. I., 1999. Modelling sugarcane production systems I. Development and performance of the sugarcane module. F. Crop. Res. 61, 253–271. doi:10.1016/S0378-4290(98)00167-1 ↩

6. Marin, F.R., Jones, J.W., Royce, F., Suguitani, C., Donzeli, J.L., Filho, W.J.P., Nassif, D.S.P., 2011a. Parameterization and Evaluation of Predictions of DSSAT/CANEGRO for Brazilian Sugarcane. Agron. J. 103, 304. doi:10.2134/agronj2010.0302 ↩ ↩²

7. Marin, F.R., Thorburn, P.J., Nassif, D.S.P., Costa, L.G., 2015. Sugarcane model intercomparison: Structural differences and uncertainties under current and potential future climates. Environ. Model. Softw. 72, 372–386. doi:10.1016/j.envsoft.2015.02.019 ↩ ↩² ↩³ ↩⁴ ↩⁵

8. Ruane, A.C., Goldberg, R., Chryssanthacopoulos, J., 2015. Climate forcing datasets for agricultural modeling: Merged products for gap-filling and historical climate series estimation. Agric. For. Meteorol. 200, 233–248. doi:10.1016/j.agrformet.2014.09.016 ↩ ↩²

9. Caldarelli, C.E., Gilio, L., 2018. Expansion of the sugarcane industry and its effects on land use in São Paulo: Analysis from 2000 through 2015. Land use policy 76, 264–274. doi:10.1016/j.landusepol.2018.05.008 ↩

10. Spera, S., VanWey, L., Mustard, J., 2017. The drivers of sugarcane expansion in Goiás, Brazil. Land use policy 66, 111–119. doi:10.1016/j.landusepol.2017.03.037 ↩

11. Marin, F.R., Jones, J.W., Singels, A., Royce, F., Assad, E.D., Pellegrino, G.Q., Justino, F., 2013. Climate change impacts on sugarcane attainable yield in southern Brazil. Clim. Change 117, 227–239. doi:10.1007/s10584-012-0561-y ↩ ↩²

12. IBGE, 2006. Mapa de Solos do Brasil. Link ↩

13. ANA - Agência Nacional de Águas, 2017. Levantamento da cana-de-açúcar irrigada na região Centro-Sul do Brasil. Link ↩

14. IAC, 2017. Censo varietal IAC de cana-de-açúcar na região Centro-Sul do Brasil - Safra 2016/2017. Link ↩

15. Walter, A., Galdos, M.V., Scarpare, F.V., Leal, M.R.L.V., Seabra, J.E.A., da Cunha, M.P., Picoli, M.C.A., de Oliveira, C.O.F., 2014. Brazilian sugarcane ethanol: Developments so far and challenges for the future. Wiley Interdiscip. Rev. Energy Environ. 3, 70–92. doi:10.1002/wene.87 ↩

16. Scarpare, F.V., Hernandes, T.A.D., Ruiz-Corrêa, S.T., Picoli, M.C.A., Scanlon, B.R., Chagas, M.F., Duft, D.G., Cardoso, T. de F., 2016. Sugarcane land use and water resources assessment in the expansion area in Brazil. J. Clean. Prod. 133, 1318–1327. doi:10.1016/j.jclepro.2016.06.074 ↩