Advertisement

Regional Environmental Change

, Volume 18, Issue 3, pp 873–883 | Cite as

Climate change impact on the potential yield of Arabica coffee in southeast Brazil

  • Priscila da Silva Tavares
  • Angélica Giarolla
  • Sin Chan Chou
  • Adan Juliano de Paula Silva
  • André de Arruda Lyra
Original Article

Abstract

The Intergovernmental Panel on Climate Change (IPCC) projections of global mean temperature rises are worrisome for coffee crop due to the intolerance of the Arabica species to high air temperature variations. The crop has a large participation in the Brazilian trade balance; therefore, in this study, the impacts of climate change on the potential yield of Arabica coffee (Coffea arabica L.) were assessed in the areas of Southeast Brazil in future climate change scenarios. Simulations of the Eta Regional Climate Model at 5-km resolution used in this study were generated from a second dynamic downscaling of the HadGEM2-ES model runs. The projections adopted two scenarios of greenhouse gas concentration, the RCP4.5 and RCP8.5, and considered the period 2011–2100. The projections indicated a large reduction of about 20 to 60% of the areas currently suitable for coffee cultivation in Southeast Brazil. In the RCP8.5 scenario, at the end of century, coffee cultivation is suitable only in elevated mountain areas, which would pose difficulties to farming management due to the operation of agricultural machinery in mountain areas. In addition, coffee cultivation in these regions could produce environmental impacts in the remnant Brazilian Atlantic Forest. Areas of high climatic risk increase due to temperature increase. The projections showed that the potential yield could be reduced by about 25% by the end of the twenty-first century. These results of potential coffee yield in the future climate indicate a need for adaptation studies of Arabica coffee cultivation.

Keywords

Climate scenarios Agroclimatic zoning Arabica coffee Eta model Brazil 

Introduction

The coffee agribusiness involves about half a billion people (DaMatta et al. 2007), approximately 8% of the world population, in a productive chain that goes from the field to the drink. It is one of the most valuable commodities marketed in the world and, in a few years, has surpassed the value of oil (ICO 2015). It is estimated that the production and commercialization of the grain handles, annually, is more than 90 billion USD (ICO 2015; DaMatta et al. 2007; Kobayashi 2007).

For over a century and a half, Brazil has been the leader in the ranking of the largest producer and exporter (Baliza et al. 2013), which currently contributes one third of all world production, with an amount of more than 48 million bags of coffee benefited, according to the harvest data for the period 2015/2016 (ICO 2017). The grain production chain generates, on average, more than 8 million jobs, provides income and access to health and education for workers and their families, and represents a factor for maintaining the workforce in rural areas (MAPA 2013). Family farming plays an important role in coffee production, which accounts for 37% of Brazilian production. It is estimated that more than 80% of rural properties that produce coffee in the country are family establishments, which makes the coffee sector more vulnerable to the effects of climate change. The impacts are mainly due to the low adaptation capacity of family coffee growers and the low tolerance of Arabica coffee to high temperatures. The Arabica coffee is the most commercialized species in Brazil and in the world.

Most of the oscillations in coffee yield are mainly associated with climatic factors because the crop needs favorable conditions throughout its vegetative and reproductive cycle. A region that meets climatic requirements, especially with respect to air temperature and water demand, is considered suitable for growing the plant. The optimum thermal conditions for Arabica coffee in Brazil are those with an annual mean temperature between 18 and 22.5 °C (Camargo 1977). Regions where the average annual temperatures range between 22.5 and 24 °C are considered areas restricted to coffee cultivation. In regions where the average annual temperatures exceed 24 °C, fruit maturation occurs early and it generally occurs during the rainy season when harvesting and drying the coffee is difficult due to frequent rainfall. This situation contributes to the degradation of the product quality due to undesired fermentations of fruit pulp (Pereira et al. 2008). Thus, warm regions are considered unsuitable for growing coffee due to the excessive heat. Moreover, on the other end of the extreme temperature, regions with annual average temperatures below 18 °C are unsuitable for coffee cultivation and, therefore, those regions are also classified as inadequate due to the thermal restriction. According to Pereira et al. (2008), under these thermal conditions, the plants cannot complete the phenological cycle before the first flowering, a situation that undermines the life cycle of the crop and contributes to the absence of a rest period, which is necessary for the induction of the new floral buds. According to the authors, temperature damages the new flowering, which ruins the next harvest, due to the process of fruits in maturation mixed with new flowers (Pereira et al. 2008). In the case of water deficit, Arabica coffee can tolerate deficits up to 150 mm per year. Some periods of water deficit can be beneficial to coffee when it coincides with the dormancy period of the crop (Thomaziello et al. 2000). However, annual water deficits greater than 150 mm can substantially affect the longevity of the crop.

In this context, the delimitation of regions with favorable climatic and pedological potential, through the use of agroclimatic zoning, is considered a crucial tool in agricultural planning (Meireles et al. 2007). Some studies carried out based on agroclimatic zoning showed direct impacts of climate change on the areas currently suitable for coffee growth (Bragança et al. 2016; Ranjitkar et al. 2016; Bunn 2015; Bunn et al. 2015; Pinto et al. 2008; Assad et al. 2004). In Brazil, highly productive regions, such as the states of Minas Gerais, São Paulo, and Espírito Santo, would present a high climatic risk if no adaptive measures are taken (Pinto et al. 2008; Zullo et al. 2006; Assad et al. 2004). According to Pinto et al. (2008), the financial losses would exceed 2.55 billion BRL and 3 billion BRL in 2070, considering the optimistic and pessimistic scenarios of climate change, respectively. Although these studies represented a milestone in the assessment of the risks that Brazilian coffee could face due to climate change, most of them considered the increase in the average air temperature homogeneously in the study area. The limitation in considering the increase of the average global temperature for any region of the world is because temperature varies in local and temporal scales. The grid size of global climate models (GCMs) ranges from 100 to 200 km, which is coarse to capture local characteristics, such as topography, watersheds, coastlines, etc., and especially agroclimatic zoning that depends on microclimatic conditions.

In studies that consider smaller regions, such as in agricultural areas, the use of regional models may provide more refined results since they allow horizontal resolutions of the order of tens of kilometers and consequently allow explicit simulations of mesoscale processes with better representation of local characteristics. In the case of Arabica coffee, located in complex topography areas in Brazil, such as in the southern region of Minas Gerais state, the use of regional models can contribute to more accurate zoning and, consequently, more careful evaluations of the impacts of climate change on the crop. One of the recent works with dynamic downscaling techniques for Brazil through regional climate models was carried out by Chou et al. (2014a, b). The results of Chou et al. (2014a), generated using the eta model nested in a global climate model, captured better the intensity and frequency of extreme events of precipitation. The projections of the future climate, in the case of the most pessimistic scenario, indicated that the increases in air temperature can exceed 7 °C in some regions of Brazil (Chou et al. 2014b). In global average terms, the temperature increase, in the most pessimistic scenario, would not exceed 6 °C, which shows the importance of the use of downscaling projections when assessing the impacts of climate change on smaller scales.

The objective of this work is to assess the impacts of climate change on the potential yield of Arabica coffee in areas of Southeast Brazil, through agroclimatic zoning and considering the scenarios projected by Intergovernmental Panel on Climate Change—Assessment Report 5 (IPCC-AR5 2013). Because coffee is grown in mountain areas, to better capture the underlying surface characteristics, simulations at very high horizontal resolutions were used. The eta model was selected because it has suitable and refined representation of topography. In addition, the model was used for the elaboration of the Third National Communication of Brazil to the United Nations Framework Convention on Climate Change (UNFCCC).

Materials and methods

The estimation of the potential yield of Arabica coffee, both in the present and in the future climate, was carried out through the agroclimatic zoning based on the simulations of the eta model. The agroclimatic zoning and the estimation of the potential yield, as well as the eta model and the selected climate change scenarios are described.

Agroclimatic zoning

The agroclimatic zoning follows the methodology proposed by Luppi et al. (2014), which considers the annual mean water and temperature restrictions of the Arabica coffee (Camargo 1977; Matiello 1991) and relates the suitability into three categories (Table 1).
Table 1

Suitability ranges based on annual mean temperature and water deficit for Arabica coffee (source—adapted from Luppi et al. 2014)

Suitability

Temperature (°C)

Water deficit (mm)

Score

Adequate

18.0–22.5

< 150

1

Restricted

22.5–24.0

150–200

2

Inadequate

< 18 and > 24

> 200

3

In the method of Luppi et al. (2014), adequate, restricted, and inadequate areas for coffee cultivation receive scores (S) 1, 2, and 3, respectively, for each climatic condition: temperature or water deficit (Table 1), and then expressed in percentages of suitability after Eq. 1 and Eq. 2. Inadequate areas receive 0% of suitability (Eq. 2).

For S = 1 or S = 2,
$$ {PS}_{CC}=\frac{100}{2S} $$
(1)
However, for S = 3
$$ {PS}_{CC}=0 $$
(2)
where PS CC is the percentage of suitability for each climatic condition (CC) and S is the score of the suitability class (Table 1). This classification according to the percentage was used as a quantitative measure which contributes to a better understanding of the suitability. The agroclimatic zoning (AZ), expressed as the percentage of crop suitability, is given by the sum of the suitability of the air temperature (PSTemp) plus the suitability by water deficit (PSWD) (Eq. 3). Table 2 shows the new proposed classification for the agroclimatic zoning of the Arabica coffee.
$$ AZ={PS}_{\mathrm{Temp}}+{PS}_{\mathrm{WD}} $$
(3)
Table 2

Percentage of suitability for the agroclimatic zoning of coffee cultivation (source—adapted from Luppi et al. 2014 and Eugenio et al. 2014)

Score

Suitability

PS

Agroclimatic zoning (%)

Classes

Temperature

Water deficit

Temperature

Water deficit

Temperature (%)

Water deficit (%)

1

1

Adequate

Adequate

50

50

100

Fully adequate

1

2

Adequate

Restricted

50

25

75

Adequate

1

3

Adequate

Inadequate

50

0

50

Regular

2

1

Restricted

Adequate

25

50

75

Adequate

2

2

Restricted

Restricted

25

25

50

Regular

2

3

Restricted

Inadequate

25

0

25

Restricted

3

1

Inadequate

Adequate

0

50

50

Regular

3

2

Inadequate

Restricted

0

25

25

Restricted

3

3

Inadequate

Inadequate

0

0

0

Inadequate

In summary, the areas fully adequate (100%) are those areas where both thermal and water conditions are favorable to the cultivation of Arabica coffee; areas with restriction to only one of the climatic conditions are classified as being 75% suitable and therefore are adequate areas; and inadequate areas (0%) are those that the climatic characteristics are not appropriate to the commercial exploitation of the crop due to the great thermal and water limitations. The method adds two classes of suitability, areas with 50 and 25%. In the first case (50%), the crop can occasionally find suitability for one of the climatic conditions or restriction for both conditions; this is classified as regular. In the production for commercial coffee, it is expected that the restriction factors be controlled. In the case of 25%, the crop is restricted and unsuitable for some of the climatic conditions, which makes this class restricted for coffee cultivation due to the higher risk of harvest loss.

The annual water deficit is estimated by the climatological water balance (CWB) using the method of Thornthwaite and Mather (1957), with modifications by Mendonça (1958). This water balance considers a soil water storage of 100 mm (Meireles et al. 2009). Potential evapotranspiration (mm), required for CWB estimation, is calculated by the FAO Penman–Monteith method (FAO 56. PM; Allen et al. 1998). This method estimates the potential evapotranspiration for a reference crop (ETo) and it is considered the most appropriate since it incorporates the aerodynamic and thermodynamic aspects (Carvalho et al. 2011; Sediyama 1996).

Potential yield

The potential yield of Arabica coffee was calculated according to the climatic suitability, following equation (Bunn 2015):
$$ {Y}_{ij}={Y}_{\max, j}\left(1-{\left(1-{AZ}_{ij}\right)}^2\right) $$
(4)
where Y ij is potential yield (kg/ha) in the grid point of the model, Y max,j is the maximum allowable yield and AZ ij is the agroclimatic zoning index (%). In the literature, regions with full development of Arabica coffee (100% fully adequate) showed yields around 2000 kg/ha (Wintgens 2004). These values are found in the optimum areas for coffee development in Brazil, such as in the State of Minas Gerais (Fig. 1). In regions of restricted climatic conditions, the potential yield is considered to be approximately half of the optimum value (Bunn 2015).
Fig. 1

Yield of coffee (kg/ha), measured as weight of whole berries, of a Brazilian municipalities, for the period 1990 to 2015, with emphasis on b the state of Minas Gerais. Source: adapted from IBGE-PAM ( 2016) (Municipal Agricultural Production/Table 106)

Eta regional climate model runs

The eta model is a complex atmospheric numerical model in the representation of the physical and dynamic processes of the atmosphere and explicitly includes the mesoscale processes. It is in the category of atmospheric models called regional or limited area and therefore they need global models to provide information of the atmosphere in the lateral boundaries. One of the characteristics of the model is the use of the vertical eta coordinate, which gives the name to the model. This coordinate is considered more suitable for simulations in regions with complex topography, which favors a better representation of microclimates as observed in the coffee regions of southeast Brazil.

The model dynamics are in finite volume (Janjić 1984; Mesinger et al. 2012). Deep and shallow convections are parametrized by the Betts–Miller scheme (Betts and Miller 1986) modified by Janjić (1994). Cloud microphysics follow the Zhao scheme (Zhao et al. 1997). The land-surface processes are represented by the NOAH scheme (Ek et al. 2003) with annual cycle of vegetation greenness. Vegetation distinguishes 13 types and soil 9 types. Radiative fluxes are treated by the Lacis–Hansen scheme (Lacis and Hansen 1974) for short waves and Fels–Schwarzkopf scheme (Fels and Schwarzkopf 1975) for long waves. The surface layer is based on the Monin–Obukhov similarity theory and Paulson (Paulson 1970) stability functions. The model solves the turbulent mixing in the atmosphere through the Mellor–Yamada scheme (Mellor and Yamada 1982) at a closure level of 2.5. The recent updates of the model are described in Mesinger et al. (2012), Chou et al. (2014a), and Lyra et al. (2017).

Adaptations in the eta model to generate scenarios of climate change over Brazil were made, initially, by Pesquero et al. (2010), Chou et al. (2012), and Marengo et al. (2012). Recently, Chou et al. (2014a, b) performed multidecadal downscaling simulations using eta model nested in three global climate models available from CMIP5 (Coupled Model Intercomparison Project Phase 5). The simulations at 20-km resolution exhibited reasonable skill in the representation of the present climate of South America and the projections of the future climate were used to support the Third National Communication of Brazil to the UNFCCC (MCTI 2016). Lyra et al. (2017) performed a second nesting of the eta model in 5-km horizontal resolution over the southeastern region of Brazil from the simulations generated by Chou et al. (2014a, b). Very-high-resolution runs have improved the present climate simulations, especially in regions of complex topography and for the extreme values of the frequency distribution of temperature and precipitation in the region. Therefore, it is expected that those projections can be used in studies of climate change that require higher resolutions, such as in coffee crop areas.

Simulations of the non-hydrostatic version of the eta model, with horizontal resolution of 5 km, double nested in HadGEM2-ES model, generated by Lyra et al. (2017) were used. Future climate simulations (2011 to 2100) are based on the scenarios of concentrations of greenhouse gases called representative concentration pathway (RCP) (Moss et al. 2010). It was considered an intermediate concentration scenario (RCP4.5) and one of high concentration (RCP8.5), similar to the previous scenarios B1 and A1F1 (IPCC AR4), respectively. In the RCP4.5 scenario, it is assumed that the earth system will store 4.5 W m−2 additional, with equivalent carbon dioxide (CO2) concentration of approximately 650 ppm, showing stabilization after the end of the twenty-first century. While in the RCP8.5 scenario, it is assumed a storage of 8.5 W m−2 with equivalent CO2 concentration above 1300 ppm by the end of the century. The changes occurred in the future scenarios were evaluated through the comparison with the baseline period (1961–1990), also called present climate.

The domain of the model covers part of southeast Brazil and encompasses the main coffee areas of the region, such as the south of Minas Gerais state—SMG (Fig. 2). The state of Minas Gerais is the main Arabica coffee producer in Brazil due to favorable climatic conditions for the growth of this species. The 5-km eta model runs are selected because they are the highest resolution climate runs available in Brazil. It is expected that the 5-km grid size will contribute to better detail the agroclimatic zoning of crop coffee. In Fig. 2, the 5-km eta model domain is shown, with emphasis on the SMG region (shaded), and the positions of the observations considered in the evaluation of the model.
Fig. 2

The domain of the eta model at 5 km, with emphasis on the southern region of Minas Gerais (SMG), shaded. Rainfall (closed circles) and other data (closed triangles) used for evaluation in the SMG

Fifteen stations were selected based on a better temporal homogeneity of the historical series available for the period 1961–2005. Data from the National Water Agency (ANA) and the National Institute of Meteorology (INMET) were used. Rainfall data (mm) and data of mean temperature (°C), minimum temperature (°C), maximum temperature (°C), wind speed (m s−1), solar radiation (MJ m−2), and surface pressure (hPa) were used to estimate ETo. This variable was used to evaluate ETo computed using eta model outputs. Data of solar radiation of the Satellite Application Facility in Climate Monitoring (CMSAF) (Karlsson et al. 2012) and surface pressure of the Climate Forecast System Reanalysis (CFSR) (Saha et al. 2010) were used to compose database to the ETo estimation for some stations with incomplete series.

The eta evaluation was carried out by comparing the time series of the model output against time series of station observations. The simulations of precipitation, ETo, and 2-m temperature were evaluated through the monthly means, the mean error (BIAS), the mean absolute error (MAE), the square root of the mean square error (RMSE), and the correlation coefficient (r).

Downscaled climate results often have a systematic bias which would require correction; however, in this study, the bias correction was not applied because the method used to estimate potential evapotranspiration (ETo) requires some variables, as cited above, that do not have available consistent long period historical dataset and with good spatial distribution. The correction of only some variables, for example precipitation and temperature, may lead to inconsistencies in the representation of the final agroclimatic zoning.

Results and discussion

Baseline evaluation

The state of Minas Gerais leads the ranking of coffee-producing states (Souza et al. 2012) with more than 50% of all national production, being the south of the state the main producing region of Arabica coffee (CONAB 2015). Eta model evaluations of the baseline period are performed using historical series of stations located in the SMG region.

The evaluation is based on the annual cycles of potential evapotranspiration, mean air temperature, and precipitation (Fig. 3). These variables are chosen because they play an important role in the estimation of the annual water deficit and in the agroclimatic zoning. Precipitation and temperature are evaluated in the period from 1961 to 2005, whereas evapotranspiration is evaluated in a shorter period, 2000 to 2005, due to the unavailability to the consistent historical series. The evaluation was computed considering an average of the grid points of the model extracted for each selected station.
Fig. 3

Annual cycle of a potential evapotranspiration (mm), b precipitation (mm), and c air temperature (°C) for the southern region of Minas Gerais (SMG). The right-hand axis in (a), (b), and (c) refers to the errors (MAE, BIAS, and RMSE) of the eta model at 5 km

The evapotranspiration variability (ETo) calculated using eta model outputs is reasonably reproduced, but the values are overestimated (Fig. 3a). In the months of austral winter, June, July, and August, where temperature and evapotranspiration is lower, ETo estimates are closer to observations, and the errors range around 38 and 44 mm/month. The mean variation of rainfall in the SMG region is well reproduced by the simulations, which show more intense values in the summer months and less intense in the winter months (Fig. 3b).

However, in these simulations, the beginning of the driest period occurs too early and causes large errors in the months between July and October. These errors vary between 57 and 80 mm/month. The SMG region exhibits a thermal amplitude of approximately 7 °C, which is well represented by the model. The highest mean temperature occurs in January, of about 22 °C, and the lowest in June, of about 16 °C (Fig. 3c). In months between August and November, the mean errors are close to zero and the mean absolute errors are close to 1 °C.

Table 3 summarizes the mean error (BIAS), the mean absolute error (MAE), the root mean square error (RMSE), and the correlation coefficient (r). In general, the model performs well and shows high correlation coefficient and small BIAS. The ETo and precipitation showed significant errors; however, they exhibited high values of correlation, indicating the reasonable performance of the model in the simulations of these variables.
Table 3

Mean error (BIAS), absolute mean error (MAE), root mean square error (RMSE), and correlation coefficient (r) of potential evapotranspiration (ETo), precipitation (Prec), and temperature (Temp) between the simulations of the Eta model at 5 km and the observations for the southern region of Minas Gerais (SMG)

Variable

BIAS

MAE

RMSE

r

ETo (mm)

58.2

59.3

67.7

0.89

Prec (mm)

35.3

89.9

114.8

0.95

Temp (°C)

− 0.3

1.6

1.9

0.98

Climate projections

The downscaling projections at 5-km resolution over southeast Brazil under RCP4.5 and RCP8.5 scenarios are shown in Fig. 4. The baseline climate simulations are included to allow comparison. The air temperature and water deficit are averaged in 30-year timeslices: 1961–1990, 2011–2040, 2041–2070, and 2071–2100.
Fig. 4

Annual mean temperature (°C) at a height of 2 m above ground level (a) and water deficit (mm) (b) simulated by the eta model at 5 km for the baseline period and future timeslices, under RCP4.5 and RCP8.5 scenarios

The projections show large changes in the air temperature in both RCP scenarios. In most of the area, where the temperature varies approximately between 18 and 26 °C in the baseline period, the temperature increases about 4 to 6 °C. An area subjected to change in terms of higher warming is located in much of São Paulo, southwest of Minas Gerais, northeast of Mato Grosso do Sul, and south of Goiás, between 18° S and 22° S and 48° W and 53° W. In this area, the warming may reach more than 8 °C by the end of the twenty-first century (2071–2100) in the highest equivalent CO2 concentration scenario. In the south of Minas Gerais and in the north of São Paulo, between 20° S and 23° S, and 44° W and 48° W, important coffee areas, where coffee growth currently occurs with temperature around 18–22 °C, may become restricted or inadequate due to the warming. However, in some mountainous regions, coffee cultivation will become suitable for temperature, but regions are often difficult for management.

The projections of the water deficit show less alarming changes. In the present climate, almost all regions that currently show adequate values for the growth of Arabica coffee, that is, below 150 mm/year, exhibit a wider range of restrictive and inadequate areas only at the end of the twenty-first century, considering the RCP8.5 scenario. In the case of the RCP4.5 scenario, the water deficit values throughout the twenty-first century remain, at great part, adequate to the maximum restricted to coffee cultivation, that is, between 150 and 200 mm/year.

Agroclimatic zoning

The agroclimatic zoning for Arabica coffee cultivation for the baseline period shows that the region exhibits 50 to 100% of suitability, that is, it varies from regular to fully adequate (Fig. 5). The restrictive factors are associated only with the air temperature, with thermal excess in the west of the region, and thermal deficit on mountainous areas. Throughout the century, the fully adequate and adequate areas of coffee production reduced considerably in both climatic scenarios, particularly in the RCP8.5 scenario. The increase of restrictive (25%) and inadequate areas is due to the increase of the air temperature of about 4 to 8 °C until the end of the twenty-first century.
Fig. 5

Agroclimatic zoning for the cultivation of Arabica coffee in Southeast Brazil, simulated by eta model at 5 km. Baseline period and future climate, RCP4.5 and RCP8.5, scenarios

As reported in previous studies (Assad et al. 2004; Pinto et al. 2008), in the higher equivalent CO2 concentration scenario, coffee cultivation will be restricted to more elevated areas. Differently from RCP8.5, in the RCP4.5 scenario there are no large inadequate areas, although the restricted and regular areas (25–50%) increase. The areas classified as regular (50%) show suitability due to one of the climatic conditions or restriction on both conditions, suggesting that Arabica coffee cultivation would not be totally unfeasible; however, for commercial purposes, it is required that restrictive factors are attenuated.

Impact on coffee in the southern Minas Gerais

Figure 6 shows the Arabica coffee suitability, in percentage of area, in the SMG region during the present climate and in the projections under the future RCP4.5 and RCP8.5 scenarios. The SMG region exhibits, in general, optimum conditions for coffee development, as 72% of the area are fully adequate and adequate. Throughout the twenty-first century, the reductions of areas with no climate restriction are clear, with reductions of about 22, 30, and 36% in the timeslices 2011–2040, 2041–2070, and 2071–2100, respectively, in the case of RCP4.5 scenario.
Fig. 6

Suitability of the Arabica coffee for the south of Minas Gerais (SMG), simulated by the eta model at 5 km for the baseline period and future climate, and for RCP4.5 and RCP8.5 scenarios

In the RCP8.5 scenario, reductions in fully adequate areas are even more significant, especially at the end of the twenty-first century. The fully adequate area, which covers 68% in the present climate, reduces to only 4% in 2071–2100. Near the end of the century, areas of regular suitability increase. The restricted areas (25%) appear only in the last timeslice. The fully adequate areas are only present in a region with higher topographies, due to the presence of the Serra da Mantiqueira (figure shown in “Climate projections” section). However, it should be noted that, in practical terms, the cultivation of coffee in these regions would be difficult to access. In the SMG region, the expansion of areas with regular suitability also serves as a warning, considering that the classification is associated with unsuitability due to the increase in air temperature, since water demand increased but did not reach restricted or inadequate values for growing coffee.

The impacts of climate changes on potential yield are inferred through Fig. 7. Under the IPCC AR5 scenarios, Arabica coffee yield declines in the SMG region due to reduction of fully adequate and adequate areas for crop growth. The average yield for the period 2011–2040 does not show significant reductions in the RCP4.5 scenario. However, in the case of the highest emission, RCP8.5 scenario, these projections show a decrease of approximately 5% in the yield, which correspond to a reduction of 89 kg/ha in the SMG region. Stronger reductions are projected in the timeslices 2041–2070 and 2071–2100, with a decrease of about 5 and 7% in the scenario of milder CO2 concentrations. In the scenario of high CO2 concentration, the projections are more worrisome, with reductions of around 9 and 25%, which correspond to a reduction of potential yield of about 169 and 459 kg/ha.
Fig. 7

Potential yield of Arabica coffee (kg/ha) for southern Minas Gerais (SMG). Baseline period and future climate, RCP4.5 and RCP8.5 scenarios

Conclusions and final considerations

Based on the double-nested downscaling of HadGEM2-ES rundown to 5-km resolution and IPCC AR5 scenarios, RCP4.5 and RCP8.5, a large part of southeast Brazil may undergo significant modifications in the areas currently suitable for the growth of Arabica coffee. Currently, most of southeast Brazil is suitable for growing coffee, ranging from fully adequate to at most regular, due to thermal excess or deficit. However, throughout the twenty-first century, considering the two scenarios of concentration of CO2 equivalent, the projections show strong reduction of the fully adequate and adequate areas, increase of the regular areas, and appearance restricted and inadequate areas. These area reductions are caused by the average air temperature increase of about 4 to 8 °C in the RCP8.5 scenario.

According to the projections, in the south of Minas Gerais (SMG), potential yield of Arabica coffee is expected to decrease by about 25% at the end of the twenty-first century. In this region, the areas fully adequate move to regions of higher altitude, which would make farming management difficult especially due to the operation of agricultural machinery in mountain areas. In addition, coffee cultivation in these regions could produce environmental impacts in the remnant Brazilian Atlantic Forest.

In the higher equivalent CO2 concentration scenario, there is an increase of areas of high climatic risk due to the thermal excess. In order to maintain coffee production in Southeast Brazil, aiming at quantity as well as quality, it is necessary to propose adaptation measures that take into account temperature rise of at least 2 to 4 °C, considering the current genetic and physiological characteristics of the Arabica coffee cultivars in this study region.

It is hoped that these assessments of the impacts of climate change on the potential yield of Arabica coffee can help to motivate further research on adaptation measures, such as development of cultivation practices and new cultivars tolerant to higher temperatures. The interactive or complementary use of agricultural practices and the use of cultivars that are more tolerant to heat and drought can contribute to adapting to climate change and, consequently, to sustaining coffee cultivation even in the areas where climate restriction occurs in the future.

Notes

Acknowledgments

The authors thank the São Paulo Research Foundation (FAPESP) for the grant 2014/00192-0, Brazilian National Council for Scientific and Technological Development (CNPq) for the grants 457874/2014-7 and 308035/2013-5, the MCTI/UNDP for the grant BRA/10/G32, and the Brazilian Coordination for the Improvement of Higher Education Personnel (CAPES) the project INCT for Climate Change (MCTI/CNPq).

References

  1. Allen RG, Pereira LS, Raes D, Smith M (1998) Crop evapotranspiration—guidelines for computing crop water requirements. FAO irrigation and drainage paper 56. FAO. http://www.fao.org/docrep/X0490E/X0490E00.htm. Accessed 31 March 2015
  2. Assad ED, Pinto HS, Junior JZ, Ávila AD (2004) Impacto das mudanças climáticas no zoneamento agroclimático do café no Brasil. Pesq Agrop Brasileira 39(11):1057–1064.  https://doi.org/10.1590/S0100-204X2004001100001 CrossRefGoogle Scholar
  3. Baliza DP, Oliveira AL, Dias RAA, Guimarães RJ, Barbosa CR (2013) Antecipação da produção e desenvolvimento da lavoura cafeeira implantada com diferentes tipos de mudas. Coffee Sci 8(1):61–68Google Scholar
  4. Betts AK, Miller MJ (1986) A new convective adjustment scheme. Part II: single column tests using GATE wave, BOMEX, ATEX and arctic air-mass data sets. Quarterly J R Meteorol Soc 112(473):693–709.  https://doi.org/10.1002/qj.49711247308 Google Scholar
  5. Bragança R, dos Santos AR, de Souza EF, de Carvalho AJC, Luppi ASL, da Silva RG (2016) Impactos das mudanças climáticas no zoneamento agroclimatológico do café arábica no Espírito Santo. Revista Agro@mbiente On-line 10(1):77–82.  10.18227/1982-8470ragro.v10i1.2809 CrossRefGoogle Scholar
  6. Bunn C (2015) Modeling the climate change impacts on global coffee production. Doctoral dissertation, Humboldt-Universität zu Berlin, Lebenswissenschaftliche FakultätGoogle Scholar
  7. Bunn C, Läderach P, Rivera OO, Kirschke D (2015) A bitter cup: climate change profile of global production of Arabica and Robusta coffee. Clim Chang 129(1–2):89–101.  https://doi.org/10.1007/s10584-014-1306-x CrossRefGoogle Scholar
  8. Camargo, AP (1977) Zoneamento de aptidão climática para a cafeicultura de arábica e robusta no Brasil. In: Fundação IBGE, recursos, meio ambiente e poluição. p.68–76Google Scholar
  9. Carvalho LG, Rios GFA, Miranda WL, Neto PC (2011) Evapotranspiração de referência: uma abordagem atual de diferentes métodos de estimativa. Pesquisa Agropecuária Tropical 41(3):456–465.  https://doi.org/10.5216/pat.v41i3.12760 Google Scholar
  10. Chou SC, Marengo JA, Lyra AA, Sueiro G, Pesquero JF, Alves LM, Kay G, Betts R, Chagas DJ, Gomes JL, Bustamante JF, Tavares P (2012) Downscaling of South America present climate driven by 4-member HadCM3 runs. Clim Dyn 38(3–4):635–653.  https://doi.org/10.1007/s00382-011-1002-8 CrossRefGoogle Scholar
  11. Chou SC, Lyra A, Mourão C, Dereczynski C, Pilotto I, Gomes J, Bustamante J, Tavares P, Silva A, Rodrigues D, Campos D, Chagas D, Sueiro G, Siqueira G, Marengo J (2014a) Assessment of climate change over South America under RCP 4.5 and 8.5 downscaling scenarios. Am J Clim Chang 3(5):512–525.  https://doi.org/10.4236/ajcc.2014.35043 CrossRefGoogle Scholar
  12. Chou SC, Lyra A, Mourão C, Dereczynski C, Pilotto I, Gomes J, Bustamante J, Tavares P, Silva A, Rodrigues D, Campos D, Chagas D, Sueiro G, Siqueira G, Nobre P, Marengo J (2014b) Evaluation of the eta simulations nested in three global climate models. Am J Clim Chang 3(5):438–454.  https://doi.org/10.4236/ajcc.2014.35039 CrossRefGoogle Scholar
  13. Companhia Nacional de Abastecimento (CONAB) (2015) Acompanhamento da safra brasileira: café. http://www.conab.gov.br. Accessed 22 April 2015
  14. DaMatta FM, Ronchi CP, Maestri M, Barros RS (2007) Ecophysiology of coffee growth and production. Braz J Plant Physiol 19(4):485–510.  https://doi.org/10.1590/S1677-04202007000400014 CrossRefGoogle Scholar
  15. Ek MB, Mitchell KE, Lin Y, Rogers E, Grummann P, Koren V, Gayno G, Tarpley JD (2003) Implementation of Noah land-surface model advances in the NCEP operational mesoscale eta model. J Geophys Res 108(D22):8851–8867.  https://doi.org/10.1029/2002JD003296 CrossRefGoogle Scholar
  16. Eugenio FC, Peluzio TMO, Pereira AAB, dos Santos AR, Peluzio JBE, Bragança R, Fiedler NC, Paula EDSO (2014) Zoning agroclimatological Coffea canephora for Espírito Santo by spatial interpolation. Coffee Sci 9(3):319–328Google Scholar
  17. Fels SB, Schwarzkopf MD (1975) The simplified exchange approximation: a new method for radiative transfer calculations. J Atmos Sci 32(7):1475–1488.  https://doi.org/10.1175/1520-0469(1975)032<1475:TSEAAN>2.0.CO;2 CrossRefGoogle Scholar
  18. Instituto Brasileiro de Geografia e Estatística (IBGE) (2016). PAM: Produção Agrícola Municipal. www.sidra.iibge.gov.br. Accessed 10 May 2016
  19. Intergovernmental Panel on Climate Change (IPCC) (2013) Twelfth session of working group I. Summary for Policymakers. http://www.climatechange2013.org/images/uploads/WGIAR5-SPM_Approved27Sep2013.pdf. Accessed 30 October 2013
  20. International Coffee Organization (ICO) (2015) Catalog coffee. http://www.ico.org/pt/coffee_storyp.asp?section=Sobre_o_café. Accessed 10 May 2015
  21. International Coffee Organization (ICO) (2017) Statistical catalog—historical data. http://www.ico.org/pt/new_historical_p.asp?section=Estat%EDstica. Accessed 05 January 2017
  22. Janjić ZI (1984) Nonlinear advection schemes and energy cascade on semi-staggered grids. Mon Weather Rev 112(6):1234–1245.  https://doi.org/10.1175/1520-0493(1984)112<1234:NASAEC>2.0.CO;2 CrossRefGoogle Scholar
  23. Janjić ZI (1994) The step-mountain eta coordinate model: further developments of the convection, viscous sublayer, and turbulence closure schemes. Mon Weather Rev 122(5):927–945.  https://doi.org/10.1175/1520-0493(1994)122<0927:TSMECM>2.0.CO;2 CrossRefGoogle Scholar
  24. Karlsson KG, Riihelä A, Müller R, Meirink JF, Sedlar J, Stengel M, Lockhoff M, Trentmann J, Kaspar F, Hollmann R, Wolters E (2012) CLARA-A1: CM SAF CLouds, albedo and RAdiation dataset from AVHRR data–edition 1–monthly means/daily means/pentad means/monthly histograms. Satell Appl Facil Climate Monit.  https://doi.org/10.5676/EUM_SAF_CM/CLARA_AVHRR/V001
  25. Kobayashi ES (2007) Consumo de Água e Produtividade de Cafeeiros Arábica na Região de Mococa, SP. Dissertation (MSc in tropical and subtropical agriculture), Instituto Agronômico de CampinasGoogle Scholar
  26. Lacis AA, Hansen JE (1974) A parameterization for the absorption of solar radiation in the earth’s atmosphere. J Atmos Sci 31(1):118–133.  https://doi.org/10.1175/1520-0469(1974)031<0118:APFTAO>2.0.CO;2 CrossRefGoogle Scholar
  27. Luppi ASL, Santos AR, Eugênio FC, Bragança R, Pelúzio JBE, Dalfi RL, Silva RG (2014) Metodologia para Classificação de Zoneamento Agroclimatológico. Revista Brasileira de Climatologia 15.  https://doi.org/10.5380/abclima.v15i0.37388
  28. Lyra AA, Tavares PS, Chou SC, Sueiro G, Dereczynski C, Sonderman M, Silva A, Marengo J, Giarolla A (2017). Climate change projections over three metropolitan regions in Southeast Brazil using the non-hydrostatic eta regional climate model at 5-km resolution. Theor Appl Climatology Accepted doi:  https://doi.org/10.1007/s00704-017-2067-z
  29. Marengo JA, Chou SC, Kay G, Alves LM, Pesquero JF, Soares WR, Santos DC, Lyra AA, Sueiro G, Betts R, Chagas DJ, Gomes JL, Bustamante JF, Tavares P (2012) Development of regional future climate change scenarios in South America using the eta CPTEC/HadCM3 climate change projections: climatology and regional analyses for the Amazon, São Francisco and the Paraná River basins. Clim Dyn 38(9–10):1829–1848.  https://doi.org/10.1007/s00382-011-1155-5 CrossRefGoogle Scholar
  30. Matiello JB (1991) O café: do cultivo ao consumo. Editora Globo. Farmer’s Collection: Grain. Globo Rural Publications, São PauloGoogle Scholar
  31. MCTI (2016) Third National Communication of Brazil to the United Nations framework convention on climate change—volume II/Ministry of Science, Technology and Innovation. Brasília: Ministério da Ciência, Tecnologiae Inovação, 2016. 229 p. ISBN: 978-85-88063-24-2Google Scholar
  32. Meireles EJL, Volpato MML, Alves HMR, Vieira TGC (2007) Zoneamento agroclimático: um estudo de caso para o café. Informe Agropecuário Belo Horizonte 28(241):50–57Google Scholar
  33. Meireles EJL, Camargo M, Pezzopane JRM, Thomaziello R, Fahl JI, Bardin L, Santos JCF, Japiassú LB, Garcia AWR, Miguel AE, Ferreira RA (2009) Fenologia do cafeeiro: condições agrometeorológicas e balanço hídrico do ano agrícola 2004–2005. Embrapa Informação Tecnológica, Embrapa Café, MAPA, Brasília (Document 5). http://www.epamig.ufla.br/geosolos/publicacoes/2007/11.pdf. Accessed 10 Feb 2012
  34. Mellor GL, Yamada T (1982) Development of a turbulence closure model for geophysical fluid problems. Rev Geophys 20(4):851–875.  https://doi.org/10.1029/RG020i004p00851 CrossRefGoogle Scholar
  35. Mendonça PV (1958) Sobre o novo método de balanço hidrológico do solo de Thornthwaite-Mather. In Congresso Luso-espanhol para o progresso das ciências 24 (Proceedings): Madrid. p. 271–282Google Scholar
  36. Mesinger F, Chou SC, Gomes JL, Jovic D, Lyra AA, Bustamante JF, Bastos PR, Lazic L, Morelli S, Ristic I (2012) An upgraded version of the eta model. Meteorog Atmos Phys 116(3–4):63–79.  https://doi.org/10.1007/s00703-012-0182-z CrossRefGoogle Scholar
  37. Ministério da Agricultura, Pecuária e Abastecimento (MAPA) (2013). Crop catalog/coffee. http://www.agricultura.gov.br/vegetal/culturas/cafe/saiba-mais. Accessed 22 April 2013
  38. Moss RH, Edmonds JA, Hibbard KA, Manning MR, Rose SK, Van Vuuren DP, Carter TR, Emori S, Kainuma M, Kram T, Meehl GA, Mitchell JFB, Nakicenovic N, Riahi K, Smith SJ, Stouffer RJ, Thomson AM, Weyant JP, Wilbanks TJ (2010) The next generation of scenarios for climate change research and assessment. Nature 463(7282):747–756.  https://doi.org/10.1038/nature08823 CrossRefGoogle Scholar
  39. Paulson CA (1970) The mathematical representation of wind speed and temperature profiles in the unstable atmospheric surface layer. J Appl Meteorol 9(6):857–861.  https://doi.org/10.1175/1520-0450(1970)009<0857:TMROWS>2.0.CO;2 CrossRefGoogle Scholar
  40. Pereira AR, Camargo AP, Camargo MBP (2008) Agrometeorologia de cafezais no Brasil. Instituto Agronômico de Campinas (IAC). 127 p. ISBN: 978-85-85564-19-3Google Scholar
  41. Pesquero JF, Chou SC, Nobre CA, Marengo JA (2010) Climate downscaling over South America for 1961–1970 using the eta model. Theor Appl Climatol 99(1–2):75–93.  https://doi.org/10.1007/s00704-009-0123-z CrossRefGoogle Scholar
  42. Pinto HS, Assad ED, Zullo Junio RJ, Evangelista SRM, Otavian AF, Ávila AMH, Evangelista BA, Marin F, Macedo Junior C, Pellegrino G, Coltri PP, Coral G (2008) Aquecimento global e a nova geografia da produção agrícola no Brasil. Embrapa, São Paulo https://www.macroprograma1.cnptia.embrapa.br/scaf/publicacoes/CLIMA_E_AGRICULTURA_BRASIL_300908_FINAL.pdf/view. Accessed 09 February 2012Google Scholar
  43. Ranjitkar S, Sujakhu NM, Merz J, Kindt R, Xu J, Matin MA, Ali M, Zomer RJ (2016) Suitability analysis and projected climate change impact on banana and coffee production zones in Nepal. PloS One 11(9):e0163916.  https://doi.org/10.1371/journal.pone.0163916 CrossRefGoogle Scholar
  44. Saha S, Moorthi S, Pan HL, Wu X, Wang J, Nadiga S, Tripp P, Kistler R, Woollen J, Behringer D, Liu H, Stokes D, Grumbine R, Gayno G, Wang J, Chuang YH, Juang HM, Sela J, Iredell M, Treadon R, Kleist D, Delst PV, Keyser D, Derber J, Ek M, Wei JM, Yang R, Lord S, Dool HVD, Kumar A, Wang W, Long C, Chelliah M, Xue Y, Huang B, Schemm JK, Ebisuzaki W, Lin R, Xie P, Chen M, Zhou S, Higgins W, Zou CZ, Liu Q, Chen Y, Han Y, Cucurull L, Reynolds WR, Rutledge G, Goldberg M (2010) The NCEP climate forecast system reanalysis. Bull Am Meteorol Soc 91(8):1015–1057.  https://doi.org/10.1175/2010BAMS3001.1 CrossRefGoogle Scholar
  45. Sediyama GC (1996) Estimativa da evapotranspiração: histórico, evolução e análise crítica. Revista Brasileira de Agrometeorologia 4(1):1–12 ISSN 0104-1347Google Scholar
  46. Souza VCO, Vieira TGC, Volpato MML, Alves HMR (2012) Espacialização e dinâmica da cafeicultura mineira entre 1990 e 2008, utilizando técnicas de geoprocessamento. Coffee Sci 7(2):122–134Google Scholar
  47. Thomaziello RA, Fazuoli LC, Pezzopane, JRM, Fahl JI, Carelli MLC (2000) Café arábica: cultura e técnicas de produção. 1ed. Campinas: Instituto Agronômico. 82p. Boletim Técnico n.187Google Scholar
  48. Thornthwaite CW, Mather JR (1957). Instructions and tables for computing potential evapotranspiration and the water balance. Drexel Institute of Technology, Laboratory of Climatology, Centerton, New Jersey. Publications in climatology, vol. 10, no. 3Google Scholar
  49. Wintgens JN (2004) Coffee: growing, processing, sustainable production. A guidebook for growers, processors, traders, and researchers. 2nd edn. WILEY-VCH. 1040 p. ISBN: 978-3-527-33253-3Google Scholar
  50. Zhao Q, Black TL, Baldwin ME (1997) Implementation of the cloud prediction scheme in the eta model at NCEP. Weather Forecast 12(3):697–712.  https://doi.org/10.1175/1520-0434(1997)012<0697:IOTCPS>2.0.CO;2 CrossRefGoogle Scholar
  51. Zullo J, Pinto HS, Assad ED (2006) Impact assessment study of climate change on agricultural zoning. Meteorol Appl 13(S1):69–80.  https://doi.org/10.1017/S135048270600257X CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany 2017

Authors and Affiliations

  • Priscila da Silva Tavares
    • 1
  • Angélica Giarolla
    • 2
  • Sin Chan Chou
    • 1
  • Adan Juliano de Paula Silva
    • 1
  • André de Arruda Lyra
    • 1
  1. 1.Center for Weather Forecasts and Climate Studies (CPTEC-INPE)Cachoeira PaulistaBrazil
  2. 2.Earth System Science Centre (CCST), National Institute for Space Research (INPE)São José dos CamposBrazil

Personalised recommendations