Simulation of Green and Blue Water Impacts Caused by Climate Changes in the Apucaraninha River Watershed , Southern Brazil

Simulação dos Impactos da Água Verde e Azul Causados pelas Mudanças Climáticas na Bacia Hidrográfica do Rio Apucaraninha, Sudeste do Brasil R E S U M O Mudanças climáticas podem gerar impactos significativos no ciclo hidrológico. É importante reconhecer as modificações na água verde (água armazenada no solo seguida de consumo pela vegetação) e água azul (água que flui em rios, lagos, pântanos e aquíferos rasos) disponibilidade em consequência das modificações resultantes das alterações climáticas. A modelagem matemática é usada para simular o efeito de cenários de mudanças climáticas nos processos hidrológicos nas bacias hidrográficas. Este estudo objetivou avaliar os impactos das mudanças climáticas na água azul e verde na bacia hidrográfica do rio Apucaraninha, Sudeste do Brasil, considerando os cenários climáticos A2 e B2, pessimista e otimista, respectivamente, sobre emissões de gases de efeito estufa desenvolvido pelo IPCC. O SWAT foi calibrado e validado utilizando vazão diária de 1987 a 2012. Os cenários climáticos A2 e B2 foram usados para simular as condições hidrológicas para o período de 20712100. O modelo mostrou ajuste satisfatório em comparação com os dados observados, permitindo a simulação das condições hidrológicas atuais e de impactos futuros das mudanças climáticas nas águas verde e azul. Verificou-se que, apesar do aumento do potencial de evapotranspiração de 19% e 12% para os cenários A2 e B2, respectivamente, causado pelo aumento na temperatura, a redução da quantidade de chuva induziu a uma redução na evapotranspiração real, que correspondem à água verde, e uma redução de 21% para o cenário A2 e 14% para o cenário B2 quanto à disponibilidade de água azul. Palavras-chave: Água verde; Água azul; Modelo SWAT; Mudança climática; Sudeste do Brasil. Introduction The interaction between climatic and hydrologic components involves multiple competing processes (Luo, * E-mail para correspondência: isaiensen@gmail.com (Iensen, I.R.R.). 2013). It is known that climate variability may influence hydrological processes, considering that the main climate variables such as precipitation, radiation, and temperature, have an inseparable role in water balance. Eckhardt and Ulbrich (2003) reported that the main consequences of Revista Brasileira de Geografia Física V. 08, N. 01 (2015) 179-186 Iensen, I.R.R.; Schultz, G.B;Santos I. dos 180 elevation in atmospheric CO2 concentration are the increase in mean temperatures and changes in temporal and spatial distribution of precipitation, which should be accompanied by an increased risk of intense events of rainfall and droughts. Climate changes are expected to have effects on precipitation variability patterns (Dufek and Ambrizzi, 2008), streamflow (Zang and Liu, 2013; Ficklin et al., 2009), evapotranspiration (Rockström et al., 2009; McKenney and Rosenberg, 1993) and hydrosedimentological processes (Nearing et al., 2005; Lu et al., 2013; Iensen et al., 2014). Long-term changes in evapotranspiration may conduct to significant modifications on the hydrologic processes, having critical implications for water availability (Chattopadhyay and Hulme, 1997). Actual evapotranspiration (AET) is a physical process of water transfer to the atmosphere by evaporation from soil and, water bodies and transpiration through vegetation. AET constitutes an important component in the hydrologic balance, which is determined by climatic factors, including temperature, radiation, humidity, and wind speed and by the availability of water in soil (McKenney and Rosenberg, 1993). Rockström and Gordon (2001) reported that most of water resources assessment tends to focus primarily on the evaluation of liquid portion in hydrologic budget, emphasizing the amount of water that flows into rivers, lakes, and groundwater. Therefore, little attention is paid to the role of evapotranspiration or vapor flows of water balance, and its importance in the hydrological cycle, sustaining terrestrial ecosystems. According to Fennessey and Kirshen (1994) in regions with humid climate, it is recognized that actual evapotranspiration may account for over than half of the average of annual water balance, and this amount of water tend not to be considered in water resources assessment. To improve the knowledge about hydroclimatic processes in watersheds Falkenmark (1995) first proposed the green and blue water concept. By definition, blue water consists of the water amount that flows through rivers, aquifers, lakes, and wetlands. Green water comprises the amount of water that comes from precipitation and is stored in soil and then consumed by vegetation. Xu (2013) describe Green water flow as equivalent to the actual evapotranspiration (AET). The concept of green and blue water are important considering the wide view of water balance inputs and outputs provided by this approach. The green and blue water concepts of Falkenmark (1995) were improved by Falkenmark and Rockström (2004), distinguishing flow from storage of green and blue water. Thus, Falkenmark and Rockström (2004) describe blue water storage as the amount of water stored in aquifers. Green water storage is defined as the soil moisture. These authors also differentiate blue water flow that refers to the water flowing in rivers, aquifers and lakes and green water flow refers to the vapor amount that return to the atmosphere as evapotranspiration. Recognizing the importance of green and blue water dynamics and to elucidate the effects of climate change on environmental processes in watershed, many studies that have been conducted concluded that any changes in climatic variables might affect significantly green and blue water availability. (Xu, 2013; Rockströmand Gordon, 2001; Zang and Liu, 2013; Faramarzi et al., 2009; Falkenmark and Rockström, 2004; Gerten et al., 2005; Hoff et al., 2010; Liu and Yang 2010). McKenney and Rosenberg (1993) reported that due to modifications in precipitation patterns it is expected that water storage in soil also decreases, and hence the reduction in actual evapotranspiration occurs, these modifications might influence the amount of water available to grow crops, to water animals, and to supply human needs. Also, Pruski and Nearing (2002) describes that changes in temperature could change evapotranspiration rates, affecting soil moisture and, therefore, infiltration and runoff. The prediction of the effect of climate change on blue and green water it is important to make a quantitative assessment thought hydrological modeling is emphasizing the importance of this assessment for future management of the watershed. The magnitude of climate change effect on environmental dynamics varies, depending on the region analyzed and considered climate scenario (IPCC, 2007). Due to the heterogeneity of climate conditions predicted by the scenarios, it is necessary to regionalize assessment of climate change impacts. Thus, this paper proposes the use of Soil and Water Assessment Tool (SWAT), to estimate potential regional impacts of climate change scenarios on blue and green water in the Apucaraninha River watershed located in southern Brazil. Material and Methods


Introduction
The interaction between climatic and hydrologic components involves multiple competing processes (Luo, * E-mail para correspondência: isaiensen@gmail.com(Iensen, I.R.R.).2013).It is known that climate variability may influence hydrological processes, considering that the main climate variables such as precipitation, radiation, and temperature, have an inseparable role in water balance.Eckhardt and Ulbrich (2003) reported that the main consequences of elevation in atmospheric CO2 concentration are the increase in mean temperatures and changes in temporal and spatial distribution of precipitation, which should be accompanied by an increased risk of intense events of rainfall and droughts.Climate changes are expected to have effects on precipitation variability patterns (Dufek and Ambrizzi, 2008), streamflow (Zang and Liu, 2013;Ficklin et al., 2009), evapotranspiration (Rockström et al., 2009;McKenney and Rosenberg, 1993) and hydrosedimentological processes (Nearing et al., 2005;Lu et al., 2013;Iensen et al., 2014).Long-term changes in evapotranspiration may conduct to significant modifications on the hydrologic processes, having critical implications for water availability (Chattopadhyay and Hulme, 1997).
Actual evapotranspiration (AET) is a physical process of water transfer to the atmosphere by evaporation from soil and, water bodies and transpiration through vegetation.AET constitutes an important component in the hydrologic balance, which is determined by climatic factors, including temperature, radiation, humidity, and wind speed and by the availability of water in soil (McKenney and Rosenberg, 1993).Rockström and Gordon (2001) reported that most of water resources assessment tends to focus primarily on the evaluation of liquid portion in hydrologic budget, emphasizing the amount of water that flows into rivers, lakes, and groundwater.Therefore, little attention is paid to the role of evapotranspiration or vapor flows of water balance, and its importance in the hydrological cycle, sustaining terrestrial ecosystems.According to Fennessey and Kirshen (1994) in regions with humid climate, it is recognized that actual evapotranspiration may account for over than half of the average of annual water balance, and this amount of water tend not to be considered in water resources assessment.
To improve the knowledge about hydroclimatic processes in watersheds Falkenmark (1995) first proposed the green and blue water concept.By definition, blue water consists of the water amount that flows through rivers, aquifers, lakes, and wetlands.Green water comprises the amount of water that comes from precipitation and is stored in soil and then consumed by vegetation.Xu (2013) describe Green water flow as equivalent to the actual evapotranspiration (AET).The concept of green and blue water are important considering the wide view of water balance inputs and outputs provided by this approach.
The green and blue water concepts of Falkenmark (1995) were improved by Falkenmark and Rockström (2004), distinguishing flow from storage of green and blue water.Thus, Falkenmark and Rockström (2004) describe blue water storage as the amount of water stored in aquifers.Green water storage is defined as the soil moisture.These authors also differentiate blue water flow that refers to the water flowing in rivers, aquifers and lakes and green water flow refers to the vapor amount that return to the atmosphere as evapotranspiration.
Recognizing the importance of green and blue water dynamics and to elucidate the effects of climate change on environmental processes in watershed, many studies that have been conducted concluded that any changes in climatic variables might affect significantly green and blue water availability.(Xu, 2013;Rockströmand Gordon, 2001;Zang and Liu, 2013;Faramarzi et al., 2009;Falkenmark and Rockström, 2004;Gerten et al., 2005;Hoff et al., 2010;Liu and Yang 2010).McKenney and Rosenberg (1993) reported that due to modifications in precipitation patterns it is expected that water storage in soil also decreases, and hence the reduction in actual evapotranspiration occurs, these modifications might influence the amount of water available to grow crops, to water animals, and to supply human needs.Also, Pruski and Nearing (2002) describes that changes in temperature could change evapotranspiration rates, affecting soil moisture and, therefore, infiltration and runoff.
The prediction of the effect of climate change on blue and green water it is important to make a quantitative assessment thought hydrological modeling is emphasizing the importance of this assessment for future management of the watershed.The magnitude of climate change effect on environmental dynamics varies, depending on the region analyzed and considered climate scenario (IPCC, 2007).Due to the heterogeneity of climate conditions predicted by the scenarios, it is necessary to regionalize assessment of climate change impacts.Thus, this paper proposes the use of Soil and Water Assessment Tool (SWAT), to estimate potential regional impacts of climate change scenarios on blue and green water in the Apucaraninha River watershed located in southern Brazil.

Study area
The Apucaraninha River watershed drains an area of 504 km² and is located in southern Brazil (50º56'W and 23º42'S) (Figure 1).The local geomorphology consists of smooth hills, with elevation ranging from 660 m to 1210 m.The average slope of the Apucaraninha River watershed is approximately 11%.
Land use in the watershed is predominantly agriculture (63%), with soybean and wheat as major crops.The climate is classified as Humid Subtropical Climate or Cfa (Köppen, 1948), characterized by high temperature and high rainfall in summer months.The average annual precipitation in the Apucaraninha River watershed is 1,634 mm.

SWAT model application
The SWAT model is a hydrological model developed in 1996 by U.S. Agricultural Research Service, Texas A & M University and other federal agencies (Neitsch et al., 2005).It is a watershed-scale model that simulates monthly and daily streamflow, nutrient loading, and sediment yield resulting from the interaction of weather, soil properties, stream channel characteristics, agricultural management and crop growth (Nearing et al. 2005).SWAT requires the input of climate data such as precipitation, temperature and humidity and therefore is used to investigate climate change effects on plant growth, evapotranspiration, snow and runoff generation (Eckhardt and Ulbrich, 2003;Zang and Liu, 2013;Luo et al., 2013).SWAT model has been successfully used to predict hydrological cycle and scenario simulation in many watersheds of different sizes and environmental conditions (Schuol et al., 2008;Gerten et al., 2005;Faramazi et al., 2009;Luo et al., 2013).
SWAT estimates runoff using the Curve Number method by SCS (Soil Conservation Service, 1972).To predict potential evapotranspiration SWAT offers three equations: Priestley-Taylor (1972), Hargreaves (Hargreaves and Samani, 1985), and Penman-Monteith (Allen et al., 1989).In this study, the Penman-Monteith method was used.This method requires data on vegetation and climatic data of radiation, humidity and wind speed.The land use data was obtained by supervised classification of digital images of LANDSAT 7 ETM +1, bands 5, 4 and 3, which spatial resolution is 30m.Definition of soil types and their physical and hydrological characteristics were obtained in EMBRAPA (1984) mapping in scale 1:600,000.The DEM had a spatial resolution of 30 m and was made contour and point elevation data of topographic maps in scale 1:50,000.
The watershed was partitioned in 43 sub-watersheds of equivalent size.Each sub-watershed was divided into Hydrological Response Units (HRU), which are combinations of homogeneous soil types, land use, slope and management (Neitsch et al., 2005).The 43 subwatersheds were divided in 350 HRU.
Streamflow from the Apucaraninha River watershed was calibrated and validated with daily measured data from 1987 to 2012.Calibration was made with observed data from 2000 to 2012 and validation from 1987 to 2000.The model was calibrated in two steps.First, a manual calibration was done to adjust the main components of water balance; then an automatic calibration was done using SWAT-CUP (Abbaspour,2011)to achieve a good model fit.

Climate change scenarios
After model calibration, climate scenarios A2 and B2 were used as input data in SWAT modeling.The scenarios were generated with spatial resolution of 50 km (0.5º latitude x 0.5º longitude) and with daily time step, using the regional climate model HadRM3P of the Hadley Centre, UK, and downscaled using Integrated System of Regional Modelling PRECIS -Providing Regional Climate for Impact Studies (Marengo et al., 2009).
Climate scenarios were based on possible trends of CO2, population growth, socio-economic development and technological changes (Marengo, 2007).The scenarios developed by IPCC compartmentalize the world in several large cells.The downscaling technique is used to regionalize data.Through downscaling, the climate and weather information are regionalized to present details of the particularities of each region.
The A2 scenario describes a pessimist scenario for climate change that the CO2 concentration approximately will fold until 2100.Also, it is predicted an increase of 2ºC to 5.4ºC in temperature by 2100 (IPCC, 2007).The B2 scenario describes an optimistic scenario for climate change in which the increase in temperature varies between 1.8ºC to 3.8ºC and CO2 concentration increase 50% by 2100.Also, A2 and B2 emissions scenarios indicate that the general tendency of global average water vapor, evaporation and precipitation are projected to increase, although at the regional scale both increases or decreases are seen (IPCC, 2007).

Blue and green water
To estimate the percentage of precipitation that is transformed in actual evapotranspiration we used the green water coefficient (Equation 1) described by Xu (2013), which herein is the equivalent of the evapotranspiration coefficient.The use of Cgw emphasizes the rate of transformation from precipitation to green water or actual evapotranspiration.
where Cgw is the coefficient of green water.Qgreen is the green water amount.P is precipitation.Qblue is the amount of blue water.
In this study, green water flow is considered the amount of actual evapotranspiration (AET).Green water storage is defined analyzing the variation of water stored in the soil.Still, blue water flow is considered the sum of surface runoff, lateral flow contribution to stream flow and groundwater contribution to streamflow.Due to limitations in SWAT output, we could not define blue water storage.Thus, this concept will not be evaluated in this paper.

Calibration and validation
According to literature, some of SWAT parameters were altered to permit an adequate reproduction the actual hydrological conditions in the Apucaraninha River watershed.The parameters and values used in the calibration process are presented inTable 1.

Climate change impacts
On both scenarios, the main variation was observed in the temperature, which increases in both scenarios and in the precipitation that decreased 365.9 mm and 451.9 mm for the scenarios A2 and B2, respectively, as could be observed in the Table 3.
In the Apucaraninha River watershed it was found that 67% of precipitation is transformed into the green water flow, or actual evapotranspiration (AET).This result indicates that more than half of water balance is not explicitly considered in water resources management in the studied area.Temperature and precipitation are not the only variables that may affect the hydrological cycle although it is considered that these are the main variables influencing water availability and evapotranspiration.Martin et al. (1989) found that the effect of temperature is unlikely to be the only climatic element affected by greenhouse warming, however, changes in other climatic elements can offset or intensify the effects of rising temperature on evapotranspiration.
The main results obtained in the simulation of climate change scenarios in the Apucaraninha River watershed are presented in Figure 2. In consequence of the increase in temperature, potential evapotranspiration (PET) increased 19% under scenario A2 and 12% under scenario B2 (Figure 2d), the most expressive variation occurred between July and October.According to Katul et al. (2012), PET is expected to increase by 6.8% for each 1°C increase in air temperature.
Although the increase in PET, it was observed a reduction in precipitation of 19% and 23% (Figure 2a) for scenario A2 and B2, respectively.With the significant reduction in precipitation amount, soil moisture became limited presenting a reduction of 13% in the A2 scenario and 7% in the B2 scenario.In this context, with the reduction in water availability, there was less water to streamflow and to be stored in soil (Figure 2c).This modification in soil moisture is alarming considering that the agricultures the predominant land use of Apucaraninha river watershed.
The decreasing trend in precipitation and soil water content, induce, reduction of 16% under A2 scenario and 33% under a B2 scenario in AET or green water flow (Falkenmark and Rockstrom, 2006).We found that blue water flow also showed a general decrease of 21% to the A2 scenario and 14% to the B2 scenario.According to McKenney and Rosenberg (1993), in many water balance models, the actual evapotranspiration (AET) is assumed to follow a linear function of soil moisture content.When soil moisture becomes limited, AET would be less sensitive to changes in temperature, solar radiation, humidity, and the wind than would PET.By implication, climate effects in the runoff would follow the same patterns of changes in PET if soil moisture content remains high and does not limit AET.In Apucaraninha River watershed, it was observed that despite the increase in PET, actual evapotranspiration did not follow the same upward tendency because of a decrease in precipitation amount.With less water stored in the soil, AET also becomes limited.
These results could be compared to those obtained by Eckhardt and Ulbrich (2003), these authors found that the increase in potential evapotranspiration associated with decreased precipitation may reduce groundwater recharge and streamflow by more than 50%.Luo et al. (2013) also reported that higher temperature is expected to increase potential evapotranspiration and decrease water yield in watersheds.Luo et al. (2013) also described that these changes in average annual streamflow predicted by SWAT were significantly correlated with precipitation changes, in addition, the seasonality of precipitation change was significantly magnified in the resultant seasonality of streamflow.
In order to evaluate the monthly time series generated by the simulations of baseline conditions A2 and B2 scenarios, it was elaborated a box plot graph (Figure 3) to analyze general variation of green and blue water flow.
Green water flow (Figure 3a) showed a reduction in the median, minimum and maximum values for both climate scenarios.This situation was caused by the reduction of precipitation, which implies that less water is available to be stored in soil and consumed by vegetation as actual evapotranspiration.
Blue water flow (Figure 3b) presented an increase in maximum and absolute minimum values for both climate scenarios indicating the occurrence of more extreme and intense events.Scenario A2 presents the largest amplitude.Scenario B2 showed a higher frequency of blue water below the median value indicating that in the climate condition predicted by this scenario is expected the increase in months characterized by low flows.

Conclusion
SWAT model was able to reproduce the current hydrological conditions of the watershed enabling the simulation of climate scenarios to assess the impacts on the hydrological dynamics.
The importance ofAET as the main output term in the water balance of most watersheds, mainly in a humid climate.As such, estimation of actual evapotranspiration conditions is a crucial component for evaluating watershed response to future climates.
By the simulation of climate change impacts, it was observed that the increase in temperature induced to an increase of 19% and 12% for A2 and B2 scenario, respectively.Although the increase in potential evapotranspiration, the decrease trend in precipitation (19% for A2 and 23% for B2) culminate to a decrease of 13% in A2 and 7% in B2 in soil moisture, which correspond to green water storage.As a consequence of less water stored in the soil, AET, which represents green water flow, reduced 16% in the A2 scenario and 33% in the B2 scenario.Blue water flow, which refers to water that flow meanwhile river and aquifers showed a reduction of 21% for A2 scenario and 14% for the B2 scenario.Also, the results showed the seasonality of climate data and its consequences in blue water flow and green water flow and green water storage in the Apucaraninha River watershed.
Analyzing green and blue water flow series it was observed the general trend of reduction in median values in green water and an increase in the amplitude in blue water for both climate change scenarios.

Figure 1 .
Figure 1.Geographical localization of the Apucaraninha River watershed in Southern Brazil.
SWAT requires spatially distributed data and time series of climatic and hydrological data.Data used in this study were obtained from the climate and rain gauge station located near the studied area.Rain gauge and Stream gauge stations are shown in Figure 1.Time series of temperature, radiation, air relative humidity and wind speed were obtained in the Instituto Agronômico do Paraná (IAPAR).The weather station is located 40 km of the Apucaraninha River watershed.

Figure 2 .
Figure 2. Simulation results for the current climate condition and scenarios A2 and B2.a) Mean Precipitation (mm).b)Mean blue water flow (mm).c) Mean green water storage (mm).d) Mean green water flow (mm).

Figure 3 .
Figure 3. Box plot for baseline climate condition and scenarios A2 and B2.a) Green water flow; b) Blue water flow.

Table 1 .
Parameters modified for simulation of runoff, their description and the adopted value.

Table 2
shows the statistical test values obtained in the process of calibration and validation for streamflow.As verified by statistical methods, the simulated streamflow has a valid fit to the observed data in calibration and validation periods.Therefore, the model with the calibrated parameters can represent the hydrological processes of Apucaraninha watershed and can be used to evaluate the current and future climate conditions.

Table 2 .
Evaluation of calibration and validation process.

Table 3 .
Comparison of values of climate variables for baseline conditions and the climate scenarios used.TMAX = Average maximum temperature; TMIN = average minimum temperature; RAD = Average radiation; HMD = Average of relative humidity of the air; WND = Average of wind speed.