Int J Biometeorol DOI 10.1007/s00484-014-0798-4
ORIGINAL PAPER
The variation of the water deficit during the winter wheat growing season and its impact on crop yield in the North China Plain Jianjun Wu & Ming Liu & Aifeng Lü & Bin He
Received: 15 January 2013 / Revised: 28 December 2013 / Accepted: 27 January 2014 # ISB 2014
Abstract The North China Plain (NCP) is one of the main agricultural areas in China. However, it is also widely known for its water shortages, especially during the winter wheat growing season. Recently, climate change has significantly affected the water environment for crop growth. Analyzing the changes in the water deficit, which is only affected by climate factor, will help to improve water management in the NCP. In this study, the Decision Support System for Agrotechnology Transfer (DSSAT) was used to investigate the variations in the water deficit during the winter wheat growing season from 1961 to 2010 in 12 selected stations in the NCP. To represent the changes in the water deficit without any artificial affection, the rainfed simulation was used. Over the past 50 years, the average temperature during the winter wheat growing season increased approximately 1.42 °C. The anthesis date moved forward approximately 7–10 days and to late April, which increased the water demand in April. Precipitation in March and May showed a positive trend, but there was a negative trend in April. The water deficit in late April and early May became more serious than before, with an increasing trend of more than 0.1 mm/year. In addition,
J. Wu State Key Laboratory of Earth Surface Processes and Resource Ecology, Beijing Normal University, Beijing 100875, China J. Wu : M. Liu (*) Academy of Disaster Reduction and Emergency Management, Beijing Normal University, Beijing 100875, China e-mail: [email protected] A. Lü Institute of Geographic Sciences and Natural Resources Research, China Academy of Sciences, Beijing 100101, China B. He College of Global Change and Earth System Science, Beijing Normal University, Beijing 100875, China
because the heading stage, which is very important to crop yield of winter wheat, moved forward, the impact of water deficit in late April was more serious to crop yield. Keywords Climate variability . Crop growth simulation . Crop water production function . DSSAT
Introduction The North China Plain (NCP) is one of the most important agricultural regions in China (Wu et al. 2006). Crop rotation is dominated by a double-cropping system of winter wheat and summer maize. Approximately 45 % of wheat planting area of China and 33 % of maize are located in the NCP (Guo et al. 2010). The NCP has a temperate continental monsoon climate, and more than 70 % of precipitation distributes from July to September (Fu et al. 2009). The lack of precipitation during the winter wheat growing season (October to June) causes a serious water deficit for crop production (Chen et al. 2010). In the past few decades, climate variability have affected the hydrological cycle in many places around the world including the NCP, which also much influence the occurrence and severity of the water deficit (Gao et al. 2006; Huntington 2006; Ma and Fu 2006; Sheffield and Wood 2008). To improve the agricultural and water management, it is necessary to analyze how the water deficit during the winter wheat growing season is affected by climate variability and the impact of the water-deficit variability on crop yield. Many drought indices, such as the Standard Precipitation Index (SPI) (McKee et al. 1993) and the Palmer Drought Severity Index (PDSI) (Palmer 1965), have been used to investigate the historical changes in drought occurrences. The SPI can be a useful tool for analyzing the soil water deficit for rainfed crops by calculating the anomalies in precipitation (Bordi et al. 2009). The PDSI involves evapotranspiration and
Int J Biometeorol
calculates the soil water deficit based on a genetic two-layer soil model (Heim 2002; Dai 2011). The frequency of dry years, which was calculated by SPI and PDSI showed statistic significant increasing trend in northern China from the 1960s (Zhai et al. 2010). Many process-based models such as the Variable Infiltration Capacity (VIC) land surface model and the soil water balance model were also used to simulate the change in soil moisture (Andreadis et al. 2005). For example, the area that suffered from a water deficit showed an increasing trend in East Asia from 1950 to 2000 based upon simulations by the VIC (Sheffield and Wood 2008). A significantly decreasing trend in soil moisture was also found in northern China based on a soil water balance model (Tao et al. 2003). However, crop water demand varies over the growth stages (Igbadun et al. 2007). Rising temperature cannot only influence the evapotranspiration rate (Meehl 2007; Easterling et al. 2007) but also affect the crop development rate, which would change the temporal dynamics of the crop water requirement (Badeck et al. 2004). However, the meteorological drought indices and the soil water balance model do not consider the impact of changing crop development rate on crop water demand. In addition, a crop yield reduction is one of the main impacts of the water deficit on agriculture (Mkhabela et al. 2010) and can indicate the severity of it. The impact of a water deficit on crop yield also varies by the growth stages (Shahid 2011). Nageswara Rao et al. (1985) found that the water stress during the seed-filling phase had a more serious impact on the peanut yield. For rice, however, Boonjung and Fukai (1996) found that the impact of water deficit during the panicle development was the most serious on yield. Zhang et al. (1999) found that the water condition during the stages of heading and flowering was more important to winter wheat. However, those meteorological drought indices cannot reflect the dynamics of the impact of the water deficits during different growth stages on crop yield and indicate the severity of the water deficit. Crop models, based on detailed descriptions of the basic physiological and ecological processes of crops (Boote et al. 2011; Jones et al. 2003; Rosenzweig and Iglesias 1998; Williams et al. 1989), can be a good tool to analyze the variation of the water deficit and the impact of the waterdeficit variability on crop yield. Crop model was always developed in one or few stations. When it was used in other regions, it may have some uncertain results because of the differences in soil, weather, and crop type (Xiong et al 2008). Model evaluation was usually the first step to crop model application. Crop models have been demonstrated to be a useful tool for investigating the changes in soil water content (Jamieson et al. 1998; Singh et al. 2008) and have a good ability to simulate crop yield under various water conditions (Liu et al. 2011). They were also used to analyze the impact of climate change and climate variability on agricultural production (Alexandrov and Hoogenboom 2001; Xiong et al. 2010;
Chavas et al. 2009; Thomson et al. 2006). In the NCP, Chen et al. (2010) simulated the impact of climate variability on the water balance through a farming system model, Agricultural Production Systems Simulator (APSIM). Mo et al. (2009) analyzed the responses of the crop yield, water consumption, and water use efficiency to climate change in winter wheat and summer maize, based on crop growth simulations under IPCC SRES A2 and B1 future climate scenarios by the Vegetation Interface Processes (VIP) in the NCP. Crop models can also simulate the impact of changing temperature on crop phenology (Guo et al. 2010; Xiao et al 2012), which can reflect their impact on the temporal dynamics of the crop water demand. By rainfed simulation, it can show the impact of climate variability on the soil water content (Eitzinger et al. 2003). Therefore, well-calibrated and validated crop models are basically very suitable for this study to analyze the impact of climate variability on the water deficit and investigate these changes’ impact on crop yield. Over the past few decades, climate change had affected the crop development rate and the temporal dynamics of crop water requirement and supplement, which also changed the temporal distribution of the water deficit. However, those climate change effects cannot be investigated by the meteorological drought indices. Thus, the objective of this study was to investigate the changes in water deficits during the growing season of winter wheat due to climate change, and their impacts on crop yield in the NCP were also analyzed based on the crop growth simulation.
Materials and methods In this study, the Crop-Environment Resource Synthesis (CERES) model in the Decision Support System for Agrotechnology Transfer (DSSAT) was chosen to simulate the crop growth of winter wheat in 12 selected stations under two treatments of automatic irrigation and rainfed during the period from 1961 to 2010. The variation in the water deficit without artificial affection, which is only affected by climate variability, was analyzed. Study region and data sources The NCP is located in northern China (31–42° E, 114–121° N) and forms part of the alluvial plain developed by the Yellow River, the Huaihe River, and the Haihe River (Liu et al. 2010). The soil types of Mollic Gleysols,Chromic Cambisols, and Eutric Cambisols are widely distributed throughout the NCP (FAO 2003). In this study, 12 stations, including Beijing, Tangshan, Huimin, Dezhou, Laiyang, Zibo, Anyang, Zhengzhou, Shangqiu, Xuchang, Suzhou, and Huaian were selected as data base (Table 1). The 12 selected stations can represent the spatial patterns of precipitation and temperature
Int J Biometeorol Table 1 Basic information about the 12 selected stations in the NCP Name
Abb.
Long. (° E)
Lat. (° N)
Subregion
Soil type
P (mm)
PET (mm)
Anyang Beijing Tangshan Dezhou Huimin Zibo Laiyang Zhengzhou Xuchang Shangqiu Suzhou Huaian
AY BJ TS DZ HM ZB LY ZZ XC SQ SZ HA
114.35 116.63 118.15 116.32 117.53 118.00 120.70 113.65 113.85 115.66 116.98 119.03
35.93 39.92 39.67 37.43 37.50 36.83 36.93 34.72 34.01 34.45 33.63 33.60
Central Northern Northern Central Central Eastern Eastern Southwestern Southwestern Southern Southeastern Southestern
Chromic Cambisols Chromic Cambisols Chromic Cambisols Mollic Gleysols Mollic Gleysols Eutric Cambisols Eutric Cambisols Chromic Cambisols Mollic Gleysols Mollic Gleysols Mollic Gleysols Mollic Gleysols
158 173 125 155 141 184 227 199 247 231 287 318
479 459 515 522 520 459 473 541 525 537 521 507
Pprecipitation during the winter wheat growing season, PET potential evapotranspiration during the winter wheat growing season based on the method of Penman
in the NCP. The average precipitation during the winter wheat growing season decreases gradually from approximately 300 mm in the southern region to less than 150 mm in the northern and central regions. The water deficit (potential evapotranspiration subtract precipitation) increases from 200 mm in the southern region to about 400 mm in the central region. The 12 stations that were evenly distributed throughout the NCP contain each gradient of water deficit in the NCP. Daily meteorological data from 1961 to 2010 of the selected stations were collected from the China Meteorological Administration, including the minimum and maximum air temperature (°C), precipitation (mm), the relative humidity (%), the sunshine duration (h), and the daily average wind speed (m/s). The sunshine duration was converted to solar radiation based on the Ångström function (Chen et al. 2010). The soil data, which included the number of soil layers and some attributes for each layer, including depth, percent sand, percent silt, percent clay, PH, percent organic carbon, and percent calcium carbonate were collected from the Gridded FAO/UNESCO Soil Units (FAO/UNESCO 1992). Model description The DSSAT is organized by modules of weather, soil, plant, the soil–plant–atmosphere interface, and crop management (Jones et al. 2003). Models for more than 28 crops are integrated into the plant module, which can simulate the physiological changes in crop growth and development (Ritchie et al. 1998; Hoogenboom et al. 2011). The DSSAT can simulate the changes in soil water and nitrogen content and represent the impact of these changes on crop growth accurately, which has been widely used in agricultural management, environmental pollution modeling, and yield forecasting (Alexandrov and
Hoogenboom 2001; Gerakis and Ritchie 1998; Xiong et al. 2008; Zhao et al. 2011). CERES models in DSSAT-Cropping System Models (CSM) have been recognized as important tools for assessing the impact of climate change and climate variability on crop growth and development (Tubiello and Ewert 2002). Based on the observation of crop growth of winter wheat under five irrigation treatments in Gucheng in the NCP in 2007–2009, we calibrated the cultivar coefficients of CERES-Wheat in the DSSAT (version 4.5) (Table 2). The relative root mean square error (RMSE) of the observed and simulated crop yield was 0.39 t/ha. The correlation of the observed and simulated biomass during the crop growing season was 4.82 %. We also compared the observation and simulated crop yield in the 12 selected stations from 2005 to 2009. The normalized RMSE was 16.5 %, which indicated CERES-Wheat had a good performance in regional simulation (Liu 2013). In this study, the crop growth of winter wheat in the 12 selected stations over the past 50 years, following treatments of automatic irrigation and rainfed, was simulated by the DSSAT. Irrigation under the automatic irrigation treatment was applied when the soil water content of the top 30 cm was less than 50 % of its capacity, which ensure that no drought stress occurs. To represent the changes in the water deficit without any artificial affection, which is only affected by climate factor, rainfed treatment was used and the difference in evapotranspiration between the two treatments was used to reflect the water deficit. Trends in temperature and precipitation Trends in temperature and precipitation in each 10-day period during the winter wheat growing season were also analyzed to
Int J Biometeorol Table 2 Value of the key cultivar coefficients after calibration
Cultivar coefficient
Introduction
Range
Value
P1V P1D P5 G1 G2
Days required for vernalization (optimum temperature) Photoperiod response (% reduction in rate/10 h drop in pp) Grain filling (excluding lag) phase duration (°Cd) Kernel number per unit canopy weight at anthesis (no./g) Standard kernel size under optimum conditions(mg)
5–60 25–120 350–600 17–30 20–55
40 55 400 20 50
G3 PHINT
Standard, nonstressed mature tiller weight (g dwt) Interval between successive leaf tip appearances(°Cd)
1.0–1.5 60–90
show climate variability during the growing season. These trends were obtained using the method developed by Hirsch et al. (1982). The trend magnitude was defined as (Eq. 1): X jk −X ik β ¼ median j−i
ðk ¼ 1; 2; ⋯; n:1961 < i < j < 2010Þ
ð1Þ where β was the trend magnitude; i and j were the years; and Xjk was the average temperature or precipitation in the kth 10-day period of the jth year. Assessing the impact of the water deficit on crop yield A reduction in crop yield is a main impact of agricultural drought. This study used the Yield Reduction by Water Deficit (YRWD) indicator to assess the impact of the water deficit. The impact of the water deficit on crop growth varies over phenological stages. The crop water production function can represent the different impacts of the water deficit during different growth stages on crop yield (Igbadun et al. 2007) (Eq. 2): X ETai Ya ¼ Ai Ym ETmi i¼1 n
ð2Þ
where Ya was the crop yield for the water-stressed treatment; Ym was the crop yield for the treatment without water stress; ETai was the total evapotranspiration amount at the growth stage i in the treatment with water stress; ETmi was the total evapotranspiration amount at the growth stage i in the treatment without water stress; and Ai was the crop sensitivity index to drought stress during the growth stage i, which can represent the different impacts of water deficit during different crop growth stages on the final yield. The crop water production function (Eq. 2) is a multiply linear function. To obtain the Ai in the crop water production function at these 12 selected stations, the crop growth was simulated in 7 deficient irrigation treatments and 1 automatic irrigation treatment in the past 30 years by the DSSAT. The irrigation stages for the seven deficient irrigation treatments
1.5 85
are shown in Table 3. Irrigations were only applied on the first day of each selected growth stage and ensured the soil water content was at the field capacity. Crop yield and evapotranspiration in the treatment of automatic irrigation were considered to be Ym and ETm, and those in the treatment of deficient irrigation were considered to be Ya and ETa in Eq. 2. The multiple linear regression was used to estimate Ai. The stages in the middle of the winter wheat growing season, such as jointing, heading, and flowering in March and April, have a larger crop sensitivity index to drought stress, and the crop sensitivity index to drought stress in winter is lower during the growing season of winter wheat (Zhang et al. 1999). The crop growth simulations of winter wheat under the treatments of rainfed and automatic irrigation were used to calculate the daily YRWD for the rainfed treatment. The yield ratio (Ya/Ym) on the ith day reflected the impact on final yield of a water deficit from planting to the ith day, which was calculated by the crop water production function (Eq. 2), based on the assumption of sufficient water supplementation after the ith day until maturity. When calculating the yield ratio on the ith day, the ETa from planting to the ith day in the crop water production function was obtained from the evapotranspiration under the rainfed treatment, and the ETa from i+ 1th day to maturity was obtained from the evapotranspiration under the automatic Table 3 Description of deficient irrigation treatments for calculating the crop sensitivity index to drought stress Irrigation treatment
Time of irrigation applied
I1 I2 I3 I4 I5 I6
Heading Stem elongation and heading Stem elongation, heading, and grain filling Reviving, stem elongation, heading, flowering Before freezing, reviving, flowering, grain filling Before freezing, reviving, stem elongation, flowering, grain filling Before freezing, reviving, stem elongation, heading, flowering, and grain filling
I7
Int J Biometeorol 400
10.0 Annual Precipitation Average Precipitation Temperature
9.5 9.0 8.5
200
8.0 7.5
o
Precipitation (mm)
300
Temperature ( C)
Fig. 1 Time series of precipitation and average temperature during the period from late October to May in the selected 12 stations from 1961 to 2010 (The average precipitation in the past 50 years and the linear regression of temperature are also provided.)
7.0
100
6.5 0 1960
1970
1980
1990
6.0 2010
2000
Year
irrigation treatment. The ETm during the whole growing season was obtained from the evapotranspiration under the automatic irrigation treatment. The value of the YRWD on the ith day was calculated as the yield ratio on i−1th day minus the yield ratio on ith day. The variation of the average YRWD in each 10-day period during the growing season of winter wheat was used to represent the variation of the impact of the changing water deficit on crop yield. The results of the trends in the YRWD still provided information regarding the application of irrigation and improve irrigation efficiency. The spring season (from March to May) was the main growing season of winter wheat in the NCP. Therefore, this study chose the period from March to May to analyze the changes in the water deficit and their impacts on crop yield.
Results and discussions Climatic trends The average temperature from late October to May has a significantly increasing trend (p
ORIGINAL PAPER
The variation of the water deficit during the winter wheat growing season and its impact on crop yield in the North China Plain Jianjun Wu & Ming Liu & Aifeng Lü & Bin He
Received: 15 January 2013 / Revised: 28 December 2013 / Accepted: 27 January 2014 # ISB 2014
Abstract The North China Plain (NCP) is one of the main agricultural areas in China. However, it is also widely known for its water shortages, especially during the winter wheat growing season. Recently, climate change has significantly affected the water environment for crop growth. Analyzing the changes in the water deficit, which is only affected by climate factor, will help to improve water management in the NCP. In this study, the Decision Support System for Agrotechnology Transfer (DSSAT) was used to investigate the variations in the water deficit during the winter wheat growing season from 1961 to 2010 in 12 selected stations in the NCP. To represent the changes in the water deficit without any artificial affection, the rainfed simulation was used. Over the past 50 years, the average temperature during the winter wheat growing season increased approximately 1.42 °C. The anthesis date moved forward approximately 7–10 days and to late April, which increased the water demand in April. Precipitation in March and May showed a positive trend, but there was a negative trend in April. The water deficit in late April and early May became more serious than before, with an increasing trend of more than 0.1 mm/year. In addition,
J. Wu State Key Laboratory of Earth Surface Processes and Resource Ecology, Beijing Normal University, Beijing 100875, China J. Wu : M. Liu (*) Academy of Disaster Reduction and Emergency Management, Beijing Normal University, Beijing 100875, China e-mail: [email protected] A. Lü Institute of Geographic Sciences and Natural Resources Research, China Academy of Sciences, Beijing 100101, China B. He College of Global Change and Earth System Science, Beijing Normal University, Beijing 100875, China
because the heading stage, which is very important to crop yield of winter wheat, moved forward, the impact of water deficit in late April was more serious to crop yield. Keywords Climate variability . Crop growth simulation . Crop water production function . DSSAT
Introduction The North China Plain (NCP) is one of the most important agricultural regions in China (Wu et al. 2006). Crop rotation is dominated by a double-cropping system of winter wheat and summer maize. Approximately 45 % of wheat planting area of China and 33 % of maize are located in the NCP (Guo et al. 2010). The NCP has a temperate continental monsoon climate, and more than 70 % of precipitation distributes from July to September (Fu et al. 2009). The lack of precipitation during the winter wheat growing season (October to June) causes a serious water deficit for crop production (Chen et al. 2010). In the past few decades, climate variability have affected the hydrological cycle in many places around the world including the NCP, which also much influence the occurrence and severity of the water deficit (Gao et al. 2006; Huntington 2006; Ma and Fu 2006; Sheffield and Wood 2008). To improve the agricultural and water management, it is necessary to analyze how the water deficit during the winter wheat growing season is affected by climate variability and the impact of the water-deficit variability on crop yield. Many drought indices, such as the Standard Precipitation Index (SPI) (McKee et al. 1993) and the Palmer Drought Severity Index (PDSI) (Palmer 1965), have been used to investigate the historical changes in drought occurrences. The SPI can be a useful tool for analyzing the soil water deficit for rainfed crops by calculating the anomalies in precipitation (Bordi et al. 2009). The PDSI involves evapotranspiration and
Int J Biometeorol
calculates the soil water deficit based on a genetic two-layer soil model (Heim 2002; Dai 2011). The frequency of dry years, which was calculated by SPI and PDSI showed statistic significant increasing trend in northern China from the 1960s (Zhai et al. 2010). Many process-based models such as the Variable Infiltration Capacity (VIC) land surface model and the soil water balance model were also used to simulate the change in soil moisture (Andreadis et al. 2005). For example, the area that suffered from a water deficit showed an increasing trend in East Asia from 1950 to 2000 based upon simulations by the VIC (Sheffield and Wood 2008). A significantly decreasing trend in soil moisture was also found in northern China based on a soil water balance model (Tao et al. 2003). However, crop water demand varies over the growth stages (Igbadun et al. 2007). Rising temperature cannot only influence the evapotranspiration rate (Meehl 2007; Easterling et al. 2007) but also affect the crop development rate, which would change the temporal dynamics of the crop water requirement (Badeck et al. 2004). However, the meteorological drought indices and the soil water balance model do not consider the impact of changing crop development rate on crop water demand. In addition, a crop yield reduction is one of the main impacts of the water deficit on agriculture (Mkhabela et al. 2010) and can indicate the severity of it. The impact of a water deficit on crop yield also varies by the growth stages (Shahid 2011). Nageswara Rao et al. (1985) found that the water stress during the seed-filling phase had a more serious impact on the peanut yield. For rice, however, Boonjung and Fukai (1996) found that the impact of water deficit during the panicle development was the most serious on yield. Zhang et al. (1999) found that the water condition during the stages of heading and flowering was more important to winter wheat. However, those meteorological drought indices cannot reflect the dynamics of the impact of the water deficits during different growth stages on crop yield and indicate the severity of the water deficit. Crop models, based on detailed descriptions of the basic physiological and ecological processes of crops (Boote et al. 2011; Jones et al. 2003; Rosenzweig and Iglesias 1998; Williams et al. 1989), can be a good tool to analyze the variation of the water deficit and the impact of the waterdeficit variability on crop yield. Crop model was always developed in one or few stations. When it was used in other regions, it may have some uncertain results because of the differences in soil, weather, and crop type (Xiong et al 2008). Model evaluation was usually the first step to crop model application. Crop models have been demonstrated to be a useful tool for investigating the changes in soil water content (Jamieson et al. 1998; Singh et al. 2008) and have a good ability to simulate crop yield under various water conditions (Liu et al. 2011). They were also used to analyze the impact of climate change and climate variability on agricultural production (Alexandrov and Hoogenboom 2001; Xiong et al. 2010;
Chavas et al. 2009; Thomson et al. 2006). In the NCP, Chen et al. (2010) simulated the impact of climate variability on the water balance through a farming system model, Agricultural Production Systems Simulator (APSIM). Mo et al. (2009) analyzed the responses of the crop yield, water consumption, and water use efficiency to climate change in winter wheat and summer maize, based on crop growth simulations under IPCC SRES A2 and B1 future climate scenarios by the Vegetation Interface Processes (VIP) in the NCP. Crop models can also simulate the impact of changing temperature on crop phenology (Guo et al. 2010; Xiao et al 2012), which can reflect their impact on the temporal dynamics of the crop water demand. By rainfed simulation, it can show the impact of climate variability on the soil water content (Eitzinger et al. 2003). Therefore, well-calibrated and validated crop models are basically very suitable for this study to analyze the impact of climate variability on the water deficit and investigate these changes’ impact on crop yield. Over the past few decades, climate change had affected the crop development rate and the temporal dynamics of crop water requirement and supplement, which also changed the temporal distribution of the water deficit. However, those climate change effects cannot be investigated by the meteorological drought indices. Thus, the objective of this study was to investigate the changes in water deficits during the growing season of winter wheat due to climate change, and their impacts on crop yield in the NCP were also analyzed based on the crop growth simulation.
Materials and methods In this study, the Crop-Environment Resource Synthesis (CERES) model in the Decision Support System for Agrotechnology Transfer (DSSAT) was chosen to simulate the crop growth of winter wheat in 12 selected stations under two treatments of automatic irrigation and rainfed during the period from 1961 to 2010. The variation in the water deficit without artificial affection, which is only affected by climate variability, was analyzed. Study region and data sources The NCP is located in northern China (31–42° E, 114–121° N) and forms part of the alluvial plain developed by the Yellow River, the Huaihe River, and the Haihe River (Liu et al. 2010). The soil types of Mollic Gleysols,Chromic Cambisols, and Eutric Cambisols are widely distributed throughout the NCP (FAO 2003). In this study, 12 stations, including Beijing, Tangshan, Huimin, Dezhou, Laiyang, Zibo, Anyang, Zhengzhou, Shangqiu, Xuchang, Suzhou, and Huaian were selected as data base (Table 1). The 12 selected stations can represent the spatial patterns of precipitation and temperature
Int J Biometeorol Table 1 Basic information about the 12 selected stations in the NCP Name
Abb.
Long. (° E)
Lat. (° N)
Subregion
Soil type
P (mm)
PET (mm)
Anyang Beijing Tangshan Dezhou Huimin Zibo Laiyang Zhengzhou Xuchang Shangqiu Suzhou Huaian
AY BJ TS DZ HM ZB LY ZZ XC SQ SZ HA
114.35 116.63 118.15 116.32 117.53 118.00 120.70 113.65 113.85 115.66 116.98 119.03
35.93 39.92 39.67 37.43 37.50 36.83 36.93 34.72 34.01 34.45 33.63 33.60
Central Northern Northern Central Central Eastern Eastern Southwestern Southwestern Southern Southeastern Southestern
Chromic Cambisols Chromic Cambisols Chromic Cambisols Mollic Gleysols Mollic Gleysols Eutric Cambisols Eutric Cambisols Chromic Cambisols Mollic Gleysols Mollic Gleysols Mollic Gleysols Mollic Gleysols
158 173 125 155 141 184 227 199 247 231 287 318
479 459 515 522 520 459 473 541 525 537 521 507
Pprecipitation during the winter wheat growing season, PET potential evapotranspiration during the winter wheat growing season based on the method of Penman
in the NCP. The average precipitation during the winter wheat growing season decreases gradually from approximately 300 mm in the southern region to less than 150 mm in the northern and central regions. The water deficit (potential evapotranspiration subtract precipitation) increases from 200 mm in the southern region to about 400 mm in the central region. The 12 stations that were evenly distributed throughout the NCP contain each gradient of water deficit in the NCP. Daily meteorological data from 1961 to 2010 of the selected stations were collected from the China Meteorological Administration, including the minimum and maximum air temperature (°C), precipitation (mm), the relative humidity (%), the sunshine duration (h), and the daily average wind speed (m/s). The sunshine duration was converted to solar radiation based on the Ångström function (Chen et al. 2010). The soil data, which included the number of soil layers and some attributes for each layer, including depth, percent sand, percent silt, percent clay, PH, percent organic carbon, and percent calcium carbonate were collected from the Gridded FAO/UNESCO Soil Units (FAO/UNESCO 1992). Model description The DSSAT is organized by modules of weather, soil, plant, the soil–plant–atmosphere interface, and crop management (Jones et al. 2003). Models for more than 28 crops are integrated into the plant module, which can simulate the physiological changes in crop growth and development (Ritchie et al. 1998; Hoogenboom et al. 2011). The DSSAT can simulate the changes in soil water and nitrogen content and represent the impact of these changes on crop growth accurately, which has been widely used in agricultural management, environmental pollution modeling, and yield forecasting (Alexandrov and
Hoogenboom 2001; Gerakis and Ritchie 1998; Xiong et al. 2008; Zhao et al. 2011). CERES models in DSSAT-Cropping System Models (CSM) have been recognized as important tools for assessing the impact of climate change and climate variability on crop growth and development (Tubiello and Ewert 2002). Based on the observation of crop growth of winter wheat under five irrigation treatments in Gucheng in the NCP in 2007–2009, we calibrated the cultivar coefficients of CERES-Wheat in the DSSAT (version 4.5) (Table 2). The relative root mean square error (RMSE) of the observed and simulated crop yield was 0.39 t/ha. The correlation of the observed and simulated biomass during the crop growing season was 4.82 %. We also compared the observation and simulated crop yield in the 12 selected stations from 2005 to 2009. The normalized RMSE was 16.5 %, which indicated CERES-Wheat had a good performance in regional simulation (Liu 2013). In this study, the crop growth of winter wheat in the 12 selected stations over the past 50 years, following treatments of automatic irrigation and rainfed, was simulated by the DSSAT. Irrigation under the automatic irrigation treatment was applied when the soil water content of the top 30 cm was less than 50 % of its capacity, which ensure that no drought stress occurs. To represent the changes in the water deficit without any artificial affection, which is only affected by climate factor, rainfed treatment was used and the difference in evapotranspiration between the two treatments was used to reflect the water deficit. Trends in temperature and precipitation Trends in temperature and precipitation in each 10-day period during the winter wheat growing season were also analyzed to
Int J Biometeorol Table 2 Value of the key cultivar coefficients after calibration
Cultivar coefficient
Introduction
Range
Value
P1V P1D P5 G1 G2
Days required for vernalization (optimum temperature) Photoperiod response (% reduction in rate/10 h drop in pp) Grain filling (excluding lag) phase duration (°Cd) Kernel number per unit canopy weight at anthesis (no./g) Standard kernel size under optimum conditions(mg)
5–60 25–120 350–600 17–30 20–55
40 55 400 20 50
G3 PHINT
Standard, nonstressed mature tiller weight (g dwt) Interval between successive leaf tip appearances(°Cd)
1.0–1.5 60–90
show climate variability during the growing season. These trends were obtained using the method developed by Hirsch et al. (1982). The trend magnitude was defined as (Eq. 1): X jk −X ik β ¼ median j−i
ðk ¼ 1; 2; ⋯; n:1961 < i < j < 2010Þ
ð1Þ where β was the trend magnitude; i and j were the years; and Xjk was the average temperature or precipitation in the kth 10-day period of the jth year. Assessing the impact of the water deficit on crop yield A reduction in crop yield is a main impact of agricultural drought. This study used the Yield Reduction by Water Deficit (YRWD) indicator to assess the impact of the water deficit. The impact of the water deficit on crop growth varies over phenological stages. The crop water production function can represent the different impacts of the water deficit during different growth stages on crop yield (Igbadun et al. 2007) (Eq. 2): X ETai Ya ¼ Ai Ym ETmi i¼1 n
ð2Þ
where Ya was the crop yield for the water-stressed treatment; Ym was the crop yield for the treatment without water stress; ETai was the total evapotranspiration amount at the growth stage i in the treatment with water stress; ETmi was the total evapotranspiration amount at the growth stage i in the treatment without water stress; and Ai was the crop sensitivity index to drought stress during the growth stage i, which can represent the different impacts of water deficit during different crop growth stages on the final yield. The crop water production function (Eq. 2) is a multiply linear function. To obtain the Ai in the crop water production function at these 12 selected stations, the crop growth was simulated in 7 deficient irrigation treatments and 1 automatic irrigation treatment in the past 30 years by the DSSAT. The irrigation stages for the seven deficient irrigation treatments
1.5 85
are shown in Table 3. Irrigations were only applied on the first day of each selected growth stage and ensured the soil water content was at the field capacity. Crop yield and evapotranspiration in the treatment of automatic irrigation were considered to be Ym and ETm, and those in the treatment of deficient irrigation were considered to be Ya and ETa in Eq. 2. The multiple linear regression was used to estimate Ai. The stages in the middle of the winter wheat growing season, such as jointing, heading, and flowering in March and April, have a larger crop sensitivity index to drought stress, and the crop sensitivity index to drought stress in winter is lower during the growing season of winter wheat (Zhang et al. 1999). The crop growth simulations of winter wheat under the treatments of rainfed and automatic irrigation were used to calculate the daily YRWD for the rainfed treatment. The yield ratio (Ya/Ym) on the ith day reflected the impact on final yield of a water deficit from planting to the ith day, which was calculated by the crop water production function (Eq. 2), based on the assumption of sufficient water supplementation after the ith day until maturity. When calculating the yield ratio on the ith day, the ETa from planting to the ith day in the crop water production function was obtained from the evapotranspiration under the rainfed treatment, and the ETa from i+ 1th day to maturity was obtained from the evapotranspiration under the automatic Table 3 Description of deficient irrigation treatments for calculating the crop sensitivity index to drought stress Irrigation treatment
Time of irrigation applied
I1 I2 I3 I4 I5 I6
Heading Stem elongation and heading Stem elongation, heading, and grain filling Reviving, stem elongation, heading, flowering Before freezing, reviving, flowering, grain filling Before freezing, reviving, stem elongation, flowering, grain filling Before freezing, reviving, stem elongation, heading, flowering, and grain filling
I7
Int J Biometeorol 400
10.0 Annual Precipitation Average Precipitation Temperature
9.5 9.0 8.5
200
8.0 7.5
o
Precipitation (mm)
300
Temperature ( C)
Fig. 1 Time series of precipitation and average temperature during the period from late October to May in the selected 12 stations from 1961 to 2010 (The average precipitation in the past 50 years and the linear regression of temperature are also provided.)
7.0
100
6.5 0 1960
1970
1980
1990
6.0 2010
2000
Year
irrigation treatment. The ETm during the whole growing season was obtained from the evapotranspiration under the automatic irrigation treatment. The value of the YRWD on the ith day was calculated as the yield ratio on i−1th day minus the yield ratio on ith day. The variation of the average YRWD in each 10-day period during the growing season of winter wheat was used to represent the variation of the impact of the changing water deficit on crop yield. The results of the trends in the YRWD still provided information regarding the application of irrigation and improve irrigation efficiency. The spring season (from March to May) was the main growing season of winter wheat in the NCP. Therefore, this study chose the period from March to May to analyze the changes in the water deficit and their impacts on crop yield.
Results and discussions Climatic trends The average temperature from late October to May has a significantly increasing trend (p
