Spatiotemporal Analysis of Maize Water Requirement in the Heilongjiang Province of China during 1960–2015

: Climate change will have a signiﬁcant e ﬀ ect on crop water requirement ( ET c ). The spatial and temporal variations of water requirement of maize under climate change are essential elements when conducting a global water security assessment. In this paper, annual reference crop evapotranspiration ( ET 0 ) and the crop water requirement of maize were calculated by the single crop coe ﬃ cient method. The crop water surplus deﬁcit index ( CWSDI ) and coupling degree of ET c and e ﬀ ective precipitation ( P e ) were calculated to analyze the relationship between ET c , ET 0 , and P e . The result shows that maize average annual ET 0 , ET c , and precipitation were 552.97, 383.05, and 264.97 mm, respectively. Moreover, ET 0 , ET c , and P e decreased by 3.28, 2.56, and 6.25 mm every decade from 1960 to 2015. The ET c decreased less than P e did, which led to the decreasing of both CWSDI and the coupling degree of ET c and P e . The tendency of ET 0 , ET c decreased ﬁrst and then increased, while P e and CWSDI increased ﬁrst and then decreased, from west to east of the Heilongjiang Province. In addition, the highest ET 0 , ET c , and lowest CWSDI and P e were found in the western part of Heilongjiang Province. This study indicated that even though the water deﬁcit in the western region was alleviated and the water deﬁcit in the eastern region grew gradually serious from 1960 to 2015, the drought situation in western Heilongjiang Province should still be taken seriously. higher ET c (458.51 and 434.80 mm, respectively). Regions in which the climate tendency of CWSDI showed a decreasing trend also showed a decreasing tendency of ET c . The current results were similar; it was found that the tendency of Pe and maize ET c during the growth period in Heilongjiang Province showed the same trend as in Northeast China. The decrease of ET c in Heilongjiang Province was faster than that in Northeast China, and the decrease of P e in this study was slower than that in Northeast China in general.


Introduction
Global warming, characterized by rising temperature, is widely accepted as the main feature of climate change [1]. The Fifth Assessment Report of the United Nations Intergovernmental Panel on Climate Change (IPCC) showed that the annual Northern Hemisphere temperatures from 1983 to 2013 were the warmest for a 30-year period during the past 800 years [2,3]. Global warming due to the greenhouse effect is expected to cause major climatic changes in some areas [4]. Furthermore, climate change has a significant effect on crop water requirement (ET c ), actual evapotranspiration, and grain yield due to increased temperature and variable precipitation, wind speed, and sunshine hours, especially in areas with poor adaptability. Actual evapotranspiration, as a major process drought indicator, which indicated the most significant impacts of drought occurred during the grain-filling and heading stages [36]. In this study, crop water surplus deficit index (CWSDI) and the coupling degree of ET c and P e were used to express the degree of water deficit and surplus, and the relationship between ET c and P e during maize growth periods from 1960 to 2015. These two parameters can express the relationship between ET c and P e more intuitively.
Heilongjiang Province is one of the most important grain-producing areas in China. It is located in Northeast China and its grain production accounted for 11.3% of the nation's grain yield in 2017 [14]. As a result, Heilongjiang Province plays an important part in ensuring China's agricultural food security. In this region, precipitation is governed by the East Asian summer monsoons, and it has experienced extreme droughts and waterlogging events since the 1970s. Due to climate change and variation of precipitation, water supply and requirements vary in the Heilongjiang Province in a spatiotemporal manner. Spatiotemporal analysis of maize ET 0 , ET c , P e , and CWSDI, and the coupling degree of ET c and P e , provide a theoretical and practical basis for improving the management efficiency of water resources and formulating a reasonable irrigation system.
The purposes of this study were (1) to quantify ET 0 , ET c , and P e of maize in Heilongjiang Province; (2) to determine the spatiotemporal distribution characteristics of ET 0 , ET c , and P e during the growth stage for maize in Heilongjiang Province during 1960-2015; (3) to quantify the crop water surplus deficit index (CWSDI) using the ratio of ET c and P e based on actual records during the growth stage of maize for the 55-year period; and (4) to analyze the spatiotemporal distribution characteristics of CWSDI and the coupling degree of ET c and P e during the growth stage of maize in Heilongjiang Province during 1960-2015.

Study Area
In this study, daily meteorological data were collected from 26 agricultural meteorological stations during 1960-2015 in Heilongjiang Province, including the highest temperature, lowest temperature, average relative humidity, wind speed, sunshine hour, precipitation, latitude, longitude, and latitude of each station. The quality of above data were carefully assessed, and the missing data were estimated using methods suggested by the FAO-56 [17]. Observation data during maize growth periods were obtained from 14 agricultural observation stations in the study area from 1991 to 2008. The above data were collected from the China Meteorological Data Sharing Service System [37]. Figure 1 shows the distribution of the research area and observation stations. According to the accumulated temperature zone of crop varieties in Heilongjiang Province [38], issued by Heilongjiang Agricultural Commission, and Area Layout Planning of High-quality and High-yield Main Food Crops in Heilongjiang Province in 2015 [39], the sixth accumulated temperature zone was not suitable for maize planting; thus, the sixth accumulated temperature zone was not included in this paper.
Water 2020, 12, x FOR PEER REVIEW  3 of 15 grain-filling and heading stages [36]. In this study, crop water surplus deficit index (CWSDI) and the coupling degree of ETc and Pe were used to express the degree of water deficit and surplus, and the relationship between ETc and Pe during maize growth periods from 1960 to 2015. These two parameters can express the relationship between ETc and Pe more intuitively. Heilongjiang Province is one of the most important grain-producing areas in China. It is located in Northeast China and its grain production accounted for 11.3% of the nation's grain yield in 2017 [14]. As a result, Heilongjiang Province plays an important part in ensuring China's agricultural food security. In this region, precipitation is governed by the East Asian summer monsoons, and it has experienced extreme droughts and waterlogging events since the 1970s. Due to climate change and variation of precipitation, water supply and requirements vary in the Heilongjiang Province in a spatiotemporal manner. Spatiotemporal analysis of maize ET0, ETc, Pe, and CWSDI, and the coupling degree of ETc and Pe, provide a theoretical and practical basis for improving the management efficiency of water resources and formulating a reasonable irrigation system.
The purposes of this study were (1) to quantify ET0, ETc, and Pe of maize in Heilongjiang Province; (2) to determine the spatiotemporal distribution characteristics of ET0, ETc, and Pe during the growth stage for maize in Heilongjiang Province during 1960-2015; (3) to quantify the crop water surplus deficit index (CWSDI) using the ratio of ETc and Pe based on actual records during the growth stage of maize for the 55-year period; and (4) to analyze the spatiotemporal distribution characteristics of CWSDI and the coupling degree of ETc and Pe during the growth stage of maize in Heilongjiang Province during 1960-2015.

Study Area
In this study, daily meteorological data were collected from 26 agricultural meteorological stations during 1960-2015 in Heilongjiang Province, including the highest temperature, lowest temperature, average relative humidity, wind speed, sunshine hour, precipitation, latitude, longitude, and latitude of each station. The quality of above data were carefully assessed, and the missing data were estimated using methods suggested by the FAO-56 [17]. Observation data during maize growth periods were obtained from 14 agricultural observation stations in the study area from 1991 to 2008. The above data were collected from the China Meteorological Data Sharing Service System [37]. Figure 1 shows the distribution of the research area and observation stations. According to the accumulated temperature zone of crop varieties in Heilongjiang Province [38], issued by Heilongjiang Agricultural Commission, and Area Layout Planning of High-quality and High-yield Main Food Crops in Heilongjiang Province in 2015 [39], the sixth accumulated temperature zone was not suitable for maize planting; thus, the sixth accumulated temperature zone was not included in this paper.

Effective Precipitation
In this study, we used the method recommended by the Soil Conservation Agency of the U.S. Department of Agriculture to calculate effective precipitation [40], as follows: where P e is effective precipitation (mm) and P is precipitation (mm).

Maize Water Requirement
Maize water requirement was calculated with the CROPWAT8.0 for Windows computer program, it was designed for the calculation of ET c from existing climate, crop, and soil data. It has been appropriately used as a tool for calculating ET c under climate change and widely used in Asia, Europe, and Africa with reliable high performance [40,41]. The altitude, latitude, longitude, and daily highest temperature, lowest temperature, average relative humidity, wind speed, sunshine hour data of each station were loaded to the Climate/ET 0 module to calculate ET 0 . The planting date, harvest date, K c values, and length of each growth period were loaded to the Crop module to calculate ET c . The maize water requirement was calculated by the single crop coefficient method [10]. Total maize water required was accumulated from the daily water requirement during the growth period based on the crop coefficient in different months, as in Equation (2). Crop evapotranspiration under standard conditions, denoted ET c , is the evapotranspiration from disease-free, well-fertilized crops, grown in large fields, under optimum soil water conditions, and achieving full production under the given climatic conditions [40].
where ET c is the crop water requirement (mm), ET 0 is the reference crop evapotranspiration (mm), and K c is the crop coefficient. FAO-56 divides crop growth period into four periods: The initial stage (L ini ) runs from planting date to approximately 10% ground cover, the crop development stage (L dev ) runs from 10% ground cover to effective full cover, the mid-season stage (L mid ) runs from effective full cover to the start of maturity, and the late season stage (L late ) runs from the start of maturity to harvest or full senescence [40]. The whole growth period of maize was divided into four stages: Sowing-Seven leaf stage (L ini ), Seven leaf stage-Tasseling stage (L dev ), Tasseling stage-Milk ripening stage (L mid ), and Milk ripening stage-Mature stage (L late ). The FAO-56 method reports generalized K c values for different growing stages of maize under standard conditions. The single crop coefficient K c is described by K c-ini , K c-mid and K c-end for the initial, middle, and late growth stages, respectively, with respective recommended values of 0.3, 1.20, and 0.35 for maize. However, to improve the estimation of the K c curve for a specific site, the recommended values need to be adjusted to obtain a local K c value. In this paper, K c-ini , K c-mid and K c-end were obtained from the recommended values in the FAO-56 and were internal automatic correction by CROPWAT model based on average infiltration depth, wind speed, humidity, crop height at different growing stage, and the correction formula of K c has been embedded into the CROPWAT model [41].
ET 0 was calculated by the Penman-Monteith method recommended by the FAO, as in Equation (3).

Crop Water Surplus Deficit Index
To accurately reflect water requirements and the water supply status of maize, we established the water requirement index during the maize growth stage ET c and the water supply index P e , and constructed the crop water surplus deficit index (CWSDI) to describe maize water surplus and deficit during the maize growth stage based on maize water deficit index [42], as in Equation (4): where CWSDI is the crop water surplus deficit index, P e is effective precipitation (mm), ET c is crop water requirement (mm).

Coupling Degree of ET c and P e
The coupling degree of ET c and P e , λ i , represents the extent to which P e meets ET c , as in Equation (5).

Climate Tendency Rate
The climatic tendency of factors is expressed by a linear equation using the least squares method, as in Equation (6).
whereŶ represents the fitted values of the studied elements, t is the corresponding year, and b 0 and b are the regression coefficients. 10b 0 is the climate tendency rate, which means a meteorological element per 10-year(10a) change. A positive value indicates an increasing trend of the corresponding meteorological elements, and negative values indicate a decreasing trend.

Trend Test
The Mann-Kendall trend test is a non-parametric statistical method that can be used to reveal the changing trend of time series and more prominent adaptability to a non-normal distribution of meteorological data. Positive and negative values of the statistical variable Z represent the trend of data change; when Z has an absolute value greater than or equal to 1.64, 2.32, or 2.56, it means passing the significance test with 95%, 99%, and 99.9% reliability, respectively [30]. This paper used the Mann-Kendall trend test to test the parameter changing trend of ET 0 , ET c , precipitation, and CWSDI.

Data Processing
In this study, we used MATLAB 2004b software to (i) calculate ETc, P e , and the climate tendency rate of maize in 26 meteorological stations of Heilongjiang Province and (ii) for the Mann-Kendall trend test. We used the spatial analysis function of ArcMap 10.2 toolbox to interpolate and map the parameters, the ordinary kriging method was employed to prepare the spatial mapping of ET 0 , ET c , P e , and CWSDI and coupling degree of ETc and Pe, as it was widely used in geoscience for interpolating from sparse sample data.

Spatial and Temporal Variation of ET 0
The spatial distribution of ET 0 and its climate tendency in this study area during 1960-2015 is shown in Figure 2. ET 0 ranged from 422.78 to 697.60 mm, with an average of 552.97 mm. Average ET 0 values during 1960-1979, 1980-1999, and 2000-2015 were 563.65, 545.42, and 549.82 mm, respectively. Figure 2 shows that there was a marked decrease and a then slight increase of ET 0 from west to east in the Heilongjiang region from 1960 to 2015. The high-value areas were mainly distributed in the western region with an average of more than 584.66 mm and the low-value areas were mainly distributed in the northeast area, with an average of less than 477.90 mm. The climate tendency rate of ET 0 was between −22.81 and 13.76 mm/(10a) from 1960 to 2015, with an average of −3.28 mm/(10a). The figure shows that there was a gradual decrease of ET 0 in the western area, with a climate tendency rate of less than −4.02 mm/(10a), and the climate tendency of ET 0 in Anda reached −17.98 mm/(10a). The climate tendency of ET 0 in the northern and eastern areas showed an increasing trend, with a climate tendency rate of more than 0.434 mm/(10a).
Water 2020, 12, x FOR PEER REVIEW 6 of 15 in the Heilongjiang region from 1960 to 2015. The high-value areas were mainly distributed in the western region with an average of more than 584.66 mm and the low-value areas were mainly distributed in the northeast area, with an average of less than 477.90 mm.  1960-1999 and 1960-2015. The climate tendency rate of ET0 was a gradual decrease in the western area; that of Anda was less than −10 mm/(10a) during 1960-1999, and there were eight areas less than −10 mm/(10a) during 1980-2015. Other regions showed a slight decrease, except for the northern areas of Heilongjiang. Huma and Sunwu, in the north, showed increasing trends during 1980-2015, which were statistically significant at the level of p < 0.001.

Spatial and Temporal Variation of Maize ETc
The spatial distribution of ETc and its climate tendency in this study area during 1960-2015 is shown in Figure 3. ETc ranged from 294.82 to 482.74 mm, with an average of 383.05 mm. Average ETc values during 1960-1979, 1980-1999, and 2000-2015 were 393.33, 376.09, and 379.73 mm, respectively. Figure 3 shows that there was a marked decrease and a then slight increase in the value of ETc from west to east in the Heilongjiang region from 1960 to 2015. The high-value areas were mainly distributed in the western part, with an average of more than 401.64 mm. The low-value areas were mainly distributed in the northeast area, with an average of less than 351.85 mm. The climate tendency rate of ETc was between −13.20 and 7.00 mm/(10a) from 1960 to 2015, with an average of −2.56 mm/(10a). There was a gradual increase in the climate tendency rate of ETc from west to east, with 20 stations recording negative values. Although ETc showed a decreasing trend, the climate tendency rate of ETc in northern areas was positive and was statistically significant at the level of p < 0.01. The climate tendency rate of ET 0 was a gradual decrease in the western area; that of Anda was less than −10 mm/(10a) during 1960-1999, and there were eight areas less than −10 mm/(10a) during 1980-2015. Other regions showed a slight decrease, except for the northern areas of Heilongjiang. Huma and Sunwu, in the north, showed increasing trends during 1980-2015, which were statistically significant at the level of p < 0.001.

Spatial and Temporal Variation of Maize ET c
The spatial distribution of ET c and its climate tendency in this study area during 1960-2015 is shown in Figure 3. ET c ranged from 294.82 to 482.74 mm, with an average of 383.05 mm. Average ET c values during 1960-1979, 1980-1999, and 2000-2015 were 393.33, 376.09, and 379.73 mm, respectively. Figure 3 shows that there was a marked decrease and a then slight increase in the value of ET c from west to east in the Heilongjiang region from 1960 to 2015. The high-value areas were mainly distributed in the western part, with an average of more than 401.64 mm. The low-value areas were mainly distributed in the northeast area, with an average of less than 351.85 mm. The climate tendency rate of ET c was between −13.20 and 7.00 mm/(10a) from 1960 to 2015, with an average of −2.56 mm/(10a).
There was a gradual increase in the climate tendency rate of ET c from west to east, with 20 stations recording negative values. Although ET c showed a decreasing trend, the climate tendency rate of ET c in northern areas was positive and was statistically significant at the level of p < 0.01.  1960-1999, 1980-2015, and 1960-2015 were −1.85, −2.37, and −2.56 mm/(10a), respectively. The climate tendency rate of ETc during 1960-1999 was the lowest, compared with that during 1980-2015 and 1960-2015. The climate tendency rate in this figure shows a gradual decrease in the western area; only that of Anda was less than −10 mm/(10a) during 1960-1999, and there were seven areas with values less than −10 mm/(10a) during 1980-2015. Compared with these two periods, the highest climate tendency rate was in the eastern area from 1960 to 1999 and in the northern area during other periods.

Spatial and Temporal Variation of Pe
The spatial distribution of Pe during the maize growing period and its climate tendency in this study area during 1960-2015 is shown in  The climate tendency rate in this figure shows a gradual decrease in the western area; only that of Anda was less than −10 mm/(10a) during 1960-1999, and there were seven areas with values less than −10 mm/(10a) during 1980-2015. Compared with these two periods, the highest climate tendency rate was in the eastern area from 1960 to 1999 and in the northern area during other periods.

Spatial and Temporal Variation of P e
The spatial distribution of P e during the maize growing period and its climate tendency in this study area during 1960-2015 is shown in Figure 4

Spatial and Temporal Variation of CDSWI
The spatial distribution of the crop water surplus deficit index (CWSDI) and its climate tendency in this study area during 1960-2015 is shown in Figure 5. CWSDI ranged from −60% to 1%, with an average of −28%. Average CWSDI values during 1960-1979, 1980-1999, and 2000-2015 were −32%, −24%, and −30%, respectively. Figure 5 shows almost all areas were in a water deficit state during maize growth periods. From Figure 5, CWSDI shows a trend of slowdown in water deficit from west to east. Furthermore, the water deficit of maize was serious, with CWSDI values of less than −40% in the western and a few northeastern areas. The most serious water deficit area was Tailai in all three periods. In this study, only Yichun was in a water surplus state. The CWSDI climate tendency rate showed little change during the study period; the climate tendency rates of CWSDI were 0.14, −0.96, and 0.09%/(10a), respectively. The worst decline area was Tonghe, and the decreased trend of Tonghe was statistically significant at the level of p < 0.01. The climate tendency rate of CWSDI showed the largest decrease during 1980-2015, compared with that during 1960-1999 and 1960-2015, when it increased.

Spatial and Temporal Variation of CDSWI
The spatial distribution of the crop water surplus deficit index (CWSDI) and its climate tendency in this study area during 1960-2015 is shown in Figure 5. CWSDI ranged from −60% to 1%, with an average of −28%. Average CWSDI values during 1960-1979, 1980-1999, and 2000-2015 were −32%, −24%, and −30%, respectively. Figure 5 shows almost all areas were in a water deficit state during maize growth periods. From Figure 5, CWSDI shows a trend of slowdown in water deficit from west to east. Furthermore, the water deficit of maize was serious, with CWSDI values of less than −40% in the western and a few northeastern areas. The most serious water deficit area was Tailai in all three periods. In this study, only Yichun was in a water surplus state. The CWSDI climate tendency rate showed little change during the study period; the climate tendency rates of CWSDI were 0.14, −0.96, and 0.09%/(10a), respectively. The worst decline area was Tonghe, and the decreased trend of Tonghe was statistically significant at the level of p < 0.01. The climate tendency rate of CWSDI showed the largest decrease during 1980-2015, compared with that during 1960-1999 and 1960-2015, when it increased.

Spatial and Temporal Variation of Coupling Degree of ET c and P e
The distribution of the coupling degree of ET c and P e during the maize growth period in this study from 1960 to 2015 is shown in Figure 6. The coupling degree of ET c and P e ranged from 0.40 to 1.00, with an average value of 0.71, and first increased and then decreased from west to east. The high-value areas with an average coupling degree of ET c and P e greater than 0.85 were Yichun, Tieli, and Sunwu in the north, and the low-value areas below 0.5 were Anda, Tailai, and Fuyu in the west. The climate tendency rate of the coupling degree of ET c and P e in the growth period was between −0.024 and 0.025/(10a); the average was 0.0028/(10a), and it showed an increasing trend. There were 15 stations in which the climate tendency rate was negative. Tonghe, Baoqing, and Fujin in the east showed a declining trend, and the decreasing trend of Tonghe was statistically significant at the level of p < 0.01 from 1980 to 2015.
periods. In this study, only Yichun was in a water surplus state. The CWSDI climate tendency rate showed little change during the study period; the climate tendency rates of CWSDI were 0.14, −0.96, and 0.09%/(10a), respectively. The worst decline area was Tonghe, and the decreased trend of Tonghe was statistically significant at the level of p < 0.01. The climate tendency rate of CWSDI showed the largest decrease during 1980-2015, compared with that during 1960-1999 and 1960-2015, when it increased.

Spatial and Temporal Variation of Coupling Degree of ETc and Pe
The distribution of the coupling degree of ETc and Pe during the maize growth period in this study from 1960 to 2015 is shown in Figure 6. The coupling degree of ETc and Pe ranged from 0.40 to 1.00, with an average value of 0.71, and first increased and then decreased from west to east. The high-value areas with an average coupling degree of ETc and Pe greater than 0.85 were Yichun, Tieli, and Sunwu in the north, and the low-value areas below 0.5 were Anda, Tailai, and Fuyu in the west. The climate tendency rate of the coupling degree of ETc and Pe in the growth period was between −0.024 and 0.025/(10a); the average was 0.0028/(10a), and it showed an increasing trend. There were 15 stations in which the climate tendency rate was negative. Tonghe, Baoqing, and Fujin in the east showed a declining trend, and the decreasing trend of Tonghe was statistically significant at the level of p < 0.01 from 1980 to 2015.

Spatial and Temporal Variation of Coupling Degree of ETc and Pe
The distribution of the coupling degree of ETc and Pe during the maize growth period in this study from 1960 to 2015 is shown in Figure 6. The coupling degree of ETc and Pe ranged from 0.40 to 1.00, with an average value of 0.71, and first increased and then decreased from west to east. The high-value areas with an average coupling degree of ETc and Pe greater than 0.85 were Yichun, Tieli, and Sunwu in the north, and the low-value areas below 0.5 were Anda, Tailai, and Fuyu in the west. The climate tendency rate of the coupling degree of ETc and Pe in the growth period was between −0.024 and 0.025/(10a); the average was 0.0028/(10a), and it showed an increasing trend. There were 15 stations in which the climate tendency rate was negative. Tonghe, Baoqing, and Fujin in the east showed a declining trend, and the decreasing trend of Tonghe was statistically significant at the level of p < 0.01 from 1980 to 2015. Comparing the coupling degree during two stages and the whole research period, we found that, although the coupling degree decreased the most from 1980 to 2015, the increasing coupling Comparing the coupling degree during two stages and the whole research period, we found that, although the coupling degree decreased the most from 1980 to 2015, the increasing coupling degree during 1960 to 1999 drove the overall increasing tendency of the coupling degree of ET c and P e during 1960 to 2015.

Climate Change Impact on Water Requirement Relationship
The meteorological factors were tested by the Mann-Kendall trend test in the study area from May to September during the maize growing seasons of 1960-2015. The average max and min temperature were increased, and average humidity, average wind speed, and average sunshine hours were decreased, with climate tendency rates of 0.43, 0.20, −0.23, −0.18, and 3.10 mm/(10a), respectively (Table 1). Average values and climate tendency rates of ET 0 , ET c , P e , CWSDI, and coupling degree of ET c and P e during the maize growing seasons of 1960-2015 in the study area are shown in Tables 2 and 3. The different impacts of climate factors on the water requirement of maize and the relationship between P e and ET c in this study region are shown in Figure 7. ET 0 was reduced by the combined action of wind speed, humidity, sunshine hours, and temperature during the maize growing period, which caused ET c to decrease. The coupling degree of ET c and P e , and CWSDI were increased by the combined action with ET c reduced and P e increased. Annotations: *, **, and *** are significant at the levels of p < 0.05, p < 0.01, and p < 0.001, respectively.

Discussion
This paper aims to estimate the relationship between maize ET0, ETc, and Pe, and the impact of climate change on the water requirement of maize from 1960 to 2015 in Heilongjiang Province. The results showed that climate change during the recent 55-year timespan caused a decrease of ET0 in maize growing periods. The increase in the highest and lowest temperatures did not cause an increase in ET0. The decrease in ET0 may be due to the decrease in wind speed and sunshine. In a previous study, Adrian Piticar

Discussion
This paper aims to estimate the relationship between maize ET 0 , ET c , and P e , and the impact of climate change on the water requirement of maize from 1960 to 2015 in Heilongjiang Province. The results showed that climate change during the recent 55-year timespan caused a decrease of ET 0 in maize growing periods. The increase in the highest and lowest temperatures did not cause an increase in ET 0 . The decrease in ET 0 may be due to the decrease in wind speed and sunshine. In a previous study, Adrian Piticar et al. found that ET 0 experienced positive trends in the Republic of Moldova from 1981 to 2012 [25]. In Adrian Piticar's study, the maximum and minimum air temperatures showed increasing trends, and wind speed and relative humidity showed decreasing trends, results that were the same as those of the current study. In his study, the sunshine duration increased in all the analyzed locations, in contrast to the current study, in which they decreased. Wu Xia et al. found that the decrease of ET 0 in China from 1961 to 2015 was due to the combined action of the decrease of wind speed, sunshine, and vapor pressure. Comparing these studies, we found wind speed and sunshine hours influenced the ET 0 change, which explained the different tendency change of ET 0 between this study and that of Adrian Piticar [2]. Therefore, the impact of various climatic factors should be considered comprehensively when discussing ET 0 . Only considering the effect of temperature rise on ET 0 will result in a large deviation between the estimated values and the true value [43]. The decrease in sunshine hours in this study may be due to an increase in cloud cover or aerosols, resulting in radiation change [44]. The cloud temperature was not decreased by the negative trend of the radiation. On the other hand, the greenhouse gas effects of clouds and aerosols are more conducive to capturing heat, reducing solar radiation, and increasing air temperature, resulting in a decrease in crop evapotranspiration.
Wang et al. [45] pointed out in his paper that under the condition of sufficient irrigation in Zhaozhou county, the maize ET c was 381.0 mm in 2014. In this study, the calculated maize ET c was 375.1 mm in the same region in 2014, which was 1.5% less than his results. Li et al. [46] measured the maize ET c by Lysimeter as 370 mm in Haerbin in 2014, this paper calculated that maize ET c in Haerbin in 2014 was 392.3 mm, which was 6% larger than the experimental result. According to a study of a contour map of water demand of the main crops in China, the average ET c in 17 stations of Heilongjiang Province from 1961 to 1980 was 427 mm and the average water shortage was 103 mm [47]. The average value of maize ET c in this study was 393.33 mm from 1961 to 1980, 7.88% lower than the value of the previous study [42]. The range of ET c in Heilongjiang Province was 294.82 to 482.74 mm from 1961 to 2015. This is mainly because the K c value for each station during the whole growth period in the previous study was 0.8. The crop type, climate data, soil evaporation, and growth determined the K c value. In the current study, K c was divided and corrected according to different growth stages of maize and different regions, which increased the accuracy of the calculation. Previous research (Gao et al.) found that the climate tendency of maize ET c in Northeast China (Heilongjiang, Jilin, and Liaoning Provinces) from 1961 to 2010 was 1.2 mm/(10a) and P e showed a downward trend [42]. In the current study, the climate tendency of maize ET c was −2.27 mm/(10a) and the climate tendency of P e in Heilongjiang Province was −2.04 mm/(10a). The length of the growth period also affects ET c . ET c requires more water during dry years than in rainy years. This is consistent with FAO-56, which notes that crops need more water in dry seasons than in rainy seasons during growth periods [28]. The current study showed the same phenomenon. Regions with the greater water deficits, such as Tailai and Anda, had the lowest CWSDI (−55% and −48% for these two regions, respectively) and higher ET c (458.51 and 434.80 mm, respectively). Regions in which the climate tendency of CWSDI showed a decreasing trend also showed a decreasing tendency of ET c . The current results were similar; it was found that the tendency of Pe and maize ET c during the growth period in Heilongjiang Province showed the same trend as in Northeast China. The decrease of ET c in Heilongjiang Province was faster than that in Northeast China, and the decrease of P e in this study was slower than that in Northeast China in general.
Previous results show that the increase of temperature in recent years has led to an advance in the sowing date of maize, an increase in the total accumulated temperature of the maize growing season, an extension of the growing period, and a shift in the planting boundary to the north and expansion to the east. To realize the effective utilization of heat resources and improve yield, many new cultivars have emerged, cultivation measures have changed, and early cultivars have been gradually replaced by middle-late cultivars [48,49]. These problems extend the growth period length and increase the maize leaf number, leading to an increase of ET c increases. In this study, the sowing time and duration of each growth stage of maize were controlled by CROPWAT, which reduces the differences of maize plant characteristics caused by different cultivars in different regions, and reduces the influence of variety differences on water requirement distribution. Therefore, the rate of ET c decrease was not faster than this study, as the CWSDI decreased in the western region of Heilongjiang Province, and more irrigation water was needed in these regions. The analysis of ET c, CWSDI, P e , and the coupling degree of ET c and P e showed that the situation of maize irrigation water requirement was still severe.
Water-saving irrigation measures should be given attention. Agricultural water-saving measures, such as lower irrigation or sprinkling irrigation, and drip irrigation under mulch, are beneficial to improve water use efficiency and ensure large yields in areas of water shortage. In future research, the relationship between irrigation water requirement, ET c , and ET 0 will be examined.

Conclusions
During the past 55 years, wind speed, humidity, and sunshine hours have decreased, and the temperature has increased. These factors impact ET 0 , ET c , and P e . Results of the current study show that reference crop evapotranspiration (ET 0 ) decreased, with a declining trend of 3.29 mm/(10a) in Heilongjiang Province. ET c also showed a decreasing trend of 2.37 mm/(10a). The results showed that P e was reduced, with a climate tendency rate of −5.42 mm/(10a). Moreover, CWSDI and coupling degree of ET c and P e was increased, with climate tendency rates of 0.0888 and 0.0028/(10a), respectively. Climate change was found to have had a negative impact on the water security of Heilongjiang Province, especially for the western region, as shown by strong ET c and low precipitation. In future research, we will focus on relieving these negative influences and continue to focus on the irrigation water requirement in Heilongjiang Province, with the aim of producing an irrigation schedule and offering advice for local food security.