Temporal-spatial evolution patterns of the annual precipitation considering the climate change conditions in the Sanjiang Plain

In recent years, global climate change has become a focus of attention. Climate change alters the hydrological cycle, which affects the management of ecological environments and the utilization of water resources (Chen et al. 2010; Yang et al. 2013). Global climates have experienced a warming trend over the past 100 years (Holden et al. 2011; Ling et al. 2012), with the global mean temperature increasing by 0.74 ± 0.18 °C and multi-year mean precipitation exhibiting an average decreasing trend of 1.1 ± 1.5 mm every 10 years. However, global climate change has not been consistent, and temperature and precipitation changes have exhibited certain regional differences (Yang et al. 2013). The mean temperature in China has only increased by 0.4–0.5 °C in the past 100 years, whereas precipitation started gradually decreasing after the 1950s (Lu et al. 2014).

China is one of the 13 countries with the most limited water resources per capita in the world. According to the 2012 Statistical Bulletin on Chinese Water Activities, China had 29526.9 × 10⁸m³ in total water resources, and the total population of China was 13.54 × 10⁸ according to the Sixth Nationwide Population Census, which was published in 2011. Therefore, the per capita amount of water resources was only 2180.65 m³, which was one quarter of the average per capita amount of water resources worldwide. As industry and agriculture show great development, human interference in the environment increases and water-intensive industries continuously grow, water resource scarcity has caused a bottleneck that limits sustainable economic, social and environmental development in China. For instance, in August 2013, flood disasters caused direct economic losses of 1.10 × 10⁸yuan in Heilongjiang and 3.146 × 10¹¹yuan in China overall. In addition, 8.76 × 10⁶ people suffered from the severe spring drought in Yunnan, and 4.94 × 10⁶ people experienced difficulty gaining access to drinking water, which severely affected the lives of these people. As an important water resource, the efficient and reasonable utilization of atmospheric precipitation can alleviate the losses caused by flood and drought disasters and solve water resource shortage problems. Thus, it is important to understand the patterns of precipitation caused by climate change to better utilize rainwater resources, and this topic has drawn extensive attention from researchers at home and abroad.

As human activities increase and abnormal climate patterns appear more frequently, precipitation, which is one of the most prominent manifestations of climate change, is closely related to severe flood and drought disasters. In the 20^th century, extensive studies addressed the cause of precipitation (Braham 1959), the relationship between spring precipitation and the pollution of wheat with strontium-90 (Rickard 1964), control of precipitation (Langmuir 1950), temporal variations in the chemical substances within precipitation (Hall et al. 1993; Kalina et al. 1999; Tsakovski et al. 2000), the relationship between precipitation and disease (Parmenter et al. 1999), and estimations of precipitation (Wiemann et al. 1998). Since the 21st century, increasing climate change has inspired research in areas such as the relationship between precipitation and temperature (Hamilton et al. 2001), factors that influence precipitation (Sevruk & Mieglitz 2002), risk analysis of extreme precipitation (Palmer & Raisanen 2002), precipitation simulations under various scale conditions (Kim & Kim 2010; Mishra & Srinivasan 2010) and pattern and prediction of changes in the annual precipitation under global warming conditions (Chou & Lan 2012; Kumar & Krishnamurti 2012). However, research on the temporal-spatial pattern of changes in regional precipitation is lacking. With the intense climate change conditions that have occurred in recent decades, the temporal-spatial patterns of regional precipitation have changed significantly. From a regional perspective, studies on the temporal and spatial characteristics of precipitation changes have important theoretical significance and can help identify potential development trends of precipitation under climate change conditions and more efficiently utilize rainwater resources. Therefore, the Sanjiang Plain, which is an important commodity grain production base, was used as the study area. Wavelet theory, the Mann-Kendall trend test and ArcGIS spatial analysis theory were used to analyze annual precipitation and mean temperature data, which were collected from 1956 to 2013, at seven national weather stations in the Sanjiang Plain to identify the temporal-spatial evolution patterns of annual precipitation. These results can provide a scientific basis for the efficient utilization of rainwater resources in the Sanjiang Plain and the prevention of flood and drought disaster.

The multi-year mean precipitation in the Sanjiang Plain is approximately 520 mm, and the temperate humid and semi-humid continental monsoon climate causes an uneven annual precipitation distribution across the Sanjiang Plain, with significant inter-annual variations. The precipitation during the rainy season represents more than 70% of the annual precipitation, whereas it is relatively low in the winter and spring seasons. The precipitation is relatively intense from June to August, and flood disasters are severe, which significantly limits the development of the regional economy and agriculture industry. In addition, because rice has been planted for many years, the groundwater resources have been overexploited, resulting in a series of immeasurable loss phenomena, such as land subsidence, water inflow attenuation and pump suspension at well locations. These phenomena severely threaten the sustainable utilization of water resources and sustainable socioeconomic development in the Sanjiang Plain. According to the Integrated Scheme for the Comprehensive Coordinated Reform of Modern Agriculture and Finances in the ‘Two Great Plains’ in Heilongjiang Province, which was approved by the State Council of the People's Republic of China in August 2013, the Sanjiang Plain will be the core region of the commodity grain production bases in China over the next three years. Therefore, studying the temporal-spatial patterns of annual precipitation changes in the Sanjiang Plain caused by climate change conditions has important theoretical and practical significance for solving many of the water resource problems, such as the effective utilization of rainwater resources and the scientific and reasonable development of surface water resources.

The ArcGIS Platform was developed by the Environmental Systems Research Institute (ESRI) and has powerful functions, such as inputting, editing, processing, searching, analyzing, mapping and outputting data. The ArcGIS Platform is powerful software with relatively advanced technical concepts compared with the related general geographic information system (GIS) software products. The ArcGIS Geostatistical Analyst Module that is included in the ArcGIS software is an extension of the ArcGIS desktop tool. The Geostatistical Analyst Module provides various tools for such tasks as data exploration, identification of outliers, evaluation of the uncertainty of prediction and the generation of data surfaces. This module is primarily used to study data variability, examine overall data variation trends, search for unreasonable data, calculate probabilities that are greater than a certain threshold value, and prepare Q-Q plots. This module can realize such functions as spatial data pre-processing, contour analysis, geo-statistical analysis and final processing (Clarkson & Bellas 2014; Tayyebi et al. 2014). Currently, the ArcGIS spatial analysis technique has been widely used in such fields as remote sensing, land and resource management, meteorology, water conservancy and disease control (Chen et al. 2014; Khormi & Kumar 2014; Valiakos et al. 2014; Zulu et al. 2014).

The data used in the present study, which included the annual precipitation and mean temperature data collected at seven national weather stations, were obtained from the China Meteorological Data Sharing Service System (http://www.cma.gov.cn/2011qxfw/2011qsjgx/). Because of partial data loss, the length of the data series collected at the Hegang Station differed from the lengths of data from the other six stations. The seven selected weather stations covered the majority of the Sanjiang Plain area. Figure 1 shows the locations of the weather stations, and Table 1 presents detailed information on each weather station and the length of the collected data series at each station. To avoid data errors related to observations and recording, all data were subjected to the graphical method (Cluis 1983) to test consistency and the Von Neumann Ratio method (Bartels 2012) to check at the 95% confidence level, and the results met the analysis requirements.

The annual mean temperature data collected at the seven weather stations in the Sanjiang Plain were used to analyze the climate change in the Sanjiang Plain over the past 50 years. Table 2 presents the detailed results.

Table 2 shows that the annual mean temperatures at all of the weather stations exhibited an increasing trend. Additionally, the correlation coefficients of the trend lines fitted with time were relatively high and all passed the significance test with a confidence level of (⁠⁠). Thus, the Sanjiang Plain exhibited a significant warming trend over the past 50 years. Climate warming in Baoqing was the most significant, with a temperature increase of 0.36 °C/10 years on average, whereas the warming in Fujin was the least significant, with a temperature increase of 0.19 °C/10 years on average. Since the 1960s, the temperature of the Sanjiang Plain has increased by 1.35 °C, which can be explained by the effect of global warming and large-scale agricultural development in the Sanjiang Plain over this time period.

The variation of the UF curve of each station showed that the UF curves for the annual precipitation at the seven stations were all between the upper and lower admissible limits of the confidence level ⁠, i.e., ⁠. Therefore, the overall increasing or decreasing trend of the annual precipitation in the Sanjiang Plain was not significant. However, the UF curves for the seven stations were all generally below the 0 level line, i.e. ⁠, indicating that the overall annual precipitation in the Sanjiang Plain has exhibited a slow decreasing trend since the 1960s. In addition, an abrupt change analysis revealed that the UF curves of the Yilan, Jixi and Hegang stations had more intersection points with UB, which were all between the upper and lower admissible limits, indicating that their annual precipitation underwent multiple abrupt changes. Combined with Figure 2, the annual precipitation of these three stations (Yilan, Jixi and Hegang stations) had significant fluctuations near the abrupt change points.

Wetlands help maintain microclimates and regulate precipitation. The Yilan, Jixi and Hegang stations are all situated in the west region of the Sanjiang Plain, which has relatively small wetland areas and experienced relatively severe climate change, resulting in additional abrupt change points of annual precipitation. All of the other stations except Jiamusi station are situated in the east region of the Sanjiang Plain, which has relatively large wetland areas that represent almost 80% of the total wetland area in the Sanjiang Plain; therefore, these areas experienced smaller annual precipitation fluctuations and had fewer abrupt change points.

Figure 4 shows the time scale variations and phase structure of the annual precipitation of each station. The positive wavelet coefficients in the figure correspond to the periods that had above average precipitation, whereas the negative wavelet coefficients correspond to the periods that had below average precipitation, and 0 corresponds to an abrupt change. The distribution intensities of the annual precipitation of each station over various time scales and periods were generally the same. On a relatively large scale, an approximately 20–25-year time scale may occur for the annual precipitation in the Sanjiang Plain. The wavelet coefficients underwent an abundant-dry alternating process that was 2.5 times longer than the entire time domain of the study, i.e., a high precipitation period before 1962, a lower precipitation period from 1970 to 1975, a greater precipitation period from 1989 to 1992, a lower precipitation period from 1998 to 2000 and a greater precipitation period after 2010. At the middle and small scales, the 10–15-year time scale was also prominent; the central scale was approximately 12 years, the positive and negative phases alternated and the wavelet coefficients underwent 6 abundant-dry alternating processes throughout the study period. However, on even smaller time scales (the 2–4-year time scale), although the abundant-dry alternating variation occurred, this pattern was somewhat disordered. For both the large 23-year scale and middle-to-small 12-year scale, the contours of the positive phases were not completely closed; therefore, the greater precipitation period is estimated to continue after 2013.

To further examine the spatial distribution characteristics of the annual precipitation in the Sanjiang Plain, a statistical analysis was performed. The coefficient of variation, C_v, and the coefficient of skewness, C_s, of each weather station were calculated. Additionally, the spatial analysis function of the ArcGIS software was used to identify the spatial distribution patterns of C_v and C_s.

In the present study, wavelet theory, the M-K test and ArcGIS spatial analysis theory were used to analyze the annual precipitation and mean temperature data, which were collected from 1956 to 2013, at seven national weather stations in the Sanjiang Plain to determine the temporal-spatial patterns of annual precipitation changes caused by climate change conditions. The main conclusions are as follows.

1. Under the condition of global warming, the climate in the Sanjiang Plain exhibited a significant warming trend over the past 50 years. The temperature increased by an average of 0.27 °C/10 years, and the temperature has increased by 1.35 °C since the 1960s. Climate warming at Baoqing Station was the greatest, with average temperature increases of 0.36 °C/10 years. Climate warming in Fujin Station was the lowest, with average temperature increases of 0.19 °C/10 years. This warming trend can improve the lack of heat and prolong the growing seasons for crops in the Sanjiang Plain. Meanwhile, it will improve the single planting pattern and promote the adjustment of planting structures in the Sanjiang Plain.

2. The annual precipitation at each station in the Sanjiang Plain exhibited certain trends. The annual precipitation of the Fujin, Yilan and Jixi stations exhibited a slight decreasing trend of 5.16 mm/10 years on average. The annual precipitation of the Baoqing and Hegang stations exhibited a relatively significant decreasing trend of 17.34 mm/10 years on average. The annual precipitation of the Jiamusi and Hulin stations exhibited a slight increasing trend of 5.6 mm/10 years on average. However, none of the aforementioned trends were significant. In addition, an abrupt change analysis showed that the Yilan, Jixi and Hegang stations had additional annual precipitation abrupt changes due to the small wetland area. Therefore, to adjust climate, wetland protection should be strengthened to avoid unbalanced agriculture development due to climate differences.

3. The annual precipitation in the Sanjiang Plain had 12-year and 23-year oscillation periods. After 2013, the annual precipitation is estimated to continue to increase. The annual precipitation also had a larger main period at an annual scale, but the specific period was not shown because of limitations of the data series lengths used in the present study. Therefore, to avoid grain reduction, we will further intensify efforts to prevent farmland floods and develop rainfall resources to improve the utilization efficiency of water resources.

4. The spatial distribution of the mean annual precipitation in the Sanjiang Plain varied for different years. The multi-year mean annual precipitation was relatively even and exhibited a spatial distribution trend that increased, decreased and then increased again in succession from south to north. In addition, the inter-annual variation of the annual precipitation in the center of the Sanjiang Plain was larger and had a positive skew, whereas the variation was smaller and had a negative skew in the south of the Sanjiang Plain. There was no significant difference in the C_v of the annual precipitation, which fluctuated between 0.20 and 0.24; the C_s value fluctuated between −0.1 and 0.9. The inter-annual fluctuation of annual precipitation was smaller and had a normal distribution. Therefore, for the center of Sanjiang Plain, we should build a reservoir to adjust the rainfall resources of different years to avoid floods or droughts due to too much or too little rainfall in some years. For the north and south of Sanjiang Plain, we should improve flood prevention during the rainy season.

The authors thank the National Natural Science Foundation of China (No. 51179032, 51279031); the Ministry of Water Resources’ Special Funds for Scientific Research on Public Causes (No. 201301096); the Province Natural Science Foundation of Heilongjiang (No. E201241); the New Century Talent Supporting Project by Education Ministry; the Yangtze River Scholars Support Program of Colleges and Universities in Heilongjiang Province; the Heilongjiang Province Water Conservancy Science and Technology project (No. 201318); and the Prominent Young Person of Heilongjiang Province (JC201402).