Study On Variation Of Runoff Along With Its Prediction And Interpolation In The Yellow River Basin Under The Global Change Background

Climate change is the heart problem of global change. Scientific research shows that over the past century, the Earth is undergoing a significant change to global warming as the main feature. Existing prediction results indicate that the next 50 to 100 years of global climate warming will continue to develop in the direction. Climate change has already produced a certain amount of influence to Hydrology and Water Resources in China over the past few decades. Measured runoff of most of the six major rivers showed a downward trend in the past 50 years. Although human activities, population growth and socio-economic development causes an increase in water consumption is declining runoff reasons, but in recent years the impact of rising temperatures and precipitation changes on river runoff can’t be ignored. So, we will probe into the variation and prediction of the runoff in Yellow River Basin under the background of global change in this paper. It will contribute to better planning of water resources. The work has important theoretical and practical significance.Based on measured runoff data and meteorological data in the Yellow River basin, we use a variety of methods, such as mathematical statistics, linear trend estimate, moving average, total anomaly, MK test, wavelet analysis et al, to reveal variation characteristics of runoff on different time scales. Besides, the response of runoff to climate factors will also be studied to analyze the contribution value of climate factors quantitatively. Finally, we will study on prediction of interpolation of runoff based on improved theory of optimal window width.(1) The Yellow River lies in the arid and semi-arid areas, which leads to its water systems are very sensitive to climate change. During the past 60 years, there was clear downward trend of runoff, and research showed that was a mutation phenomenon. The Yellow River Basin was in the wet season in 1950s and 1960s and transferred into normal water period in 1970s and 1980s. In 1990s, it entered into dry season. This stage was still in the dry season, but there were signs of transferring from dry to the abundance in the 21st century. The runoff showed significant inter-annual oscillations. The closer the downstream, the more significant inter-annual oscillations were. Most of dry years appear in the 1990s when it was the global warming period. There was a higher probability of occurrence of dry years than that of normal water periods and wet years. Consecutive dry years were up to 21 years, and Continuous wet years were only 8 years. The change of runoff during the year was quite dramatic. Runoff mainly concentrated in the flood season. Distribution of runoff during the year was very uneven. The closer the downstream, the more uneven the distribution was. Runoff could lag up to eleven months and the summer half year was in higher stability.(2) The temperature of the Yellow River Basin showed a significant upward trend, the warming rate was 0.29℃/10a, and winter warming was the most obvious, reaching 0.45℃/10a. In other word, warm winter phenomenon was obvious. There was a sudden change of temperature in the late 1980s. Precipitation had a weak downward trend, and exhibited periodic characteristics. Precipitation was abundant in 1950s,1960s and 1970s then decreased in 1980s and 1990s. In the 21st century, there was a growing trend in precipitation. This inter-annual variability of rainfall and change characteristics of runoff were basically consistent. Precipitation and runoff had positive correlation, and there was a negative correlation between temperature and runoff. With the appearance of warming and wetting phenomenon, runoff tended to increase in the 21st century.(3) The correlation between runoff and climatic factors was not linear but non-linear. Overall, runoff was more sensitive to changes in precipitation than temperature. Precipitation and temperature both contributed to the variation of runoff, but the former was a positive contribution and the latter was opposite through quantitative analysis. And they were more pronounced in summer and autumn, relatively weak in winter and spring.(4) Based on information diffusion theory, we improved the optimal window width theory in order to conduct prediction of interpolation of runoff of Lijin Hydrological station. The forecast results could well simulate the waveform changes of the actual runoff sequence, especially accurate in wet years (as 2007) and dry years (as 2009). The average relative error of Long-term forecasts (10a) was only 11.59%, which had a greater improvement compared with traditional methods. We also proved that this algorithm was applicable to the estimates and forecasts of annual runoff of hydrological stations which had different geographical factors, underlying surface, drainage area and so on.