| The continuous ecological restoration on the Loess Plateau over past 60 years is known for two strategies: the integrated soil conservation project and the “Grain for Green” project. Climate change and intense human impacts resulted in rainfall–streamflow generation–sediment production in a basin. Understanding and quantifying the impacts of land use/cover change and climate variability on hydrological responses are important to the design of water resources and land use management strategies for adaptation to climate change, especially in water-limited areas. The scaling behavior was also characterized to understand the physical processes of streamflow and sediment load, and their scaling characteristics were also used for extreme precipitation/streamflow event estimation and causal simulation. Streamflow from 32 hydrological stations and sediment from 15 stations(9 catchments) in the middle reaches of the Yellow River(MRYR) were analyzed.(1) Thirty-two hydrological station were selected to investigate streamflow regime variations during the two project periods. Mean annual streamflow varied from 16.3 mm to 105.0 mm. Streamflow of multi-time scale were decreased in a spatial scale varying from 100 km2 to more than 105 km2. Annual streamflow significantly decreased in 31 out of 32 catchments, varying from-0.11 to-1.83 mm/a. Annual streamflow was reduced due to high flow decreases, whereas in the low flows increased in southern MRYR during both the integrated soil conservation period(PII) and the “Grain for Green” period(PIII). However, both daily and event streamflow decreased during the two periods in the northern MRYR. The extent of streamflow reduction can be linked to the soil conservation measures and the accumulative area treated. A stable streamflow regime can be achieved in vegetated areas, and streamflow moderation is dependent on ecological management practices.(2) Both sediment load and suspended sediment concentration(SSC) in the 15 stations varied with two order of magnitudes, each varying from 199 ~ 10738 t·km-2·a-1、4.2 ~ 344.2 kg·m-3. The sediment load can be lower to less than soil loss tolerance in vegetated areas, i.e. Huangling station, 199·km-2·a-1. The multi-time scale sediment load and SSC decreased, and the annual sediment load and SSC decreased, varying from-4.2 ~-386.8·t·km2·a-1,-0.08 ~-7.52 kg·m-3, respectively. The multi-time scale streamflow, sediment load and SSC changes showed that the ecological restoration changed the rainfall–streamflow, rainfall–sediment yield and water discharge–SSC dynamics. The sediment yield was decreased by reduction in both streamflow and SSC in catchment scales varying from 100 ~ 104 km2. The changes in SSC decreased the sediment yield by 1008 t·km-2·a-1, 2049 t·km-2·a-1 in the 9 catchments during the two periods, accounting for 28.3%, 25.2% in percentage, respectively. Moreover, the increased annual sediment yield at the Zhangcunyi station exposed a risk of increased erosion in areas where forests have been well preserved.(3) The streamflow changes attributiong to climate variability and human activities were analyzed based on the elasticity method. Based on the fractal idea and water discharge–SSC distribution, the contribution of changes in SSC to sediment changes were calculated, then a method based on the elasticity and water discharge–SSC distribution were proposed to partition the effect of climate variability and human activities on sediment yield changes. Since the ecological restoration have moderated daily streamflow by increasing low flow in the southern MRYR, human activities accounted for less than 50% of streamflow reduction, while accounted for more than 50% in the northern MRYR. Then, human activities contributed 56% ~ 87% in PIII in the whole MRYR. Increased vegetation cover consumed more water by increasing evaportranspiration, runoff coefficient decreased during the whole period. Human activitiy was the main factor for runoff coefficient reducing, accounting for 48% ~ 71%, 62% ~ 82% during the two periods, respectively. Sediment load in the 9 catchments attributing to human activities and climate variability are 2249 t·km-2·a-1, 1313 t·km-2·a-1 in PII, while 4720 t·km-2·a-1, 1508 t·km-2·a-1, in PIII, respectively. Obviously, human activities were the main factor reducing sediment load, accounting for 63.1%, 75.8% in percentage, respectively.(4) Multifractal analysis showed that the daily river flow series were multifractal over a range of scales spanning at least 20 to 214 days. Although no outer limit to the scaling was found for most of the rivers, there is a break in the scaling regime at a period of about 25 which is comparable to the atmospheric synoptic maximum. MFDFA showed that daily streamflow series are long-term correlated, and the persistence of the runoffs is generally caused by storage effects. The fluctuation of streamflow series is predominated by large fluctuation in the MRYR, which is different from those studies in other basins. The multifraclity of runoff is similar to that of precipitation, indicating that the multifractility of precipitation is retained in runoff series. A priori knowledge of multifractility shows that daily streamflow in the MRYR has larger indices, C1, α, and lower H index. These indices indicating that, the main components contributing to the mean of streamflow are floods, catchments in the MRYR have higher extremes, streamflow regime is more variable. The spatial agglomeration ability of a basin can be enhanced by a fully-structured vegetation cover, and hence determine the scaling behavior in synoptic scale(< 32 days). Ecological restoration reduced streamflow, but has not been detected in regulating scaling behavior, since the scaling behavior are formed in a long history and generally stable.(5) Multifractal techniques were applied to calculate the physically meaningful EP threshold(EPT). The multifractal approach considers the physical processes and probability distribution, thereby providing a formal framework to determine the EPT independent of empirical adjustments. This advantageous method can objectively obtain a unique set of EPTs. Results showed that the EPTs ranged from 50 mm/d in the southeastern to 17 mm/d in the northwestern Loess Plateau regions. During the past 53 years, the EP frequency(EPF) spatially varied at 50–111 days. Notable occurrences of EPF were mainly observed in the mid-southern and southeastern portions of the Loess Plateau. Note that 95% of the total EP events in the study period were concentrated from May 21 to September 18. The highest EP intensity(EPI) values mainly occurred in the Central Loess Plateau, whereas the most intense EP severity index(EPSI) occurred in the mid-southeastern Loess Plateau. The spatially lower precipitation, highly concentrated intra-annual EPF(the maximum of which also occurs 11 days earlier than the maximum daily precipitation), the highest EPI, and the relatively higher EPSI are responsible for the most serious soil erosion in the Central Loess Plateau. Approximately 60% of the Loess Plateau area demonstrated a downward trend of annual precipitation but experienced upward trends of one or more EP indices. This evidence showed that the EP was intensified over the Loess Plateau, especially in the mid-southern region.(6) The value and utility of applying multifractal analysis techniques to systematically calculate physically meaningful estimates of probable maximum precipitation(PMP) from observations on the Loess Plateau is assessed in this study. The multifractal approach is advantageous because it provides a formal framework to infer the magnitude of extreme events independent of empirical adjustments. Results showed that the estimated result is little influenced by the length of data series, and is of higher precision than traditional probability distribution function, e.g. GEV distribution. The 1-day and 3-day PMP with a return period of 100 year increasing from west to east, ranges from 80 ~ 310 mm, 90 ~ 350 mm, respectively. There is a PMP plateau in the Central Loess Plateau with a south-north direction. The western boundary of the area is located along a line of Hohhot–Etuoke–Suide–Yan’an–Wuqi–Huanxian–Xifeng–Huaxian, and the east boundary, a line of Hohhot–Wuzai–Jiexiu–Linfen–Luanchuan. The 1-day PMP with a return priod of 100 and 200 year in the PMP plateau ranges from 170 ~ 220 mm, 230 ~ 300 mm.(7) The multifractal properties of precipitation and streamflow(particularly their scaling and intermittency) are best understood in terms of mulplicative cascade processes(anisotropic and scaling). The cascades provide a framework not only for theoretically and empirically investigating these fields, but also for constructing physically based stochastic models. This physical basis is provided by cascade scaling and intermittency, which is of broadly the same sort as that specified by the dynamical(nonlinear, partial differential) equations. Cascade processes have been used for simulating turbulent systems – including clouds, rain, streamflow – as well as for solid earth fields. Realistic simulations are continuous in scale, but suffer from strong ‘‘finite size effects’ ’ i.e. deviations from pure power law scaling, which can take surprisingly large ranges of scale to disappear. The FIF model corrected the simulations for the leading finite size effects for both causal and acausal simulations, and hence improved the precision in small scale simulation. Based on the FIF model and the addictive priniple, the field with different scaling behavior can be simulated efficiently. In the context of hydroligcal regime changes, the model is of great significance for data comparison. |
PDF Full Text Request