Yangtze Delta floods and droughts of the last millennium: Abrupt changes and long term memory

Summary Climate variability and flood events in the Yangtze Delta, which is a low-lying terrain prone to flood hazards, storm tides and typhoons, are studied in terms of a trend and detrended fluctuation analysis of historical records. The data used in this paper were extracted from historical records such as local annuals and chronologies from 1000–1950 and supplemented by instrumental observations since 1950. The historical data includes frequencies of floods, droughts and maritime events on a decadal basis. Flood magnitudes increase during the transition from the medieval warm interval into the early Little Ice Age. Fluctuating climate changes of the Little Ice Age, which are characterised by arid climate events, are followed by wet and cold climate conditions with frequent flood hazards. For trend analysis, the Mann-Kendall test is applied to determine the changing trends of flood and drought frequency. Flood frequency during 1000–1950 shows a negative trend before 1600 A.D. and a positive trend thereafter; drought frequency increases after 1300. The detrended fluctuation analysis of the flood and drought frequencies reveals power law scaling up to centuries; this is related to long-term memory and is similar to the river Nile floods.     

Trends in flood seasonality of the River Ouse (Northern England) from archive and instrumental sources since AD 1600

The last decade has witnessed an increase in the application of historical records (historical and documentary) in developing a more complete understanding of high-magnitude flood frequency; but little consideration has been given to the additional information that documentary accounts contain, particularly relating to flood seasonality. This paper examines the methods and approaches available in long-term flood seasonality analysis and applies them to the River Ouse (Yorkshire) in Northern England since AD 1600. A detailed historical flood record is available for the City of York consisting of annual maxima flood levels since AD 1877, with documentary accounts prior to this. A detailed analysis of long-term flood seasonality requires confidence in the accuracy and completeness of flood records; as a result the augmented flood series are analysed using three strategies: firstly, considering all recorded floods since AD 1600; secondly, through detailed analysis of the more complete record since AD 1800; and finally, applying a threshold to focus on high-magnitude flood events since AD 1800. The results identify later winter flooding, particularly in the second half of the twentieth century, with a notable reduction in summer flood events at York during the twentieth century compared to previous centuries. Flood generating mechanisms vary little between the periods considered, with a general pattern of stability in the ratio of floods incorporating a snowmelt component.     

基于降水数据的云南省近61年旱涝特性研究

选择标准化降水指数(SPI)刻画旱涝特征,基于云南省1954—2014年间32个气象站点逐月降水量资料,采用经验正交函数(EOF)方法、径向基函数(RBF)空间插值法、小波分析法,分析了近61年来云南省SPI序列、旱涝情态的时间特征和空间格局。结果表明:近61年来,云南省整体呈现干旱趋势(SPI变率为-0.009 1),SPI序列在2000年之后变化更加剧烈、速率加快。旱涝等级时间序列中,偏涝至偏旱年份占88.52%,大旱占4.92%,重旱和大涝均占3.28%,且自2003年以后,发生干旱的次数和强度明显增加。旱灾易发区主要分布于2个片区:景洪-思茅-元江站片区,以及沾益站东北地区;洪涝易发区主要分布于3个片区:临沧-大理-华坪沿线的西部片区,昭通站北部区域,及云南省东南部片区。干旱与洪涝事件发生的频率具有较好的对应性,但干旱事件发生的频率要略高于洪涝事件。EOF分析的第一个模态表明云南省整体呈现一致性的变涝或变旱特征,可能受到大尺度气候特征影响,第二模态可能受到地形因素的控制,第三模态可能与季风、大气环流等多种因素的影响有关;相应的时间系数也印证了云南省整体具有干旱趋势。SPI序列存在准2 a、准6 a、准8 a、准18 a、准28 a的周期性特征,且以准28 a为主周期。      

Historic variation of warm-season rainfall,Southern Colorado Plateau,Southwestern U.S.A.

Rainfall during the warm season (June 15–October 15) is the most important of the year in terms of flood generation and erosion in rivers of the southern Colorado Plateau. Fluvial erosion of the plateau decreased substantially in the 1930s to early 1940s, although the cause of this change has not been linked to variation of warm-season rainfall. This study shows that a decrease of warmseason rainfall frequency was coincident with and probably caused the decreased erosion by reducing the probability of large floods. Warm-season rainfall results from isolated thunderstorms associated with the Southwestern monsoon and from dissipating tropical cyclones and (or) cutoff low-pressure systems that produce widespread, general rainfall. Warm-season rainfall is typically normal to above normal during warm El Niño-Southern Oscillation (ENSO) conditions. A network of 24 long-term precipitation gages was used to develop an index of standardized rainfall anomalies for the southern Colorado Plateau for the period 1900–85. The index shows that the occurrence of anomalously dry years increased and the occurrence of anomalously wet years decreased after the early 1930s, although 1939–41, 1972, and 1980–84 were anomalously wet. The decrease in warm-season rainfall after the early 1930s is related to a decrease in rainfall from dissipating tropical cyclones, shifts in the incidence of meridional circulation in the upper atmosphere, and variability of ENSO conditions.     

Magnitude and Frequency of Flooding in the Tagus Basin (Central Spain) over the Last Millennium

A relational database using Access, and an interface based on a Geographical Information System (GIS) with ArcView program, were created to allow spatial-temporal analysis of documentary flood data collected for the Tagus basin (Central Spain). High flood frequencies were registered in the periods: AD 1160–1210 (3%), 1540–1640 (11%; peak at 1590–1610), 1730–1760 (5%), 1780–1810 (4%), 1870–1900 (19%), 1930–1950 (17%) and 1960–1980 (12%). Flood magnitudes of those documentary events were estimated at four sites (Aranjuez, Toledo, Talavera and Alcántara) along the Tagus River using HEC-RAS Computer program, matching the calculated water surface profiles with historical references of flood stage. The largest floods took place during the periods AD 1168–1211 (Toledo and Talavera), AD 1658–1706 (Talavera), AD 1870–1900, and AD 1930–1950 (Aranjuez, Toledo, Talavera and Alcántara). The temporal distribution of flood magnitude and frequency is discussed within the context of climatic variability experienced by the Iberian Peninsula. Although flood producing mechanisms of Atlantic and Mediterranean basins of the Iberian Peninsula are related to different, independent atmospheric patterns, there is a clear coincidence between most periods showing high flood frequencies. These periods of high flooding seem to correspond to the initial and final decades of periods with climate deterioration described at the continental scale. This suggests that climatic variability over the last millennium has induced a response in hydrological extremes (positive or negative), irrespective of the flood-producing mechanism.     

近40年长江中游地区旱涝特点分析

对Z指数确定的旱涝等级及各站月降水距平百分率的标准差分析表明:长江中游地区大范围的旱涝主要集中在盛夏和秋季。不同季节, 旱涝的主要分布区域不同, 但洞庭湖区始终是一个旱涝多发地。对1960年1月—2001年12月各月长江中游地区50个代表站的旱涝等级进行经验正交函数分析, 并在此基础上进行多通道奇异谱分析, 得到长江中游地区的旱涝分布主要存在4种空间型, 且具有明显的年代际和年际尺度变化, 在20世纪80年代以前, 长江中游地区以偏旱型为主, 而到80年代以后则转变以偏涝型为主; 另外, 长江中游地区的旱涝具有准8年、准5年、准2年和准4年的周期振荡性质, 但从80年代初开始, 旱涝的准2年周期特征在长江中游地区表现不再明显。     

气侯变化背景下黑河上游春季融雪洪水预估研究

气候变暖将导致高山区冰冻圈加剧融化,一方面融水资源时空分布的不确定性增大;另一方面,融水洪水灾害发生的频度和强度也将发生改变。基于气象、水文数据和MODIS积雪覆盖数据,利用融雪径流模型(SRM),对1990—2012年共23年祁连山黑河札马什克控制区融雪期径流进行模拟与验证。结果表明：SRM在该流域具有较高的模拟精度（纳什系数为0.91),可用于分析和预估控制区径流强度变化。为此,采用黑河流域气温、降水降尺度数据,预估了未来气候变化背景下积雪范围变化及不同重现期洪水变化趋势。结果显示,与基准期相比,在RCP2.6、RCP4.5和RCP8.5情景下,最大积雪范围可减小3%~7%,且随着海拔升高,变化愈剧烈。RCP2.6情景下因气温和降水变化幅度较小,到21世纪末各重现期洪水强度保持在10%以内波动;RCP4.5情景下,各重现期洪水强度最高增大约20%;在RCP8.5情景下,各重现期洪水强度最高可增大超30%。相关分析结果显示,不同重现期洪水径流与气温和降水均具有较强相关性：重现期越长,洪峰与气温的相关性越大;重现期越短,洪峰与降水的相关性越大。通过预估气候变化背景下的融雪性洪水事件强度及重现期变化,有助于有效开展区域洪水风险管理、提高洪水资源的利用价值。     

Three centuries of change in the Peace–Athabasca Delta, Canada

The Peace–Athabasca Delta in northern Alberta, Canada, is a dynamic wetland ecosystem. Climatic, hydrologic, biological, and historical data are synthesized to elucidate how the ecosystem has changed over the past 300 years. Annual temperature is now higher than it has been in the past 300 years. For much of the 1700s, the Delta was colder in winter and had a lower flood frequency than that of the last 30 years. The 1800s were characterized by long and cold winters, 4–12 year-long episodes of high or low water, and repeated human epidemics. The early twentieth century was relatively moist and cool. Since mid-twentieth century the Delta has experienced periods of both intense warmth and cold, desiccation and recharge. Since the mid-1960s, local and regional mean annual temperatures have increased 0.30°C to 0.48°C per decade while winter temperatures have increased 0.68°C to 0.92°C per decade; annual snowfall has decreased 12 to 41 cm per decade while winter snowfall has decreased 12 to 34 cm per decade. Major events in the past 45 years include climatic changes favoring a warmer, drier ecosystem; cultural and socioeconomic changes; building of the Bennett Dam; prevention of the Athabasca River mainstem avulsion in 1972; the Cree Creek avulsion of 1982; large fluctuations in water, vegetation, and wildlife; and the development of the Alberta Tar Sands. Increased rates of basin desiccation and wildfire activity and upstream land disturbances may combine to alter the Delta’s biotic composition. There appears to be no relevant historical analogue of the present Delta.     

山洪灾损预估方法研究

为了预估四川省的山洪灾害造成的人口和经济影响,本文基于四川省数字高程模型数据(DEM),依据山洪沟的判别标准,提取出全省范围的山洪沟,并结合坡度与降水耦合的关系建立了四川山洪危险性评价模型。以2011年1月~2013年12月发生的21次山洪灾害为样本,利用2013年四川省人口和经济数据,建立了山洪灾害对人口和经济影响的预估模型。结论如下:(1)提取出影响生命财产安全的山洪沟64346条,其主要分布在盆周山区、川西高原和攀西地区,最大影响面积约38万km2,约占全省面积的78%;(2)建立了山洪危险性评估模型,评估结果能客观地预警四川范围内的山洪等级和影响范围;(3)建立了山洪灾害影响预估模型,模型对人口的影响预估结果比较可靠。      

On the link between extreme floods and excess monsoon epochs in South Asia

This paper provides a synoptic view of extreme monsoon floods on all the nine large rivers of South Asia and their association with the excess (above-normal) monsoon rainfall periods. Annual maximum flood series for 18 gauging stations spread over four countries (India, Pakistan, Bangladesh and Nepal) and long-term monsoon rainfall data were analyzed to ascertain whether the extreme floods were clustered in time and whether they coincided with multi-decade excess monsoon rainfall epochs at the basin level. Simple techniques, such as the Cramer’s t-test, regression and Mann–Kendall (MK) tests and Hurst method were used to evaluate the trends and patterns of the flood and rainfall series. MK test reveals absence of any long-term tendency in all the series. However, the Cramer’s t test and Hurst-Mandelbrot rescaled range statistic provide evidence that both rainfall and flood time series are persistent. Using the Cramer’s t-test the excess monsoon epochs for each basin were identified. The excess monsoon periods for different basins were found to be highly asynchronous with respect to duration as well as the beginning and end. Three main conclusions readily emerge from the analyses. Extreme floods (>90th percentile) in South Asia show a tendency to cluster in time. About three-fourth of the extreme floods have occurred during the excess monsoon periods between ~1840 and 2000 AD, implying a noteworthy link between the two. The frequency of large floods was higher during the post-1940 period in general and during three decades (1940s, 1950s and 1980s) in particular.     

Diabatic heating and the low frequency dynamics in the tropics

Summary Using a General Circulation Model developed at FSU (FSUGCM), the role of the diabatic heating on the 30–60-day oscillation is investigated. To concentrate on the radiation and the moist convection processes, an aqua planet model is employed in this paper. We have obtained a 40-day oscillation with relatively lower frequency than other GCMs without strong heating in the lower troposphere. Unlike some GCMs and simple models, the convective area does not move eastward along with the oscillation. Adiabatic cooling due to the upward motion is mostly compensated by diabatic heating. This implies that Kelvin CISK theory might not explain our 40-day oscillation. We have also examined the impact of radiative heating on the low frequency oscillation. When we reduce the radiative cooling rate, our 40-day mode does not appear and a Kelvin CISK mode appears with a faster phase speed. The impact of the different convection schemes is also investigated. With an enhanced convection scheme, zonal wave number two with a 40-day period is generated.With 12 FiguresOn leave from Japan Meteorological Agency.     

1959-2008年广西西江流域洪涝气候特征

 利用1959-2008年西江流域气象测站降水资料、西江干流及其支流年最高水位资料,建立逐年洪涝发生站次序列,并分析其时空特征和演变规律。结果表明：西江干流发生洪涝的频率比各支流大,各支流的中下游发生洪涝频率比上游大;西江流域洪涝的发生站次具有较显著的阶段性和突变性特征,主要突变发生在1967年和1993年;近16 a洪涝发生频率显著偏多,流域性大洪涝基本集中在这一时期。了解西江流域洪涝变化的规律,有利于提高洪涝灾害评估和预测水平,为防灾减灾提供科学依据。     

区域性地震过程对中国大陆洪水影响之初探

为探讨地震、洪水（暴雨）间可能存在的链状关系,以1979--1990年为例,对中国及周围区域肘≥7地震的活动格局与中国大陆洪涝灾害的时空分布特征进行了分析研究。结果发现：类似于2005年西江洪水,不少大洪水（暴雨）也位于地震迁移线上或其附近。洪涝的发生时间有可能与震链的后震或前震比较接近,或与震链推移之进程基本吻合,反映了链程中的场态效应。震链之进程往往构成三角之类的几何体或兼具振荡特征,所围限区域在气象条件允许时,有可能发生强降雨;中国大陆若完全在“几何体”,可能当年洪涝灾害相对较弱。洪涝（暴雨）并非都与地震有关。某些震链“几何体”对降雨的影响可能与季风有关。强震对降雨往往有时空近程影响。震群或震链多线程指向点、端点、交会点、角夹持地带可能加剧洪涝。震链回迁或转折处往往是不同体系、单元、性质构造的复合点,震源应力变异点,形态与活动方式复杂。     

郑州市近58a降雨量变化分析

采用水文序列中常用的谱分析、有序聚类分析、Kendall秩次相关检验和Man-Kendall检验等方法,分别对郑州市1951-2008年的年降雨量、汛期和非汛期降雨量序列进行周期性、突变性和趋势性检验分析,结果显示:总体上郑州市年降雨量保持平稳,但各年汛期和非汛期的降雨量分别呈现出增加和减少的趋势,而且非汛期降雨量减少的趋势性较为显著;年降雨量和汛期降雨量有较为显著的近似10 a的周期,而非汛期降雨量表现为不显著的3 a的周期;年降雨量、汛期和非汛期降雨量序列的突变点均在20世纪70年代末和80年代初,但并不显著。     

1997年华南汛期降水异常与大气低频振荡的关系

利用国家气象中心提供的逐日降水资料及NCEP逐日再分析资料,分析了1997年华南地区汛期异常降水低频特征与大气低频振荡的关系。研究表明,1997年华南前汛期和后汛期降水表现为不同的振荡特征;前汛期降水主要以10～20天准双周振荡为主,而后汛期降水的低频特征并不明显。进一步对降水和其它要素的低频振荡特征进行分析发现,该年华南地区前汛期降水和风场的低频振荡现象是普遍存在的;低频纬向风的传播变化与降水的时间分布有较好的对应。并且,高、低纬度低频风场同时向华南地区传输,会产生极强的降水。在对大气低频扰动动能的分析中也发现,华南前汛期降水伴随着低频扰动动能在该地区的集中释放。     

Spatial and temporal variability of winter and summer precipitation over Serbia and Montenegro

Summary The main characteristics of the spatial and temporal variability of winter and summer precipitation observed at 30 stations in Serbia and Montenegro were analysed for the period 1951–2000. The rainfall series were examined spatially by means of Empirical Orthogonal Functions (EOF) and temporally by means of the Mann-Kendall test and spectral analysis. The Alexandersson test was used to detect the inhomogeneity of the data set.The EOF analysis gave three winter and summer dominant modes of variations, which explained 89.7% and 70.4% of the variance, respectively. The time series associated with the first pattern showed a decreasing trend in winter precipitation. The spectral analysis showed a 16-year oscillation for the dominant winter pattern, around a 3-year oscillation for the dominant summer pattern, and a quasi-cycle of 2.5 years for the winter third pattern.     

North-Eastward propagating 25–50-day mode and its link with Indian summer monsoon activity

Summary Spatio-temporal characteristics of the 25–50-day oscillations are investigated using the empirical orthogonal function (EOF) decomposition and spectral analysis with the Maximum Entropy Method (MEM). Daily pressure values over India during 1978 are used in this study. Power spectra of the temporal coefficients of the two leading EOFs show the existence of a low frequency oscillation with a period range 25–50-day over all India. An analysis using extended empirical orthogonal function (EEOF) indicates North-Eastward propagating 25–50-day mode. The analysis EOF has allowed to establish a relationship between the 25–50-day oscillation and the activity of the summer monsoon. The North-Eastward propagation of this mode is also confirmed by the simple EOF analysis.With 9 Figures     

DIAGNOSTIC ANALYSIS OF PERSISTENT DROUGHT/FLOOD EVENTS IN SUMMER OVER THE TWO-LAKE REGION OF CHINA

Based on the daily regional mean rainfall,the Z-index method is used to identify persistent flood and drought events lasting for at least 10 days over a region where Dongting Lake and Poyang Lake sit(referred to as the"two-lake region"hereafter).The National Centers for Environmental Prediction(NCEP)reanalysis data are then utilized to perform a preliminary diagnostic analysis on these events.The results indicate that the composite standardized geopotential height at 500 hPa presents two different meridional wave trains from north to south over the East Asian-Pacific region,i.e.,a"-+-"pattern for the droughts and a"+-+"pattern for the floods,respectively.The developing,maintaining and decaying phases in the drought and flood events are closely related to the intensity and location of a subtropical high and an extra-tropical blocking high.It is shown that the East Asian summer monsoon is strong(weak)with the occurrence of persistent drought(flood)events.Droughts(floods)are accompanied by a weak(strong)tropical convergent system and a strong(weak)subtropical convergent system.Furthermore,the persistent drought(flood)events are associated with a divergence(convergence)of vertically integrated water vapor flux.In the vertical profile of water vapor flux,divergence(convergence)in the mid-and lower-levels and convergence(divergence)in the higher levels are evident in the droughts(floods).Both the divergence in the droughts and the convergence in floods are strongest at 850 hPa.     

Projections of climate change impacts on river flood conditions in Germany by combining three different RCMs with a regional eco-hydrological model

A general increase in precipitation has been observed in Germany in the last century, and potential changes in flood generation and intensity are now at the focus of interest. The aim of the paper is twofold: a) to project the future flood conditions in Germany accounting for various river regimes (from pluvial to nival-pluvial regimes) and under different climate scenarios (the high, A2, low, B1, and medium, A1B, emission scenarios) and b) to investigate sources of uncertainty generated by climate input data and regional climate models. Data of two dynamical Regional Climate Models (RCMs), REMO (REgional Model) and CCLM (Cosmo-Climate Local Model), and one statistical-empirical RCM, Wettreg (Wetterlagenbasierte Regionalisierungsmethode: weather-type based regionalization method), were applied to drive the eco-hydrological model SWIM (Soil and Water Integrated Model), which was previously validated for 15 gauges in Germany. At most of the gauges, the 95 and 99 percentiles of the simulated discharge using SWIM with observed climate data had a good agreement with the observed discharge for 1961–2000 (deviation within ±10 %). However, the simulated discharge had a bias when using RCM climate as input for the same period. Generalized Extreme Value (GEV) distributions were fitted to the annual maximum series of river runoff for each realization for the control and scenario periods, and the changes in flood generation over the whole simulation time were analyzed. The 50-year flood values estimated for two scenario periods (2021–2060, 2061–2100) were compared to the ones derived from the control period using the same climate models. The results driven by the statistical-empirical model show a declining trend in the flood level for most rivers, and under all climate scenarios. The simulations driven by dynamical models give various change directions depending on region, scenario and time period. The uncertainty in estimating high flows and, in particular, extreme floods remains high, due to differences in regional climate models, emission scenarios and multi-realizations generated by RCMs.     

RECOGNISING THE UNCERTAINTY IN THE QUANTIFICATION OF THE EFFECTS OF CLIMATE CHANGE ON HYDROLOGICAL RESPONSE

A widely used method of evaluating effects of climate change on flow regime is to perturb the climate inputs to a rainfall–runoff model and examine the effect on a statistic of the modelled flows. Such studies require four elements: a method of perturbing the climate, a rainfall–runoff model, a study catchment and a flow index. In practice the direction and magnitude of the estimated effects depend on each of the four elements, leading to concern over the usefulness and generality of the results. To investigate these uncertainties two climate scenarios and eight climate sensitivity tests have been applied to three UK catchments using two conceptual rainfall–runoff models in order to quantify effects of climate change on three flow indices representing mean runoff, flood magnitudes and low flows. The sensitivity tests were found to be useful to assess the suitability of the models to simulate flows outside the conditions experienced in their calibration. Both models gave internally consistent results but, on close examination, one model was found inappropriate for this application. Results show that the effect of climate change on flow varies between catchments and that different flow response indices can change in opposite directions, e.g. floods increased in magnitude while low flows reduced. Contrasting results were obtained from the two climate scenarios.