Identification of physical processes inherent in artificial neural network rainfall runoff models

The emergence of artificial neural network (ANN) technology has provided many promising results in the field of hydrology and water resources simulation. However, one of the major criticisms of ANN hydrologic models is that they do not consider/explain the underlying physical processes in a watershed, resulting in them being labelled as black‐box models. This paper discusses a research study conducted in order to examine whether or not the physical processes in a watershed are inherent in a trained ANN rainfall‐runoff model. The investigation is based on analysing definite statistical measures of strength of relationship between the disintegrated hidden neuron responses of an ANN model and its input variables, as well as various deterministic components of a conceptual rainfall‐runoff model. The approach is illustrated by presenting a case study for the Kentucky River watershed. The results suggest that the distributed structure of the ANN is able to capture certain physical behaviour of the rainfall‐runoff process. The results demonstrate that the hidden neurons in the ANN rainfall‐runoff model approximate various components of the hydrologic system, such as infiltration, base flow, and delayed and quick surface flow, etc., and represent the rising limb and different portions of the falling limb of a flow hydrograph. Copyright © 2004 John Wiley & Sons, Ltd.     

Runoff responses to afforestation in a watershed of the Loess Plateau,China

Afforestation has been suggested as a means of improving soil and water conservation in north‐western China, especially on the Loess Plateau. Understanding of the hydrological responses to afforestation will help us develop sustainable watershed management strategies. A study was conducted during the period of 1956 to 1980 to evaluate runoff responses to afforestation in a watershed on the Loess Plateau with an area of 1·15 km², using a paired watershed approach. Deciduous trees, including locust (locusta L.), apricot (praecox L.) and elm (ulmus L.), were planted on about 80% of a treated watershed, while a natural grassland watershed remained unchanged. It was estimated that cumulative runoff yield in the treated watershed was reduced by 32% as a result of afforestation. A significant trend was also observed that shows annual runoff reduction increases with the age of the trees planted. Reduction in monthly runoff occurred mainly from June to September, which was ascribed to greater rainfall and utilization by trees during this period. Afforestation also resulted in reduction in the volume and peak flow of storm runoff events in the treated watershed with greater reduction in peak flow. Copyright © 2003 John Wiley & Sons, Ltd.     

The hydrological influence of black locust plantations in the loess area of northwest China

Black locust (Robina pseudoacacia) has become one of the most important shelter species in the loess area of northwest China. This paper summarizes recent research concerning its hydrological influence, including canopy interception, litter absorption capacity, its effect on rainfall kinetic energy, infiltration rates, surface runoff, soil moisture, and evapotranspiration, and its role in soil conservation. Several predictive models are listed. on the basis of existing results, optimum characteristics for an effective plantation are defined, and problems requiring further research are identified.     

Wavelet analysis of inter‐annual variability in the runoff regimes of glacial and nival stream catchments,Bow Lake,Alberta

Continuous wavelet analyses of hourly time series of air temperature, stream discharge, and precipitation are used to compare the seasonal and inter‐annual variability in hydrological regimes of the two principal streams feeding Bow Lake, Banff National Park, Alberta: the glacial stream draining the Wapta Icefields, and the snowmelt‐fed Bow River. The goal is to understand how water sources and flow routing differ between the two catchments. Wavelet spectra and cross‐wavelet spectra were determined for air temperature and discharge from the two streams for summers (June–September) 1997–2000, and for rainfall and discharge for the summers of 1999 and 2000. The diurnal signal of the glacial runoff was orders of magnitude higher in 1998 than in other years, indicating that significant ice exposure and the development of channelized glacial drainage occurred as a result of the 1997–98 El Niño conditions. Early retreat of the snowpack in 1997 and 1998 led to a significant summer‐long input of melt runoff from a small area of ice cover in the Bow River catchment; but such inputs were not apparent in 1999 and 2000, when snow cover was more extensive. Rainfall had a stronger influence on runoff and followed quicker flow paths in the Bow River catchment than in the glacial catchment. Snowpack thickness and catchment size were the primary controls on the phase relationship between temperature and discharge at diurnal time scales. Wavelet analysis is a fast and effective means to characterize runoff, temperature, and precipitation regimes and their interrelationships and inter‐annual variability. The technique is effective at identifying inter‐annual and seasonal changes in the relative contributions of different water sources to runoff, and changes in the time required for routing of diurnal meltwater pulses through a catchment. However, it is less effective at identifying changes/differences in the type of the flow routing (e.g. overland flow versus through flow) between or within catchments. Copyright © 2003 John Wiley & Sons, Ltd.     

Pressures of Crystallization of Icelandic Magmas

Iceland lies astride the Mid-Atlantic Ridge and was createdby seafloor spreading that began about 55 Ma. The crust is anomalouslythick (20–40 km), indicating higher melt productivityin the underlying mantle compared with normal ridge segmentsas a result of the presence of a mantle plume or upwelling centeredbeneath the northwestern edge of the Vatnajökull ice sheet.Seismic and volcanic activity is concentrated in 50 km wideneovolcanic or rift zones, which mark the subaerial Mid-AtlanticRidge, and in three flank zones. Geodetic and geophysical studiesprovide evidence for magma chambers located over a range ofdepths (1·5–21 km) in the crust, with shallow magmachambers beneath some volcanic centers (Katla, Grimsvötn,Eyjafjallajökull), and both shallow and deep chambers beneathothers (e.g. Krafla and Askja). We have compiled analyses ofbasalt glass with geochemical characteristics indicating crystallizationof ol–plag–cpx from 28 volcanic centers in the Western,Northern and Eastern rift zones as well as from the SouthernFlank Zone. Pressures of crystallization were calculated forthese glasses, and confirm that Icelandic magmas crystallizeover a wide range of pressures (0·001 to 1 GPa), equivalentto depths of 0–35 km. This range partly reflects crystallizationof melts en route to the surface, probably in dikes and conduits,after they leave intracrustal chambers. We find no evidencefor a shallow chamber beneath Katla, which probably indicatesthat the shallow chamber identified in other studies containssilica-rich magma rather than basalt. There is reasonably goodcorrelation between the depths of deep chambers (> 17 km)and geophysical estimates of Moho depth, indicating that magmaponds at the crust–mantle boundary. Shallow chambers (<7·1 km) are located in the upper crust, and probablyform at a level of neutral buoyancy. There are also discretechambers at intermediate depths (11 km beneath the rift zones),and there is strong evidence for cooling and crystallizing magmabodies or pockets throughout the middle and lower crust thatmight resemble a crystal mush. The results suggest that themiddle and lower crust is relatively hot and porous. It is suggestedthat crustal accretion occurs over a range of depths similarto those in recent models for accretionary processes at mid-oceanridges. The presence of multiple stacked chambers and hot, porouscrust suggests that magma evolution is complex and involvespolybaric crystallization, magma mixing, and assimilation. KEY WORDS: Iceland rift zones; cotectic crystallization; pressure; depth; magma chamber; volcanic glass     

Spatial Variability of Rock Depth in Bangalore Using Geostatistical, Neural Network and Support Vector Machine Models

Geospatial technology is increasing in demand for many applications in geosciences. Spatial variability of the bed/hard rock is vital for many applications in geotechnical and earthquake engineering problems such as design of deep foundations, site amplification, ground response studies, liquefaction, microzonation etc. In this paper, reduced level of rock at Bangalore, India is arrived from the 652 boreholes data in the area covering 220 km². In the context of prediction of reduced level of rock in the subsurface of Bangalore and to study the spatial variability of the rock depth, Geostatistical model based on Ordinary Kriging technique, Artificial Neural Network (ANN) and Support Vector Machine (SVM) models have been developed. In Ordinary Kriging, the knowledge of the semi-variogram of the reduced level of rock from 652 points in Bangalore is used to predict the reduced level of rock at any point in the subsurface of the Bangalore, where field measurements are not available. A new type of cross-validation analysis developed proves the robustness of the Ordinary Kriging model. ANN model based on multi layer perceptrons (MLPs) that are trained with Levenberg–Marquardt backpropagation algorithm has been adopted to train the model with 90% of the data available. The SVM is a novel type of learning machine based on statistical learning theory, uses regression technique by introducing loss function has been used to predict the reduced level of rock from a large set of data. In this study, a comparative study of three numerical models to predict reduced level of rock has been presented and discussed.     

受陆岸影响浅水区台风风浪的计算方案

在我国港口工程技术规范的波浪成长理论基础上，考虑到台风波浪的基本特点、等效水深的概念、浅水区波浪的折射、不规则波的能量分布特性、受陆岸和岛屿影响角度范围内的有效能量风区长度等因素，提出了受陆岸、岛屿影响浅水区（包括海湾区）台风风浪的计算方案。本方案只需台风中心位置、中心气压和计算点于各方位的风区长度（受陆岸影响部分）等资料，便可快速地得到计算点的波浪要素。经实测资料验证，效果良好。     

The Darwin Rise demise: the western Pacific guyot heights trace the trans-Pacific Mendocino fracture zone

The Darwin Rise has been proposed so many times and in so many forms and places that the time has come to make a more comprehensive examination of the region. Lying on the NW Pacific Plate between the Geisha Guyots, the Mid-Pacific Mountains, the equator, and the trenches, the region is roughly bounded by magnetic anomaly M20 (147 Ma). It was subjected to a massive outpouring of lava about 105 to 120 Ma, which created the guyots and seamounts in that region. Guyots are excellent tools for studying events of long ago because they eroded in the same lowstand in the Cretaceous and guyot relief, therefore, is a surrogate for paleo-sealevel. The relief is derived by subtracting the break depth of the summit plateau of a guyot from the regional depth. Guyot relief would necessarily be less in the center than to the periphery if the feature formed on a pre-existing rise, as has been postulated. The existence of a paleo-Darwin Rise would give concentric contours for the region in question. Of the sixty guyots used in this study, thirty-seven of these guyots were surveyed using SASS multibeam in the Marcus-Wake seamount group. Twenty-three guyots were surveyed using random track single-beam sonar surveys. An entirely different scenario is shown. Data revealed a major fracture passing through the area coevally or after the guyots formed. Because the depths to the summit are not the same now, vertical tectonics occurred after subaerial erosion. This means the fracture formed during and after the erosion (roughly 105 Ma) and influenced the normal sequence of events in guyot formation. Depending on how one deciphers trends through the Hess Rise morass, SASS bathymetry shows a continuation of the Surveyor/Mendocino fracture zone swarm inside the M20 region to the NE of these data. The fracture swarm continues to the western Pacific trench system. Based on this information, if the Darwin Rise ever existed, it had to have done so elsewhere.     

黄河流域天然径流量突变性与周期性特征

突变性和周期性是水文时间序列的两个重要特征。黄河流域面积广阔，各区域水文水资源系统演变规律各不相同，它们的突变和周期变化及其形成的物理机制遍异，因此系统分析各区域水资源突变性和周期性特征及其影响机制具有重要意义。把黄河流域划分为15个区域，计算出各区域1951—1998年的年天然径流量系列。利用Mann-Kendall非参数检验方法检测黄河流域各区域年天然径流量的突变年份，结果表明各区域的突变年份不完全一致，主要在1953—1955年、1979—1983年、1991—1993年等发生了突变，这些突变与北半球气候突变具有一致性，且由于下垫面改变、人类活动等影响而复杂化。利用Morlet小波分析各区域年天然径流量的变化周期，发现主要存在3～4a、7～9a、11a的周期，形成这些周期的物理因子有太阳黑子、海—气相互作用和下垫面因素等。通过分析黄河流域主要产流区不同时段小波系数变化，发现20世纪80年代之后年径流量主要以短周期变化。