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.     

Variations in the natural abundance of 15N in estuarine suspended particulate matter: A specific indicator of biological processing

In this study, the ¹⁴N:¹⁵N ratio of suspended particulate material collected from the Tamar river estuary, south-west England, is described. Three populations of particles, distinguishable by their ¹⁵N content, were observed. This investigation has shown that populations of estuarine particles are generated by biological transformations in situ and that the ¹⁵N content of estuarine particles does not merely reflect hydrodynamic mixing of the freshwater and seawater source particulate material.     

Particulate organic matter in the coastal and estuarine waters of Goa and its relationship with phytoplankton production

In the coastal and estuarine waters of Goa, particulate organic carbon (POC) varied from 0.52 to 2.51 mg l^?1 and from 0.28 to 5.24 mg l^?1 and particulate phosphorus (PP) varied from 0.71 to 5.18 μg l^?1 and from 0.78 to 20.34 μg l^?1, respectively. The mean values of chlorophyll and primary productivity were 1.94 mg m^?3 and 938.1 mg C m^?2 day^?1 in the coastal waters and 4.3 mg m^?3 and 636.5 mg C m^?1 day^?1 in the estuarine waters, respectively.$\text{POC}\text{chl}$ ratios were low in June and October even when POC values were quite high. The POC in surface waters was linearly correlated with the chlorophyll content. Also PP increased when chlorophyll and primary productivity remained high. The results suggest that the phytoplankton was sharply increasing and contributed to POC and PP content. The percentage of detritus calculated from the intercept values of chlorophyll on POC varied from 46 to 76% depending on season. Results indicate that the major portion of POC and PP during postmonsoon (October–January) is derived from phytoplankton production while the allochthonous matter predominate during monsoon (June–September).     

Methods of array element localization for a towed underwateracoustic array

The location of the hydrophones on a towed underwater acoustic array as a function of time (array element localization) is needed for signal processing. Methods to perform this localization using least squares polynomial fitting to data from depth sensors, heading sensors, and sensors detecting a ping from a single source are discussed. Arc distance along the array is used as the independent parameter so that all solutions are constrained to be space curves. Examples of application to real data are presented, and techniques to discriminate against bad sensor data are discussed     

Metamorphism in Archaean greenstone belts: calculated fluid compositions and implications for gold mineralization

An inescapable consequence of the metamorphism of greenstone belt sequences is the release of a large volume of metamorphic fluid of low salinity with chemical characteristics controlled by the mineral assemblages involved in the devolatilization reactions. For mafic and ultramafic sequences, the composition of fluids released at upper greenschist to lower amphibolite facies conditions for the necessary relatively hot geotherm corresponds to those inferred for greenstone gold deposits (X_CO2= 0.2–0.3). This result follows from the calculation of mineral equilibria in the model system CaO–MgO–FeO–Al₂O₃–SiO₂–H₂O–CO₂, using a new, expanded, internally consistent dataset. Greenstone metamorphism cannot have involved much crustal over-thickening, because very shallow levels of greenstone belts are preserved. Such orogeny can be accounted for if compressive deformation of the crust is accompanied by thinning of the mantle lithosphere. In this case, the observed metamorphism, which was contemporaneous with deformation, is of the low-P high-T type. For this type of metamorphism, the metamorphic peak should have occurred earlier at deeper levels in the crust; i.e. the piezothermal array should be of the ‘deeper-earlier’type. However, at shallow crustal levels, the piezothermal array is likely to have been of ‘deeper-later’type, as a consequence of erosion. Thus, while the lower crust reached maximum temperatures, and partially melted to produce the observed granites, mid-crustal levels were releasing fluids prograde into shallow crustal levels that were already retrograde. We propose that these fluids are responsible for the gold mineralization. Thus, the contemporaneity of igneous activity and gold mineralization is a natural consequence of the thermal evolution, and does not mean that the mineralization has to be a consequence of igneous processes. Upward migration of metamorphic fluid, via appropriate structurally controlled pathways, will bring the fluid into contact with mineral assemblages that have equilibrated with a fluid with significantly lower X_CO2. These assemblages are therefore grossly out of equilibrium with the fluid. In the case of infiltrated metabasic rocks, intense carbonation and sulphidation is predicted. If, as seems reasonable, gold is mobilized by the fluid generated by devolatilization, then the combination of processes proposed, most of which are an inevitable consequence of the metamorphism, leads to the formation of greenstone gold deposits predominantly from metamorphic fluids.     

Response of circular footings and anchor plates in non-homogeneous elastic soils

The axisymmetric elastic response of circular footings and anchor plates in a linearly non-homogeneous elastic soil is analysed. It is assumed that footings/anchors are flexible and subjected to axisymmetric vertical loads. The response of the footing/anchor is modelled by using the classical Poisson–Kirchhoff thin plate theory. A variational technique is used to analyse the interaction problem. A representation for the contact stress is established by using a fundamental solution corresponding to a unit vertical pressure acting over an annular region in the interior of the non-homogeneous soil. The fundamental solution can be derived by using rigorous analytical procedures. The influence of the footing flexibility and the degree of soil non-homogeneity on the displacements, bending moments and contact stresses of a surface footing is examined over a wide range of governing parameters. In the case of anchor plates the influence of depth of embedment, degree of soil non-homogeneity and anchor flexibility on the anchor displacement is investigated.     

Small angle neutron scattering by opals

Micro- and non-crystalline opals, chalcedony and flint show diffuse small angle neutron scattering (SANS). Precious opals give rise to two additional intensity maxima at very small scattering angles which are due to Bragg reflections from the closest packed non-crystalline silica spheres. A small angle texture diagram reveals that the closest packing is faulty. Synthetic non-crystalline opals yield much less intense small angle scattering due to lower contrast between silica spheres and interstitial cement or particles; in this case intensity maxima were not observed. The outer part of the scattering curves of opal-CT and microcrystalline quartz deviates from Porod's law. The specific surface of natural non-crystalline opals ranges from 0.006 to 0.018 nm^–1. In microcrystalline opals, the specific surface is about 10 times larger than in non-crystalline opals.     

Comparison of results from several AOGCMs for global and regional sea-level change 1900–2100

 Sea-level rise is an important aspect of climate change because of its impact on society and ecosystems. Here we present an intercomparison of results from ten coupled atmosphere-ocean general circulation models (AOGCMs) for sea-level changes simulated for the twentieth century and projected to occur during the twenty first century in experiments following scenario IS92a for greenhouse gases and sulphate aerosols. The model results suggest that the rate of sea-level rise due to thermal expansion of sea water has increased during the twentieth century, but the small set of tide gauges with long records might not be adequate to detect this acceleration. The rate of sea-level rise due to thermal expansion continues to increase throughout the twenty first century, and the projected total is consequently larger than in the twentieth century; for 1990–2090 it amounts to 0.20–0.37 m. This wide range results from systematic uncertainty in modelling of climate change and of heat uptake by the ocean. The AOGCMs agree that sea-level rise is expected to be geographically non-uniform, with some regions experiencing as much as twice the global average, and others practically zero, but they do not agree about the geographical pattern. The lack of agreement indicates that we cannot currently have confidence in projections of local sea-level changes, and reveals a need for detailed analysis and intercomparison in order to understand and reduce the disagreements. Received: 1 September 2000 / Accepted: 20 April 2001     

Grain boundary diffusion in enstatite

We have investigated grain boundary diffusion rates in enstatite by heating single crystals of quartz packed in powdered San Carlos olivine (Mg_0.90Fe_0.10)₂SiO₄ at controlled oxygen fugacities in the range 10^?5.7 to 10^?8.7?atm and temperatures from 1350° to 1450?°C for times from 5 to 100?h at 1?atm total pressure. Following the experiments, the thickness of the coherent polycrystalline reaction rim of pyroxene that had formed between the quartz and olivine was measured using backscatter scanning imaging in the electron microprobe. Quantitative microprobe analysis indicated that the composition of this reaction phase is (Mg_0.92Fe_0.08)₂Si₂O₆. The rate of growth of the pyroxene increases with increasing temperature, is independent of the oxygen fugacity, and is consistent with a parabolic rate law, indicating that the growth rate is controlled by ionic diffusion through the pyroxene rim. Microstructural observations and platinum marker experiments suggest that the reaction phase is formed at the olivine-pyroxene interface, and is therefore controlled by the diffusion of silicon and oxygen. The parabolic rate constants determined from the experiments were analyzed in terms of the oxide activity gradient across the rim to yield mean effective diffusivities for the rate-limiting ionic species, assuming bulk transport through the pyroxene layer. These effective diffusivities are faster than the lattice diffusivities for the slowest species (silicon) calculated from creep experiments, but slower than measured lattice diffusivities for oxygen in enstatite. Thus, silicon grain boundary diffusion is most likely to be the rate-limiting process in the growth of the pyroxene rims. Also, as oxygen transport through the pyroxene rims must be faster than silicon transport, diffusion of oxygen along the grain boundaries must be faster than through the lattice. The grain boundary diffusivity for silicon in orthopyroxenite is then given by D¯gbSiδ=(3.3±3.0)×10^?9f0.0O₂e^?400±65/RT?m³s^?1, where the activation energy for diffusion is in kJ/mol, and δ is the grain boundary width in m. Calculated growth rates for enstatite under these conditions are significantly slower than predicted by an extrapolation from similar experiments performed at 1000?°C under high pressure (hydrous) conditions by Yund and Tullis (1992), perhaps due to water-enhancement of diffusion in their experiments.