1Recent extensive water quality surveys in Ireland revealed that diffuse phosphorus (P) 2 pollution originating from agricultural land and transported by runoff and subsurface 3 flows is the primary cause of the deterioration of surface water quality. P transport from 4 land to water can be described by mathematical models that vary in modelling approach,

The Multi-Layer Feed-Forward Neural Network (MLFFNN) is applied in the context of river flow forecast combination, where a number of rainfall-runoff models are used simultaneously to produce an overall combined river flow forecast. The operation of the MLFFNN depends not only on its neuron configuration but also on the choice of neuron transfer function adopted, which is non-linear for the hidden and output layers. These models, each having a different structure to simulate the perceived mechanisms of the runoff process, utilise the information carrying capacity of the model calibration data in different ways. Hence, in a discharge forecast combination procedure, the discharge forecasts of each model provide a source of information different from that of the other models used in the combination. In the present work, the significance of the choice of the transfer function type in the overall performance of the MLFFNN, when used in the river flow forecast combination context, is investigated critically. Five neuron transfer functions are used in this investigation, namely, the logistic function, the bipolar function, the hyperbolic tangent function, the arctan function and the scaled arctan function. The results indicate that the logistic function yields the best model forecast combination performance.

The performance of three artificial neural network (NN) methods for combining simulated river flows, based on three different neural network structures, are compared. These network structures are; the simple neural network (SNN), the radial basis function neural network (RBFNN) and the multi-layer perceptron neural network (MLPNN). Daily data of eight catchments, located in different parts of the world, and having different hydrological and climatic conditions, are used to enable comparisons of the performances of these three methods. In the case of each catchment, each neural network combination method synchronously uses the simulated river flows of four rainfall-runoff models operating in design non-updating mode to produce the combined river flows. Two of these four models are black-box, the other two being conceptual models. The results of the study show that the performances of all three combination methods are, on average, better than that of the best individual rainfall-runoff model utilized in the combination, i.e. that the combination concept works. In terms of the NashSutcliffe R 2 model efficiency index, the MLPNN combination method generally performs better than the other two combination methods tested. For most of the catchments, the differences in the R 2 values of the SNN and the RBFNN combination methods are not significant but, on average, the SNN form performs marginally better than the more complex RBFNN alternative. Based on the results obtained for the three NN combination methods, the use of the multi-layer perceptron neural network (MLPNN) is recommended as the appropriate NN form for use in the context of combining simulated river flows.

Grid-oriented, physically based catchment models calculate fields of various hydrological variables relevant to phosphorous detachment and transport. These include (i) for surface transport: overland flow depth and flow in the coordinate directions, sediment load, and sediment concentration and (ii) for subsurface transport: soil moisture and hydraulic head at various depths in the soil. These variables can be considered as decoupled from any chemical phosphorous model since phosphorous concentrations, either as dissolved or particulate, do not influence the model calculations of the hydrological fields. Thus the phosphorous concentration calculations can be carried out independently from and after the hydrological calculations. This makes it possible to produce a separate phosphorous modelling component which takes as input the hydrological fields produced by the catchment model and which calculates, at each step the phosphorous concentrations in the flows. This paper summarise the equations and structure of Grid Oriented Phosphorous Component (GOPC) developed for simulating the phosphorus concentrations and loads using the outputs of a fully distributed physical based hydrological model. Also the GOPC performance is illustrated by am example of an experimental catchment (created by the author) subjected to some ideal conditions. Keywords: Phosphorous modelling; Soil phosphorous, Phosphorus transport; Grid Oriented Phosphorous Component IntroductionModelling of phosphorous (P) loss from agriculture land and its transport consists of two parts. The first part deals with simulating most of the chemical transformations and movements in the soil phosphorous cycle, whereas the second part focuses on the transport of phosphorous over and beneath the land surface until it reaches the water bodies. As it has been classified by Stevenson and Cole, (1999) the soil P compounds comprise of soluble inorganic and organic P, weakly adsorbed (labile) inorganic P, sparingly soluble P, insoluble organic P, strongly adsorbed and/or occluded P by hydrous oxides of Fe and Al, and fixed P of silicate minerals. The detached and dissolved P compounds from the parent material can be transported by the storm runoff in which they normally exist in dynamic equilibrium between the dissolved and sediment-bound or particulate forms (Lee et al., 1989).

Publication informationJournal of Hydrology, For many good and practical reasons, lumped rainfall-runoff models are widely used 10 to represent a catchment"s response to rainfall. However, they have some 11 acknowledged limitation, some of which are addressed here using a neuro-fuzzy 12 model to combine, in an optimal way, a number of lumped-sub-models. For instance, 13 to address temporal variation, one of the sub-models in the combination may perform 14 well under flood conditions and another under drier conditions and the neuro-fuzzy 15 system would combine their outputs for each time-step in a manner depending on the 16 prevailing conditions. Similarly to address spatial variation, one of the sub-models 17 may perform well for the upland parts of the catchment and another for the lowland 18 parts and again the neuro-fuzzy system is expected to combine the different outputs