Application of neural approaches to one-step daily flow forecasting in Portuguese watersheds

Pulido‐Calvo, Inmaculada; Portela, María Manuela

doi:10.1016/j.jhydrol.2006.06.015

Cited by 96 publications

(51 citation statements)

References 54 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In this way, high correlations can be achieved by mediocre or poor models. Similar conclusions were obtained in the forecasting of different kinds of time variables [44][45][46][47][48].…”

Section: Discussionsupporting

confidence: 80%

Deep water longline selectivity for black spot seabream (Pagellus bogaraveo) in the Strait of Gibraltar

et al. 2009

View full text Add to dashboard Cite

Species and size selectivity of the deep water longline traditionally used in commercial fishing of the black spot seabream (Pagellus bogaraveo) were studied in the Strait of Gibraltar with four sizes of hooks. Black spot seabream contributed up to 88% of the catch by number. Catch and by-catch rates differed for the different hooks and fishing trials. Significant differences in average fish length between all hooks, except in one case, were found. The comparison of two experimental fishing trials within 4 years indicates a displacement towards smaller sizes in the size frequency distributions. The results of this study show that the fishing gear can be size selective depending on hook size. The fitted selectivity models for each experiments were very different despite having two hooks in common. This is probably due to the very different catch size distributions in the two periods, which suggests that the population size structure changed significantly between

show abstract

“…In this way, high correlations can be achieved by mediocre or poor models. Similar conclusions were obtained in the forecasting of different kinds of time variables [44][45][46][47][48].…”

Section: Discussionsupporting

confidence: 80%

Deep water longline selectivity for black spot seabream (Pagellus bogaraveo) in the Strait of Gibraltar

et al. 2009

View full text Add to dashboard Cite

show abstract

“…In fact, continuous measurements of precipitation and stream discharge can be more easily obtained without cost compared to continuous measurements of soil characteristics, initial soil moisture, infiltration, and groundwater characteristics. Therefore, a 'black box' approach that operates based only on the first set measurements can be much more suitable for operational forecasting purposes than a physical based model that also requires the latter set of measurements (Tokar and Johnson 1999;Inmaculada and Maria 2007).…”

Section: Introductionmentioning

confidence: 99%

Projected streamflow in the Huaihe River Basin (2010–2100) using artificial neural network

Gao

Gemmer

Zeng

et al. 2009

Stoch Environ Res Risk Assess

View full text Add to dashboard Cite

Climate projections for the Huaihe River Basin, China, for the years 2001-2100 are derived from the ECHAM5/MPI-OM model based on observed precipitation and temperature data covering 1964-2007. Streamflow for the Huaihe River under three emission scenarios (SRES-A2, A1B, B1) from 2010 to 2100 is then projected by applying artificial neural networks (ANN). The results show that annual streamflow will change significantly under the three scenarios from 2010 to 2100. The interannual fluctuations cover a significant increasing streamflow trend under the SRES-A2 scenario (2051-2085). The streamflow trend declines gradually under the SRES-A1Bscenario (2024)(2025)(2026)(2027)(2028)(2029)(2030)(2031)(2032)(2033)(2034)(2035)(2036)(2037), and shows no obvious trend under the SRES-B1 scenario. From 2010 to 2100, the correlation coefficient between the observed and modeled streamflow in SRES-A2 scenario is the best of the three scenarios. Combining SRES-A2 scenario of the ECHAM5 model and ANN might therefore be the best approach for assessing and projecting future water resources in the Huaihe basin and other catchments. Compared to the observed period of streamflows, the projected periodicity of streamflows shows significant changes under different emission scenarios. Under A2 scenario and A1B scenario, the period would delay to about 32-33a and 27-28a, respectively, but under B1 scenario, the period would not change, as it is about 5-6a and the observed period is about 7-8a. All this might affect drought/flood management, water supply and irrigation projects in the Huaihe River basin.

show abstract

“…De entre os vários métodos de aprendizagem (calibração e validação) das ANNs, no presente estudo foi utilizado o algoritmo de cálculo de retro-propagação do erro (Pulido-Calvo e Portela, 2007). A aprendizagem por retro-propagação do erro consiste num algoritmo iterativo desenvolvido para resolver sistemas de equações lineares aplicado a redes neuronais com três ou mais camadas e que implementa o sistema de cálculo sucessivo das derivadas parciais numa direcção contrária à de propagação normal da informação através da rede (Rumelhart et al, 1986 O sobre ajustamento do modelo ocorre quando o erro referente ao subconjunto [CSS] é muito reduzido mas aumenta ao serem apresentados novos dados à rede.…”

Section: Redes Neuronais -Procedimento Geralunclassified

Previsão de secas na primavera em Portugal Continental com base em indicadores climáticos de larga escala

2015

View full text Add to dashboard Cite

O sucesso de uma estratégia de mitigação dos efeitos da seca passa pela implementação de um sistema de monitorização e previsão eficaz, capaz de identificar os eventos de seca e de seguir a sua evolução espácio-temporal. Neste artigo demonstrase a eficiência de redes neuronais artificiais na previsão, para Portugal, do índice de precipitação padronizada, SPI, relativo à primavera. A validação dos modelos recorreu ao hindcasting, designando-se, por tal, a técnica através da qual um dado modelo é testado mediante a sua aplicação a períodos temporais históricos, com comparação dos resultados obtidos com as respectivas observações. O índice SPI foi calculado à escala temporal de 6 meses tendo o hindcast utilizado como indicadores climáticos a oscilação do Atlântico Norte e temperaturas da superfície do mar. O estudo evidenciou a mais valia da inclusão dos anteriores predictores externos no modelo de previsão. Elaboraram-se, ainda, mapas de probabilidade de ocorrência de seca os quais constituem importantes ferramentas no planeamento integrado e na gestão de recursos hídricos.

show abstract

Application of neural approaches to one-step daily flow forecasting in Portuguese watersheds

Cited by 96 publications

References 54 publications

Deep water longline selectivity for black spot seabream (Pagellus bogaraveo) in the Strait of Gibraltar

Deep water longline selectivity for black spot seabream (Pagellus bogaraveo) in the Strait of Gibraltar

Projected streamflow in the Huaihe River Basin (2010–2100) using artificial neural network

Previsão de secas na primavera em Portugal Continental com base em indicadores climáticos de larga escala

Contact Info

Product

Resources

About