Streamflow Estimation in a Mediterranean Watershed Using Neural Network Models: A Detailed Description of the Implementation and Optimization
This study compares the performance of three different neural network models to estimate daily streamflow in a watershed under a natural flow regime. Based on existing and public tools, different types of NN models were developed, namely, multi-layer perceptron, long short-term memory, and convoluti...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-03-01
|
| Series: | Water |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2073-4441/15/5/947 |
| _version_ | 1797614231222222848 |
|---|---|
| author | Ana Ramos Oliveira Tiago Brito Ramos Ramiro Neves |
| author_facet | Ana Ramos Oliveira Tiago Brito Ramos Ramiro Neves |
| author_sort | Ana Ramos Oliveira |
| collection | DOAJ |
| description | This study compares the performance of three different neural network models to estimate daily streamflow in a watershed under a natural flow regime. Based on existing and public tools, different types of NN models were developed, namely, multi-layer perceptron, long short-term memory, and convolutional neural network. Precipitation was either considered an input variable on its own or combined with air temperature as another input variable. Different periods of accumulation, average, and/or delay were considered. The models’ structures were optimized and automatically showed that CNN performed best, reaching, for example, a Nash–Sutcliffe efficiency of 0.86 and a root mean square error of 4.2 m<sup>3</sup> s<sup>−1</sup>. This solution considers a 1D convolutional layer and a dense layer as the input and output layers, respectively. Between those layers, two 1D convolutional layers are considered. As input variables, the best performance was reached when the accumulated precipitation values were 1 to 5, and 10 days and delayed by 1 to 7 days. |
| first_indexed | 2024-03-11T07:06:49Z |
| format | Article |
| id | doaj.art-b1a765cc7bfc43e9b25ab2cc439e05bf |
| institution | Directory Open Access Journal |
| issn | 2073-4441 |
| language | English |
| last_indexed | 2024-03-11T07:06:49Z |
| publishDate | 2023-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Water |
| spelling | doaj.art-b1a765cc7bfc43e9b25ab2cc439e05bf2023-11-17T08:55:17ZengMDPI AGWater2073-44412023-03-0115594710.3390/w15050947Streamflow Estimation in a Mediterranean Watershed Using Neural Network Models: A Detailed Description of the Implementation and OptimizationAna Ramos Oliveira0Tiago Brito Ramos1Ramiro Neves2Centro de Ciência e Tecnologia do Ambiente e do Mar (MARETEC-LARSyS), Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1, 1049-001 Lisboa, PortugalCentro de Ciência e Tecnologia do Ambiente e do Mar (MARETEC-LARSyS), Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1, 1049-001 Lisboa, PortugalCentro de Ciência e Tecnologia do Ambiente e do Mar (MARETEC-LARSyS), Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1, 1049-001 Lisboa, PortugalThis study compares the performance of three different neural network models to estimate daily streamflow in a watershed under a natural flow regime. Based on existing and public tools, different types of NN models were developed, namely, multi-layer perceptron, long short-term memory, and convolutional neural network. Precipitation was either considered an input variable on its own or combined with air temperature as another input variable. Different periods of accumulation, average, and/or delay were considered. The models’ structures were optimized and automatically showed that CNN performed best, reaching, for example, a Nash–Sutcliffe efficiency of 0.86 and a root mean square error of 4.2 m<sup>3</sup> s<sup>−1</sup>. This solution considers a 1D convolutional layer and a dense layer as the input and output layers, respectively. Between those layers, two 1D convolutional layers are considered. As input variables, the best performance was reached when the accumulated precipitation values were 1 to 5, and 10 days and delayed by 1 to 7 days.https://www.mdpi.com/2073-4441/15/5/947neural networksMLPLSTMCNNstreamflow estimation |
| spellingShingle | Ana Ramos Oliveira Tiago Brito Ramos Ramiro Neves Streamflow Estimation in a Mediterranean Watershed Using Neural Network Models: A Detailed Description of the Implementation and Optimization Water neural networks MLP LSTM CNN streamflow estimation |
| title | Streamflow Estimation in a Mediterranean Watershed Using Neural Network Models: A Detailed Description of the Implementation and Optimization |
| title_full | Streamflow Estimation in a Mediterranean Watershed Using Neural Network Models: A Detailed Description of the Implementation and Optimization |
| title_fullStr | Streamflow Estimation in a Mediterranean Watershed Using Neural Network Models: A Detailed Description of the Implementation and Optimization |
| title_full_unstemmed | Streamflow Estimation in a Mediterranean Watershed Using Neural Network Models: A Detailed Description of the Implementation and Optimization |
| title_short | Streamflow Estimation in a Mediterranean Watershed Using Neural Network Models: A Detailed Description of the Implementation and Optimization |
| title_sort | streamflow estimation in a mediterranean watershed using neural network models a detailed description of the implementation and optimization |
| topic | neural networks MLP LSTM CNN streamflow estimation |
| url | https://www.mdpi.com/2073-4441/15/5/947 |
| work_keys_str_mv | AT anaramosoliveira streamflowestimationinamediterraneanwatershedusingneuralnetworkmodelsadetaileddescriptionoftheimplementationandoptimization AT tiagobritoramos streamflowestimationinamediterraneanwatershedusingneuralnetworkmodelsadetaileddescriptionoftheimplementationandoptimization AT ramironeves streamflowestimationinamediterraneanwatershedusingneuralnetworkmodelsadetaileddescriptionoftheimplementationandoptimization |