scholarly journals Is the Deep-Learning Technique a Completely Alternative for the Hydrological Model?: A Case Study on Hyeongsan River Basin, Korea

2021 ◽  
Author(s):  
Jaewon Kwak ◽  
Heechan Han ◽  
Soojun Kim ◽  
Hung Soo Kim
Keyword(s):  
Deep Learning ◽  
River Basin ◽  
Hydrological Model ◽  
Short Term Memory ◽  
Hydrologic Model ◽  
Hydrologic Models ◽  
Runoff Simulation ◽  
Runoff Modeling ◽  
Learning Technique ◽  
Relationship Of

Abstract It is no doubt that the reliable runoff simulation for proper water resources management is essential. In the past, the runoff was generally modeled from hydrologic models that analyze the rainfall-runoff relationship of the basin. However, since techniques have developed rapidly, it has been attempted to apply especially deep-learning technique for hydrological studies as an alternative to the hydrologic model. The objective of the study is to examine whether the deep-learning technique can completely replace the hydrologic model and show how to improve the performance of runoff simulation using deep-learning technique. The runoff in the Hyeongsan River basin, South Korea from 2013 to 2020 were simulated using two models, 1) Long Short-Term Memory model that is a deep learning technique widely used in the hydrological study and 2) TANK model, and then we compared the runoff modeling results from both models. The results suggested that it is hard to completely replace the hydrological model with the deep-learning technique due to its simulating behavior and discussed how to improve the reliability of runoff simulation results. Also, a method to improve the efficiency of runoff simulation through a hybrid model which is a combination of two approaches, deep-learning technique and hydrologic model was presented.

Towards deep learning based flood forecasting for ungauged basins

2020 ◽  
Author(s):  
Frederik Kratzert ◽  
Daniel Klotz ◽  
Guy Shalev ◽  
Sella Nevo ◽  
Günter Klambauer ◽  
...  
Keyword(s):  
Deep Learning ◽  
Low Income ◽  
Short Term Memory ◽  
Flood Forecasting ◽  
Low Income Countries ◽  
Water Model ◽  
Ungauged Basins ◽  
Streamflow Forecasting ◽  
Runoff Simulation ◽  
Streamflow Forecasts

<p>Floods are among the most destructive natural hazards in the world. To reduce flood induced damages and casualties, streamflow forecasts should be as accurate as possible.</p><p>As of today, streamflow forecasts are usually made with either conceptual or process-based hydrological models. The problem these models usually have is that they perform best when calibrated for a specific basin, and performance degrades drastically if the models are used in places without historic streamflow measurements. To make things worse, some of the most devastating floods occur in developing and low-income countries, where historic records of streamflow measurements are scarce. Therefore, a central task for enhancing flood forecasts and helping local authorities to manage these areas is to provide high-quality streamflow forecasts in ungauged rivers. Although the IAHS dedicated an entire decade (2003-2012) to advance the problem of Prediction in Ungauged Basins the central goal remains largely a challenge.</p><p>In this talk, we will present a novel approach for tackling the problem of prediction in ungauged basins using a data-driven approach. More concretely, we show that the Long Short-Term Memory network (LSTM), which is a special type of a deep learning model, can serve as a generalizable rainfall-runoff simulation model. We will present recent results indicating that the LSTM gives on average better out-of-sample predictions (ungauged prediction) than e.g. the SAC-SMA in-sample (gauged) or the US National Water Model (Kratzert et al., 2019).</p><p>One place where these research results are already finding their way into operation is Google’s Flood Forecasting Initiative. The goal of this initiative is to provide (enhanced) flood warnings, where needed, starting with a pilot project in India. And as mentioned above, historic streamflow records in those regions are scarce, which motivates new and innovative approaches for enhanced streamflow forecasting.</p><p>References:</p><p>Kratzert, F., Klotz, D., Herrnegger, M., Sampson, A. K., Hochreiter, S., & Nearing, G. S.: Toward improved predictions in ungauged basins: Exploiting the power of machine learning. Water Resources Research, 55. https://doi.org/10.1029/2019WR026065, 2019.</p>

scholarly journals A note on leveraging synergy in multiple meteorological datasets with deep learning for rainfall-runoff modeling

2020 ◽  
Author(s):  
Frederik Kratzert ◽  
Daniel Klotz ◽  
Sepp Hochreiter ◽  
Grey S. Nearing
Keyword(s):  
Deep Learning ◽  
Short Term Memory ◽  
Rainfall Runoff ◽  
Data Set ◽  
Simulation Accuracy ◽  
Meteorological Forcing ◽  
Runoff Modeling ◽  
Long Short Term Memory ◽  
Rainfall Runoff Model ◽  
Runoff Model

Abstract. A deep learning rainfall-runoff model can take multiple meteorological forcing products as inputs and learn to combine them in spatially and temporally dynamic ways. This is demonstrated using Long Short Term Memory networks (LSTMs) trained over basins in the continental US using the CAMELS data set. Using multiple precipitation products (NLDAS, Maurer, DayMet) in a single LSTM significantly improved simulation accuracy relative to using only individual precipitation products. A sensitivity analysis showed that the LSTM learned to utilize different precipitation products in different ways in different basins and for simulating different parts of the hydrograph in individual basins.

scholarly journals Multiple solutions, multi-site, and parameter transfer to calibrate DHSVM hydrological model

2021 ◽  
Vol 43 ◽  
pp. e7
Author(s):  
Roberto Avelino Cecílio ◽  
Helder De Amorim Mendes ◽  
Sidney Sara Zanetti
Keyword(s):  
Multiple Solutions ◽  
Hydrological Model ◽  
Hydrologic Model ◽  
The Other ◽  
Hydrologic Models ◽  
Environmental Attributes ◽  
Input Parameters ◽  
Best Parameters

The application of hydrologic models often needs sets of input parameters related to environmental attributes which are not always available. This leads to the necessity of calibrating the input parameters. However, due to the non-linearity of the hydrologic phenomena, there may be multiple “best” solutions for the calibration. This paper proposes a method for calibrating the DHSVM hydrologic model using the concepts of multiple solutions, multi-site, and parameter transfer among catchments. Eight watersheds were calibrated, resulting in obtaining five sets of “best” parameters (clusters) for each one. Afterward, each watershed was modeled using the parameters of the other catchments in order to verify if the transfer of the calibrated parameters could promote satisfactory modeling of the streamflows. The results show that clusters calibrated for one watershed may be suitable for other catchments. Besdes, the calibrated parameters of the smaller catchments were satisfactory to simulate the streamflow of the bigger catchments. The proposed method can be useful in calibrating and extrapolating the input parameters to regions that do not have information about them.

scholarly journals Evaluation of TMPA Satellite Precipitation in Driving VIC Hydrological Model over the Upper Yangtze River Basin

Water ◽  
2020 ◽  
Vol 12 (11) ◽  
pp. 3230
Author(s):  
Bin Zhu ◽  
Yuhan Huang ◽  
Zengxin Zhang ◽  
Rui Kong ◽  
Jiaxi Tian ◽  
...  
Keyword(s):  
River Basin ◽  
Time Scales ◽  
Satellite Data ◽  
Yangtze River ◽  
Hydrological Model ◽  
Yangtze River Basin ◽  
Runoff Simulation ◽  
Satellite Precipitation ◽  
Upper Yangtze River ◽  
The Upper Yangtze River

Although the Tropical Rainfall Measurement Mission (TRMM) has come to an end, the evaluation of TRMM satellite precipitation is still of great significance for the improvement of the Global Precipitation Measurement (GPM). In this paper, the hydrological utility of TRMM Multi-satellite Precipitation Analysis (TMPA) 3B42 RTV7/V7 precipitation products was evaluated using the variable infiltration capacity (VIC) hydrological model in the upper Yangtze River basin. The main results show that (1) TMPA 3B42V7 had a reliable performance in precipitation estimation compared with the gauged precipitation on both spatial and temporal scales over the upper Yangtze River basin. Although TMPA 3B42V7 slightly underestimated precipitation, TMPA 3B42RTV7 significantly overestimated precipitation at daily and monthly time scales; (2) the simulated runoff by the VIC hydrological model showed a high correlation with the gauged runoff and lower bias at daily and monthly time scales. The Nash–Sutcliffe coefficient of efficiency (NSCE) value was as high as 0.85, the relative bias (RB) was −6.36% and the correlation coefficient (CC) was 0.93 at the daily scale; (3) the accuracy of the 3B42RTV7-driven runoff simulation had been greatly improved by using the hydrological calibration parameters obtained from 3B42RTV7 compared with that of gauged precipitation. A lower RB (14.38% vs. 66.58%) and a higher CC (0.87 vs. 0.85) and NSCE (0.71 vs. −0.92) can be found at daily time scales when we use satellite data instead of gauged precipitation data to calibrate the VIC model. However, the performance of the 3B42V7-driven runoff simulation did not improve in the same operation accordingly. The cause might be that the 3B42V7 satellite products have been adjusted by gauged precipitation. This study suggests that it might be better to calibrate the parameters using satellite data in hydrological simulations, especially for unadjusted satellite data. This study is not only helpful for understanding the assessment of multi-satellite precipitation products in large-scale and complex areas in the upper reaches of the Yangtze River, but also can provide a reference for the hydrological utility of the satellite precipitation products in other river basins of the world.

scholarly journals Deep Learning with a Long Short-Term Memory Networks Approach for Rainfall-Runoff Simulation

Water ◽  
2018 ◽  
Vol 10 (11) ◽  
pp. 1543 ◽  
Author(s):  
Caihong Hu ◽  
Qiang Wu ◽  
Hui Li ◽  
Shengqi Jian ◽  
Nan Li ◽  
...  
Keyword(s):  
Deep Learning ◽  
Short Term Memory ◽  
Network Models ◽  
Machine Learning Techniques ◽  
Rainfall Runoff ◽  
Ann Model ◽  
Runoff Simulation ◽  
Simulation Performance ◽  
Artificial Neural Network Ann ◽  
Runoff Events

Considering the high random and non-static property of the rainfall-runoff process, lots of models are being developed in order to learn about such a complex phenomenon. Recently, Machine learning techniques such as the Artificial Neural Network (ANN) and other networks have been extensively used by hydrologists for rainfall-runoff modelling as well as for other fields of hydrology. However, deep learning methods such as the state-of-the-art for LSTM networks are little studied in hydrological sequence time-series predictions. We deployed ANN and LSTM network models for simulating the rainfall-runoff process based on flood events from 1971 to 2013 in Fen River basin monitored through 14 rainfall stations and one hydrologic station in the catchment. The experimental data were from 98 rainfall-runoff events in this period. In between 86 rainfall-runoff events were used as training set, and the rest were used as test set. The results show that the two networks are all suitable for rainfall-runoff models and better than conceptual and physical based models. LSTM models outperform the ANN models with the values of R 2 and N S E beyond 0.9, respectively. Considering different lead time modelling the LSTM model is also more stable than ANN model holding better simulation performance. The special units of forget gate makes LSTM model better simulation and more intelligent than ANN model. In this study, we want to propose new data-driven methods for flood forecasting.

scholarly journals Sentimental analysis using recurrent neural network

2018 ◽  
Vol 7 (2.27) ◽  
pp. 88 ◽  
Author(s):  
Merin Thomas ◽  
Latha C.A
Keyword(s):  
Deep Learning ◽  
Short Term Memory ◽  
Learning Approaches ◽  
Short Term ◽  
Indian Language ◽  
Natural Languages ◽  
South Indian ◽  
Learning Technique ◽  
Long Short Term Memory ◽  
Traditional Approaches

Sentiment analysis has been an important topic of discussion from two decades since Lee published his first paper on the sentimental analysis in 2002. Apart from the sentimental analysis in English, it has spread its wing to other natural languages whose significance is very important in a multi linguistic country like India. The traditional approaches in machine learning have paved better accuracy for the Analysis. Deep Learning approaches have gained its momentum in recent years in sentimental analysis. Deep learning mimics the human learning so expectations are to meet higher levels of accuracy. In this paper we have implemented sentimental analysis of tweets in South Indian language Malayalam. The model used is Recurrent Neural Networks Long Short-Term Memory, a deep learning technique to predict the sentiments analysis. Achieved accuracy was found increasing with quality and depth of the datasets. 

scholarly journals Sentiment Analysis of Persian Movie Reviews Using Deep Learning

Entropy ◽  
2021 ◽  
Vol 23 (5) ◽  
pp. 596
Author(s):  
Kia Dashtipour ◽  
Mandar Gogate ◽  
Ahsan Adeel ◽  
Hadi Larijani ◽  
Amir Hussain
Keyword(s):  
Deep Learning ◽  
Sentiment Analysis ◽  
English Language ◽  
Short Term Memory ◽  
Support Vector ◽  
Feature Engineering ◽  
Short Term ◽  
Machine Learning Technique ◽  
Learning Technique ◽  
Long Short Term Memory

Sentiment analysis aims to automatically classify the subject’s sentiment (e.g., positive, negative, or neutral) towards a particular aspect such as a topic, product, movie, news, etc. Deep learning has recently emerged as a powerful machine learning technique to tackle the growing demand for accurate sentiment analysis. However, the majority of research efforts are devoted to English-language only, while information of great importance is also available in other languages. This paper presents a novel, context-aware, deep-learning-driven, Persian sentiment analysis approach. Specifically, the proposed deep-learning-driven automated feature-engineering approach classifies Persian movie reviews as having positive or negative sentiments. Two deep learning algorithms, convolutional neural networks (CNN) and long-short-term memory (LSTM), are applied and compared with our previously proposed manual-feature-engineering-driven, SVM-based approach. Simulation results demonstrate that LSTM obtained a better performance as compared to multilayer perceptron (MLP), autoencoder, support vector machine (SVM), logistic regression and CNN algorithms.

scholarly journals Rainfall-runoff simulations in the Lukovska River Basin with the HEC-HMS model

2019 ◽  
pp. 33-60
Author(s):  
Ranka Eric ◽  
Andrijana Todorovic ◽  
Jasna Plavsic ◽  
Vesna Djukic
Keyword(s):  
River Basin ◽  
Optimal Parameter ◽  
Hydrologic Model ◽  
Model Parameters ◽  
Rainfall Runoff ◽  
Flood Events ◽  
Runoff Generation ◽  
Hydrologic Models ◽  
Generation Mechanisms ◽  
Event Based

Hydrologic models are important for effective water resources management at a basin level. This paper describes an application of the HEC-HMS hydrologic model for simulations of flood hydrographs in the Lukovska River basin. Five flood events observed at the Mercez stream gauge were available for modelling purposes. These events are from two distinct periods and two seasons with different prevailing runoff generation mechanisms. Hence the events are assigned to either ?present? or ?past?, and ?spring? or ?summer? group. The optimal parameter sets of each group are obtained by averaging the optimal parameters for individual events within the group. To assess model transferability, its applicability for simulation of flood events which are not considered in the model calibration, a cross-validation is performed. The results indicate that model parameters vary across the events, and that parameter transfer generally leads to considerable errors in hydrograph peaks and volumes, with the exception of simulation of summer events with ?spring? parameters. Based on these results, recommendations for event-based modeling are given.

Sign in / Sign up

Export Citation Format

Share Document