Faculty of Environment: Martin Neruda, Tomáš Přikryl, Jitka Fikarová River Board Povodí Ohře: Lucie Tichá Belfast, Questor, 9.3.-12.3.2010.

Slides:



Advertisements
Similar presentations
Michael R. Smith, Mark Clement, Tony Martinez, and Quinn Snell
Advertisements

...if the last time you sampled your lake plankton during the summer holiday, this is like monitoring a temperate forest shortly after the last ice age.
FLOOD FORECASTING IN THE REPUBLIC OF MACEDONIA TRAINING WORKSHOP ON FLOOD RISK ASSESSMENT September 27 – October , Istanbul, TURKEY FLOOD FORECASTING.
A Brief Overview of Neural Networks By Rohit Dua, Samuel A. Mulder, Steve E. Watkins, and Donald C. Wunsch.
Multilayer Perceptrons 1. Overview  Recap of neural network theory  The multi-layered perceptron  Back-propagation  Introduction to training  Uses.
4 th International Symposium on Flood Defence Generation of Severe Flood Scenarios by Stochastic Rainfall in Combination with a Rainfall Runoff Model U.
Charles Rodenkirch December 11 th, 2013 ECE 539 – Introduction to Artificial Neural Networks PREDICTING INDIVIDUAL PLACEMENT IN COLLEGIATE WATERSKI TOURNAMENTS.
APPLICATION OF NEURO-GENETIC OPTIMIZER FOR SEDIMENT FORECASTING IN LAM PHRA PHLOENG RESERVOIR Thanyalak Iamnarongrit Assoc. Prof. Kampanad Bhaktikul, Assoc.
Forecasting impact of climate change on runoff coefficient in Limpopo basin using ANN MARCH , 2004AF_42 DAKAR WORKSHOP WELCOME TO THE SECOND AIACC.
Applications of Scaling to Regional Flood Analysis Brent M. Troutman U.S. Geological Survey.
J. K. Gietl * and O. Klemm Institute of Landscape Ecology, University of Münster, Germany Corresponding author:
The Need for a New Paradigm Including for Rainfall Forecasting Jennifer Marohasy PhD
Nidal Salim, Walter Wildi Institute F.-A. Forel, University of Geneva, Switzerland Impact of global climate change on water resources in the Israeli, Jordanian.
1 STARMAP: Project 2 Causal Modeling for Aquatic Resources Alix I Gitelman Stephen Jensen Statistics Department Oregon State University August 2003 Corvallis,
Neural Networks Neural Network Application For Predicting Stock Index Volatility Using High Frequency Data Project No CFWin03-32 Presented by: Venkatesh.
1 Chapter 11 Neural Networks. 2 Chapter 11 Contents (1) l Biological Neurons l Artificial Neurons l Perceptrons l Multilayer Neural Networks l Backpropagation.
Introduction to Neural Networks Simon Durrant Quantitative Methods December 15th.
An Overview of State-of- the-Art Data Modelling Introduction.
Satellite based mapping of lakes and climatic variations in the Ruizi and Katonga Catchments, Uganda Bernard Barasa December, 2014.
Soft Computing Colloquium 2 Selection of neural network, Hybrid neural networks.
Stream-Based Electricity Load Forecast Authors: Joao Gama Pedro Pereira Rodrigues Presented by: Viktor Botev.
Multiple-Layer Networks and Backpropagation Algorithms
ENWAMA meeting in Budapest Corvinus University.
Artificial Intelligence Lecture No. 29 Dr. Asad Ali Safi ​ Assistant Professor, Department of Computer Science, COMSATS Institute of Information Technology.
Center for Hydrometeorology and Remote Sensing, University of California, Irvine Basin Scale Precipitation Data Merging Using Markov Chain Monte Carlo.
From Machine Learning to Deep Learning. Topics that I will Cover (subject to some minor adjustment) Week 2: Introduction to Deep Learning Week 3: Logistic.
Supervisor: Dr. Eddie Jones Co-supervisor: Dr Martin Glavin Electronic Engineering Department Final Year Project 2008/09 Development of a Speaker Recognition/Verification.
1 Chapter 11 Neural Networks. 2 Chapter 11 Contents (1) l Biological Neurons l Artificial Neurons l Perceptrons l Multilayer Neural Networks l Backpropagation.
Statistical Tools for Solar Resource Forecasting Vivek Vijay IIT Jodhpur Date: 16/12/2013.
1 GMDH and Neural Network Application for Modeling Vital Functions of Green Algae under Toxic Impact Oleksandra Bulgakova, Volodymyr Stepashko, Tetayna.
Multi-Layer Perceptron
The NOAA Hydrology Program and its requirements for GOES-R Pedro J. Restrepo Senior Scientist Office of Hydrologic Development NOAA’s National Weather.
Programming for Geographical Information Analysis: Advanced Skills Online mini-lecture: Introduction to Neural Nets Dr Andy Evans.
05/09/2009 Slide 1 of 19 Practical Linear and Nonlinear Modelling of Environmental Data: A Case Study for River Flow Forecasting Hua-Liang Wei, Stephen.
Akram Bitar and Larry Manevitz Department of Computer Science
ICS 586: Neural Networks Dr. Lahouari Ghouti Information & Computer Science Department.
HYDROASIA 2008 FLOOD ANALYSIS STUDY AT INCHEON GYO CATCHMENT TEAM GREEN NGUYEN HOANG HUYSUN YABIN GWON YONGHYEON SUZUKI ATSUNORI LI WENTAO LEE CHANJONG.
An Artificial Neural Network Approach to Surface Waviness Prediction in Surface Finishing Process by Chi Ngo ECE/ME 539 Class Project.
Macroecology & Conservation Unit
Floods: “Rain Rain, go away”!! Brooke Porter Science 1st Mr. Shepard.
Dr.Abeer Mahmoud ARTIFICIAL INTELLIGENCE (CS 461D) Dr. Abeer Mahmoud Computer science Department Princess Nora University Faculty of Computer & Information.
Dynamic Neural Network Control (DNNC): A Non-Conventional Neural Network Model Masoud Nikravesh EECS Department, CS Division BISC Program University of.
Colorado Basin River Forecast Center Greg Smith Senior Hydrologist National Weather Service Colorado Basin River Forecast Center January 25, 2011 Navajo.
LOAD FORECASTING. - ELECTRICAL LOAD FORECASTING IS THE ESTIMATION FOR FUTURE LOAD BY AN INDUSTRY OR UTILITY COMPANY - IT HAS MANY APPLICATIONS INCLUDING.
EMMA - Hydrology EUMETNET representatives visit, 21 st – 22 nd October 2013, CHMI, Praha Tomáš Vlasák.
Data Mining: Concepts and Techniques1 Prediction Prediction vs. classification Classification predicts categorical class label Prediction predicts continuous-valued.
Aquatic Biomes. Determined by Salt content Flow rate Size (sometimes) 2 major categories of aquatic biomes: Salt water system Freshwater.
Prepared by Fayes Salma.  Introduction: Financial Tasks  Data Mining process  Methods in Financial Data mining o Neural Network o Decision Tree  Trading.
Classification of models
Artificial Neural Networks
Identification on Significant Pressures - Surface Water Bodies
Artificial neural networks
The National Institute of Hydrology and Water Management
Artificial Intelligence (CS 370D)
CSE 473 Introduction to Artificial Intelligence Neural Networks
NEURAL NETWORK APPROACHES FOR AUTOMOBILE MPG PREDICTION
Audio Content Description
Application of satellite-based rainfall and medium range meteorological forecast in real-time flood forecasting in the Upper Mahanadi River basin Trushnamayee.
TOWARDS HIGH-RESOLUTION GLOBAL SATELLITE PRECIPITATION ESTIMATION
Overview of Models & Modeling Concepts
Artificial Neural Network & Backpropagation Algorithm
ARTIFICIAL NEURAL NETWORKS
A Comparative Study of the National Water Model Forecast to Observed Streamflow Data By leah Huling.
Introduction to Radial Basis Function Networks
Temporal Back-Propagation Algorithm
Fig. 1. A diagram of artificial neural network consisting of multilayer perceptron. This simple diagram is for a conceptual explanation. When the logistic.
Kreshna GOPAL C. Prakash KHEDUN Anoop SOHUN
Structure of a typical back-propagated multilayered perceptron used in this study. Structure of a typical back-propagated multilayered perceptron used.
CS621: Artificial Intelligence Lecture 18: Feedforward network contd
Akram Bitar and Larry Manevitz Department of Computer Science
Presentation transcript:

Faculty of Environment: Martin Neruda, Tomáš Přikryl, Jitka Fikarová River Board Povodí Ohře: Lucie Tichá Belfast, Questor,

Content Monitoring in Chabařovice and Most lakes Rehabilitation of Černý potok stream Rainfall-runoff modelling with Artificial Neural Networks

Chabařovice lake Measuring profiles: Eutrophication reservoir: 3 places: − 1) inside the reservoir - in the other end of the reservoir − 2) inside the reservoir - around the outlet − 3) outside of the reservoir - below the outlet in the channel Chabařovice lake: 1 place – inside the lake in the north – west part

Chabařovice lake – area of interest

Measuring profiles

Methodology Catch a small fish and plankton with special net Make a conservation in a small plastic bottle with a chemicals Idea is to say which species of fish come to the lake because of their influence to the fish stock From the last year: research of plankton species to have a detailed picture about the lake‘s ecosystem

Eutrophication lake - in the other end of the reservoir – fish results: 10 pieces of Scardinius erythrophthalmus + 1 piece of Tinca tinca

Eutrophication reservoir - around the outlet – fish results: 10 pieces of Scardinius erythrophthalmus – plankton results: Cyclops vicinus, Moina macrocopa, Ceriodaphnia reciculata – lots of them

Eutrophication lake - below the outlet in the channel No fish species Plenty of the invertebrates (juveniles of insects), many species of plankton – failed the conservation

Chabařovice lake – inside the lake in NW part Aimed only to plankton: Metacyclops gracilis, Melosira varicus, cyclops sp. Daphnia magna – many of them

Most lake – new one Profiles: 2 places 1) In side „ Most “ lake 2) Small lake above Most lake

Most lake - inside small species of plankton Chydorus sphaericus, Daphnia magna, Acanthocyclops nanus – many of them

Small lake above the Most lake 5 pieces of Scardinius erythrophthalmus Small species of plankton (algae)

Černý potok stream In Ore mountains – north from Usti – boarder with Germany 2 phases of rehabilitation: 1) 1st August- 31st December ) 1st August 2010 – 31st December 2010 Organised by Agency for Nature Conservation and Landscape Protection of the Czech Republic in Ústí nad Labem Pictures from October 2009:

Information about Rainfall-runoff modelling Main goal: improving flood warning system in Czech Hydrometeorological Institute For small size river basins Based on several time series of hourly measured data For short time runoff predictions (1-6 hours) based on runoff and rainfall data

R-R modelling Significant data for prediction (made by correlation analysis): runoff data, short time rainfall history, API values Multilayer perceptron Radial basis function Results made in 6 weeks internship (Petr Paščenko) within the scope of the project HPC Europa 2 at the EPCC Department of the Edinburgh University

Results Ploučnice River Main problem: - overtraining of the network – poor generalization - small number of extreme events which makes it difficult for a model to predict the amplitude of the event

Results Experiments with absolute and relative runoff prediction Neural model shows about 5 % improvement in terms of efficiency coefficient over linear models

Results B est behavior - Multilayer perceptrons with one hidden layer trained by back propagation algorithm and predicting relative runoff Genetically evolved input filter improves the performance of yet another 5 %

Future Experiments in on-line prediction using real-time data from Smeda River (Jizerské hory mountains) with 6 hours lead time forecast Following the operational reality we will focus on classification of the runoffs into flood alert levels

Results

The two diagrams show the statistically relevant input variables for the two linear regression models. The red cells variables have 0.001, the violet have 0.01, the blue 0.1 and the white cells stand for less significant inputs.

Thank you for your attention