Hydroinformatics: Data Mining in Hydrology

agreeablesocietyΤεχνίτη Νοημοσύνη και Ρομποτική

29 Οκτ 2013 (πριν από 3 χρόνια και 11 μήνες)

79 εμφανίσεις

IIHR SEMINAR (DECEMBER
3,
2010
)
EVAN ROZ

Hydroinformatics
: Data Mining in
Hydrology


UNESCO
-
IHE,
Delft, Dr.
Solomatine


Hydroinformatics


t
echniques
were adopted from computational intelligence
(CI)/intelligent systems/machine
learning
hydroinformatics


c
onceptual
model
: data for calibration
.


data
-
driven model: data for training/validation
.


Shortcomings:


knowledge extraction


Strengths:


models quickly developed


highly accurate short term forecast


feature selection algorithms






Data Mining in
Hydroinformatics


Rainfall
-
runoff
modeling/Short term
forecasts
(
Vos

&
Rientjes

2007)


R
ain
-
fall
-
runoff and
groundwater model
calibration
-
Genetic
Algorithm (
Franchini

1996)


Flood forecasting
(Yu & Chen 2005)


Evapotranspiration
(Kisi 2006) and infiltration
estimation (
Sy

2006)




Deltares


Vegetation Induced
Resistance (
Keijer

et
al.
2005)



Genetic programming
identifies a more
concise relationship
between vegetation
and resistance

1DV model versus GP

Equations of the 1DV model

Equation derived from genetic
programming

Imperial
College

of
London

Value of High
Resolution Precipitation Data


1.
Short
Term

Prediction

of
Urban

Pluvial
Floods

(Maureen
Coat

2010)


Objective:
Interpolate

available

rain

gauge data


2.
Real
-
time Forecasting of Urban Pluvial
Flooding
(
Angélica

Anglés

2010)


Objective: Improved
analysis of the
existing rainfall data obtained by both
rain
gauges and radar networks.



=

𝑅
𝑏

Physical
meteorology

Statistics based

Maureen
Coat
-
Tipping Bucket Interpolation


Inverse
Distance
Weight


Liska’s

Method



Polygone
of
Thiessen


Most Effective
: Kriging


Teschl

(2007)


Feed
forward
neural
network
trained with
reflectivity
data
at
four altitudes
above rain gauge


Objective
:
Estimate
precipitation at
tipping bucket.

IPWRSM Inspired Future Work

Combine:


1.
Radar reflectivity
data from
Davenport, IA
(KDVN)


2.
Interpolated
precipitation data
via Kriging of
tipping buckets

Questions?

Franchini
, M. and
Galeati
, G. (1997). “Comparing Several Genetic Algorithm

Schemes for the Calibration of Conceptual Rainfall
-
runoff Models.”

Hydrological

Sciences Jour
nal, 42, 3, 357


379.


Keijzer
, M., Baptist, M.,
Babovic
, V., and
Uthurburu
, J.R.
(2005).
“Determining

Equations
for Vegetation Induced
Resistance using
Genetic

Programming.”
GECCO’05
,
June 25

29
, 2005, Washington, DC, USA
.


See, L., Solomatine, D., and Abrahart, R. (2007).

Hydroinformatics
:

Computational Intelligence and Technological Developments in Water

Science Applications.”
Hydrological Sciences Journal
, 52,
3, 391


396.


Vos
, N.J. and
Rientjes

,T.H.M. (2008). “
Multiobjective

Training Of Artificial

Neural Networks For Rainfall
-
runoff
Modeling.”
Water
Resources

Research
, 44, W08434.