Presentation by Maria Castrillo, Institute of Physics of Cantabria (IFCA), at the Delft3D - User Days (Day 4: Water quality and ecology), during Delft Software Days - Edition 2018. Thursday, 15 November 2018, Delft.
What Goes Wrong with Language Definitions and How to Improve the Situation
DSD-INT 2018 Hydrodynamic and Water Quality modelization of Cuerda del Pozo reservoir - Castrillo
1. Hydrodynamic and Water Quality modelization
of Cuerda del Pozo reservoir
Maria Castrillo*, Daniel García, Fernando Aguilar, Jesús Marco, Agustín Monteoliva
*Presenting
<castrillo@ifca.unican.es>
University of Cantabria (Spain)
November 15, 2018
2. 2
Table of contents
1. Who are we?
2. Framework: XDC project
3. Use case: Cuerda del Pozo reservoir
4. Hydrodynamic model
5. Water quality model
6. Conclusion
4. 4
IFCA: Institute of Physics of Cantabria
• Joint center CSIC – University of
Cantabria (UC)
• Research lines:
→ Galaxies and AGNs
→ Observational Cosmology and
Instrumentation
→ Particle Physics and Instrumentation
→ Advanced Computing and e-Science
→ Nonlinear Dynamics, Meteorology and Data Mining
5. 5
Advanced Computing and e-Science group
• Objectives:
– Develop a complete support chain for
data management, specially large data
volumes using BigData techniques.
– Integrate different components for
mutidisciplinary applications (sensors,
data acquisition, computing frameworks,
simulation, etc.).
– Foster the usage, maintain and improve
the existing computing infrastructure.
– Study the scalability of those solutions in
HPC, supercomputing, clusters, Grid and
Cloud frameworks.
• Some of our projects
7. 7
XDC Project
The eXtreme DataCloud (XDC) project (H2020)
o Scalable technologies for federating storage resources and
managing data in highly distributed computing
environments.
o XDC addresses requirements from a wide range of User
Communities belonging to several disciplines.
o Testing the developed solutions against the real life use
cases provided by the Communities represented in the
Consortium.
8. 8
Lifewatch
Biodiversity and ecosystem research:
Lifewatch ERIC
o Water quality forecasting for supplying and other human-
related uses.
o Predicting alert and warning to allow citizens and authorities
to put in place appropriate countermeasures.
o By means of specific models that rely on heterogeneous
data sources (monitoring instrumentation, satellite data, and
meteorological data).
11. 11
The Cuerda del Pozo reservoir
o Located in the header of the Duero river in
Spain
o The dam was constructed in 1941
o It supplies drinking water to small cities
around and supports the agriculture in an
extended zone.
o 2176 Ha, maximum length of 12 km.
o 10 m average depth, 36 m maximum depth.
o 229.2 hm3 of capacity
o Recreational and touristic area, comprising a
river beach, with high affluence during the
summertime.
12. 12
The Cuerda del Pozo reservoir
o The reservoir of Cuerda del Pozo
has experienced diverse HABs
episodes.
o Between 2007 and 2010 several
harmful species were detected, like
Anabaena sp. and Aphanizomenon
flos-aquae.
o In 2010 several sewage treatment
facilities were put into operation in
different points of the drainage
basin.
13. 13
Previous work
DORII project (FP7)
ROEM+ project (LIFE)
o Platform with surface sensors and
vertical profiling: physical, chemical,
biological, etc.
o Buoys with sensors and sample
campaigns in different points of the
drainage basin.
o Data visualization tool.
o Watershed model for tributaries
flowrate and nutrients
characterization.
15. 15
Hydrodynamic model with Delft3D-FLOW
Horizontal resolution:
40 × 40 m → 122 × 95 cells
Vertical resolution:
⁓1 m layers → 35 layers
Temporal resolution:
5 min
Co-ordinate system:
Z model
16. 16
Setting up of the hydrodynamic model
o Forcing functions
• solar radiation (hourly basis)
• air temperature (hourly basis)
• relative humidity (hourly basis)
• wind speed and direction (temporal resolution
of 10 minutes)
o 5 tributaries as walking discharges:
• flow rate (daily basis)
• salinity (daily basis)
• temperature (daily basis)
o 2 outlets in the dam as open boundary:
• flow rate (daily basis)
o Modelling periods:
• Calibration: April – October 2014
• Validation: April – October 2015
0
50
100
150
200
250
300
350
400
01/04/2015 21/05/2015 10/07/2015 29/08/2015 18/10/2015
Dailyaverageradiation(Wm-2)
Date
0
1
2
3
4
5
6
7
8
01/04/2015 21/05/2015 10/07/2015 29/08/2015 18/10/2015
Flowrate(m3s-1)
Date
Revinuesa
Remonicio
Ebrillos
Dehesa
Duero
17. 17
Some results
o Direct results
• Water level temporal profile
• Temperature vertical profiles in monitoring
points
• Salinity concentration
o Derived results
• Thermal stratification
• Epilimnion and hypolimnion temperatures
• Thermocline location and temperature
Rimmer et al. 2005
19. 19
Water quality model with Delft3D-ECO
Based on hydrodynamic model output
and watershed model:
o 5 tributaries:
• flow rate (daily basis)
• P, N and Si dissolved and particulate
concentrations (daily basis)
o 2 outlets in the dam:
• flow rate (daily basis)
o 1 monitoring point
Setting up the model with the PLCT
Light regime
Nitrification
and
denitrification
Organic
matter
decomposition
Phytoplankton
processes
Silicate
disolution
Other
auxiliary
processes
20. 20
Schematic overview of the model
Water
Atmosphere
POX1
DOX
PO4
NH4
CO2CO2
NO3
DO
OPAL Si
Phytoplankton
POX2
O2
CO2
21. 21
Some results
o Limiting factors
•Until July, nitrogen is the limiting factor.
•From July, light becomes the limiting factor.
•PO4-P concentration remains near constant at about 0,005 g/m3.
Phytoplankton
concentrates at
the depth of the
metalimnion.
Depth averaged (2-10 m) chlorophyll temporal profile Chlorophyll vertical profile
22. 22
Problems found
• Auxiliary process EMERSION is switched on with Zthreshold = 0.01 m. Switch
indicating emersion and submerssion works properly but cells with switch = 1 are not
deactivated.
• If discharges are located in cells that become dry, or its surrounding cells become
dry, the flow is interrupted, although they are set as walking discharges in the
hydrodynamic model.
24. 24
Conclusions and next steps
o Conclusions
• Delft3D has been succesfully implemented
for hydrodynamics and water quality
modelization in Cuerda del Pozo reservoir.
• A large amount of data is needed, coming
from highly heterogeneous data sources.
• Large models require powerful computing
and storage resources.
• Representative use case of a cloud
infrastructure: computing, storage and
virtual research environments can be
provided.
o Next steps
• Developing tools for integrated data pre-
processing to be used as input for
forecasting models.
• Implementing practices for proper
metadata management.
• Orchestration and workflow integration.
• Obtaining data to perform forecasting.