E3S Web of Conferences 7, 12005 (2016) DOI: 10.

1051/ e3sconf/2016 0712005

FLOODrisk 2016 - 3rd European Conference on Flood Risk Management

Flood protection as a key-component of the environmental restoration of

Canal del Dique, Colombia
1,a 1 1 1 1 2
Marius Sokolewicz , Eisse Wijma , Harm Nomden , Tjeerd Driessen , Quintijn van Agten and Fortunato Carvajal
1
Royal HaskoningDHV, P.O. Box 1132, 3800 BC Amersfoort, the Netherlands
2
Royal HaskoningDHV, Calle 90 No 19-41 Oficina 801, Bogotá, Colombia

Abstract. Canal del Dique is a man-made distributary of Rio Magdalena. After its  
degradation caused by abundant sediment load and changes to hydrology took a catastrophic form. In 2010, the
   In 2013 a huge effort to restore the environment in the
Canal del Dique system and to flood-proof the villages was started. An integrated approach was adopted to provide an
optimal solution for flood control, environment, fresh water supply and navigation. In order to prepare restoration
plans, an extensive hydrologic and hydraulic assessment was carried out. 1D, 2D and 3D numerical models were
developed to answer different questions, to evaluate different alternatives and to enable selection of optimal solutions.
To assess the flood risk, a hindcast of 2010 flood was carried out. A solution was designed in which the inlet of water
from the Magdalena River is regulated by a control structure, managed by an automated system based on SCADA
and Delft-FEWS flood forecasting software and advanced statistics.

1 Introduction ! )&- %  

   /  )      /
) %  &&% !   
1.1 Historical background &    )   &/     )    &
            !
             $  0   1 %   
           ! "    *  /  ! 2 
     # !$   )/)   
%&   ' ( )    ,   ) & &
( *+, *    -          & 
  '(.+!   % &    ! $  0 %/
            )! 3
  (   4  %    & 
  /  % /)   
&-  ))56!



Figure 1. Overview hydraulic behaviour of the Canal del
Dique system during high water conditions

$  /      - %  &  
)  &  ! &    - Figure 2 Caño Mahattes forming an inner delta in Ciénaga
&  &   - %/ &   &   
Capote as a result of large sediment inflow
 *  %  % 
a
Corresponding author: [Link]@[Link]

© The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution
License 4.0 ([Link]
E3S Web of Conferences 7, 12005 (2016) DOI: 10.1051/ e3sconf/2016 0712005
FLOODrisk 2016 - 3rd European Conference on Flood Risk Management

 were deployed to register continuously water levels in the

Due to human intervention a number of ciénagas were Canal and ciénagas. Furthermore, an extensive data
isolated from the canal by a lockable weir; this area is collection of ecological and water quality parameters was
known as reservoir Guajaro. carried out.
     (  &     The collected information was used to develop a set
718/    89 -      of very detailed numerical models, comprising a 1D2D
&&/    & &4 SOBEK model of the Canal del Dique system and the Rio
 0 (! 2   /       
Magdalena from Calamar to its mouth (hydrodynamics &
          &  %! )   &
   / * '" 4+   salt intrusion), 2D SOBEK-WAQ suspended sediment
 % &    :        and water quality model of the Canal system, 3D Delft3D
       56! hydrodynamic and sediment models of the bifurcation
$  4  /  (     ( &  point and the coastal zone (see Figure 4).
        &9/ &
         & ))   / 
" 9!


Figure 4. Different numerical models developed by the

environmental restoration project

Figure 3. Santa Lucia town flooded in December 2010 The hydrodynamic models were calibrated to an accuracy
of 10-15 cm RMSE making them a reliable tool for flood
1.2 Environmental restoration project hazard mapping and flood control [2].
 4   &       &   & 
 &;-&  !"    
&  /  4 9  <-     &
     )#         
- )   )         
)        /      &
&  ) /- /-  /   
 ( )) %!

2 Analysis of the system

2.1 Investigations and development of models
A thorough analysis of the problems was required to
Figure 5. Extreme discharge statistics of Rio Magdalena in
determine the environmental baseline and to find the
Calamar
optimum solution meeting all project objectives. To
achieve that, a massive program of investigations was
Establishing the extreme water level statistics turned
started to complement the database collected in previous
out to be a challenge. The hydrometric station in Calamar
studies. A detailed land level map covering nearly 2,000
on Rio Magdalena close to the entrance to the Canal, has
km2 was constructed using Lidar technology. The
!         
bathymetry in the Canal and the connecting water bodies
However, the migration of large river islands in the Rio
were measured using classic echo-sounding techniques.
Magdalena close to Calamar strongly influences the
Extensive surveys were carried out to measure
relationship between the occurring discharge and the
discharges, sediment concentrations and transport.
observed water level, making the rating curve in Calamar
Salinities in the project area were monitored monthly and
to change in time. To determine the design (1/100 per
vertical salinity profiles were collected to determine
year) water level in Calamar, first the statistics of
salinity intrusion. 16 automatic hydrographic stations

2
E3S Web of Conferences 7, 12005 (2016) DOI: 10.1051/ e3sconf/2016 0712005
FLOODrisk 2016 - 3rd European Conference on Flood Risk Management

discharge as an independent variable had to be

established (Figure 5), and only then the corresponding
1/100 per year water level could be derived (direct
estimation from measured water levels would be
unreliable as the  
large impact on the water levels upstream). This was
solved by deriving a new rating curve using all available
discharge measurements and establishing the limits of
this variability which ranged ± 10%. Selecting the
discharge measurements occurring in the period close to
 " ! #     $%
rating curve was representative for that period, therefore
that curve was used to establish the statistics. The
December 2010 discharge was found to correspond to 82 Figure 7. Example flood map
years return period.
2.3 Temporal climatic variations

2.3.1 El Niño and La Niña

The short term climatic variations in the discharge in the

Rio Magdalena are very much related to the El Niño and
la Niña climatic phenomena. The 2010 flood event for
instance was related to the La Niña period lasting from
summer 2010 - early 2013. Recent years, and in
particular 2015 and early 2016 are defined as a strong El
Niño, resulting in extremely low discharge in the Rio
Magdalena, and in drying up of ciénagas as a result of
Figure 6. Rating curves for Calamar closure of the caño connections and lack of sufficient
rainfall to feed the ciénagas. Figure 8 shows the water
level variation in the Rio Magdalena at Calamar for three
2.2 Modelling studies years representing the typical behaviour during El Niño
and La Niña climatic phenomena. Transitions from El
The developed numerical models were used to Niño to La Niña are not calendar limited, the year 1998
establish the environmental baseline for the project, to shows strong El Niño at the beginning of the year and La
Niña at the end of the year. Whereas other years 1999 are
analyse the system and to test different solution
fully influenced by La Niña.
alternatives. This analysis comprised hydrological
balance, hydrodynamics of the system, sediment
transport, salt intrusion and water quality. An important
aspect of these studies was the flood hazard assessment,
which was used to determine the location and level of
urgent protection works.
The SOBEK 1D2D model was used to simulate the
2010 flood. The results of simulation (Figure 7) were
compared to the (scarce) information from the field,
mostly reports from villages which were either flooded or
not flooded. After the solution alternatives were defined
(see Chapter 3), they were also assessed for flood hazard.

Figure 8. Water level variation in the Rio Magdalena at

Calamar for three years representing the typical behaviour of El
Niño and La Niña climatic phenomena.

3
E3S Web of Conferences 7, 12005 (2016) DOI: 10.1051/ e3sconf/2016 0712005
FLOODrisk 2016 - 3rd European Conference on Flood Risk Management

(Torres et al.) [7] found that the SLR in Cartagena for the
period 1993-2010 is 5.3 ± 0.3 mm compared to 1.7 ± 0.1
mm SLR in Cristobal at the Caribbean coast of Panama.
This indicates large influence of tectonic movements on
the relative SLR.

Figure 9. variations in the Oceanic Niño Index (ONI)

indicating the El Niño and La Niña phenomena.

Figure 9 shows the direct relations between the Oceanic

Niño Index (ONI) and the records of water levels at
Calamar. The ONI index is a measure for the cyclic
behaviour of average sea water temperature [3].

2.3.2 Impact on river discharge

On the long term, climate change further affects the

Figure 10. Trend in Sea Level Rise for Cartagena Bay,
discharge in the Rio Magdalena. The climate change in
Consorcio Dique.
the Magdalena River was studied to define the future
requirements of the system in terms of water supply and The global eustatic SLR predictions of IPCC [5] give
flood threat. The literature survey showed that climate an upper range of 82 cm in 2010. Extrapolation of this
change in the Magdalena basin is expected to bring a trend to a figure of 96 cm in 100 years. Considering the
change in the seasonal variability in climatic conditions, tectonic movements, but also large uncertainties in the
leading to higher temperatures, longer dry periods and predictions, a SLR of 1 meter in 100 years was assumed
more extreme wet periods. in the project.
An article published by Nakaegawa & Vergara [4] Such rise of sea level could have a huge impact on the
(referred to in the Fifth Assessment Report of IPCC [5]) low-lying delta, see Figure 11. It could be expected,
gives a good view on expected changes in precipitation, however, that in a sedimentary system the land in the
evaporation, surface runoff patterns in the Rio Magdalena delta will be built up with mangrove vegetation trapping
catchment and the resulting changes in river discharge sediments, which will partly compensate the impact of
(temporal as well as in magnitude) of the Magdalena rising sea level [8], [9].
River. The main conclusions based on the latter study are:
- The average river discharge in the Río Magdalena
will decrease in the coming century;
- Discharges in the months February-April will
decrease with 10% (around -400m3/s)
- Discharges during the first peak in June-July will
increase with 10% (around +850m3/s)
- Discharges during the second peak in November-
December will decrease with 12% (around -
850m3/s).

It was concluded that the maximum increase of 10%

for peak discharges should be used to include also the
effect of climate change on the river discharges of the Río Figure 11. Flooded area in the delta of Canal del Dique
Magdalena. Also 10% increase was adopted for the with SLR of 1m [9]
extreme discharges. This was in agreement with the
analysis of Future Water [6] which found an increase of In order to establish the impact of SLR, the project
10% of precipitation in the Magdalena basin. established the Mean Sea Level (MSL) in Cartagena in
the period 2008-2013 to the national datum (msnm,
meters above sea level in Buenaventura on the Pacific
2.3.3 Sea Level Rise
coast of Colombia). This relationship had never before
& ' ( )    *   Bay been derived for the Caribbean coast of Colombia. By
of Cartagena shows a trend of rise (Figure 10) which is referencing the hydrographic station in Cartagena to
nearly double of the eustatic Sea Level Rise (SLR).

4
E3S Web of Conferences 7, 12005 (2016) DOI: 10.1051/ e3sconf/2016 0712005
FLOODrisk 2016 - 3rd European Conference on Flood Risk Management

msnm, MSL was established to be 0.35 m above

msnm.[2]

3 Study of alternatives
Three basic alternatives were constructed, all
*  "+  + In all alternatives, a
control structure near Calamar regulates the discharge
entering the Canal. It is operated in such a way that it
mimics the natural hydrograph with two high water peaks
in June and November, allowing for flushing the ciénagas
while topping-off high discharges (see Figure 12),
providing for flood protection.
Figure 14. Alternative 2

In Alternative 3, another control structure is added

in the middle of the system. This creates a cascade
system consisting of two compartments in which the
water level can be fully controlled without requiring large
volumes of water. Parallel to the Canal lies a cascade
system of connected ciénagas that functions
independently of the Canal (stagnant water in the canal
while water flowing through the ciénagas).

Figure 12. Controlled hydrograph in Canal del Dique

Alternative 1 consists of one regulating structure in

Calamar, which allows for reduction of volume of water
and sediment entering the Canal. Water and sediment
follow the same path as in the present situation. The base
flow is just enough to ensure the minimum water depth
for navigation in the Canal. To prevent emptying of the
ciénagas, the open caño connections are closed. Inflow of
water into the ciénagas is regulated by an inflow section
at the upstream of the ciénagas to mimic the historic
cascade-wise flow pattern. Figure 15. Alternative 3

4 Integral solution
A solution based on principles of Alternative 2, fully
"+ +, was designed in which the
inlet of water from the Magdalena River is regulated by
the main control structure [10]. This provides a controlled
dynamics of water levels necessary for environmental
purposes while keeping the extreme flows out of the
Canal system. A second main structure directs the
majority of the flow through the delta to the open sea,
outside of the bays suffocating from sediments in the
Figure 13. Alternative 1 present situation. Both structures are combined with large
ship locks for navigation. The water exchange between
In Alternative 2 a control structure is added at the the Canal and ciénagas is controlled by fixed weirs.
downstream end of the Canal (at Puerto Badel), These structures assist in mimicking the natural
preventing any flow from the Canal into the Cartagena behaviour of water level variation in the ciénagas and
and Barbacoas bays. The base flow is directed through flow orientation through the ciénagas, allowing uniform
caño Correa in the delta to ensure the water level in flushing of ciénagas necessary to keep them in good
ciénagas and to prevent salt intrusion. At flood stage, environmental shape. In the current situation most
excess water is directed over the delta. ciénagas are fed by water coming through the caño, often

5
E3S Web of Conferences 7, 12005 (2016) DOI: 10.1051/ e3sconf/2016 0712005
FLOODrisk 2016 - 3rd European Conference on Flood Risk Management

located at the downstream end of the ciénagas, limiting

the equal distribution of the incoming water.
The first gate is located in the Canal del Dique at 3.2
km from the entrance near Calamar and is accompanied
by a ship lock (see Figure 16). The gate will consist of 3
openings of each 20 meters wide. The lower crest level is
located at -1 msnm.

Figure 16. Control structure in Calamar

5 Control system
The control structures will be managed by an
automated system. This system, based on SCADA and
Delft-FEWS flood forecasting software and advanced
statistics, uses the hydrographic stations along the Figure 17. Water level near Incora (top) and discharge
Magdalena River and in the Canal to predict water levels entering the Canal Del Dique (bottom) in the present situation
near the bifurcation point. A fixed set of operating rules and with the control structure.
has been derived. Together with real-time and forecasted
water levels, the optimum management choice will be It is not allowed to let more water into the Canal Del
determined on daily basis and the control structures can Dique system than the discharge necessary for the 7.0
be adjusted accordingly. msnm at Incora. Exceeding the maximum water level at
The operation rules are based on required water levels Incora of 7 msnm leads to flooding of villages along the
at Incora (7 km from the entrance near Calamar), a few ciénagas in the Canal del Dique system. The operational
kilometers downstream of the gate. The required water rules of the gate are such that the maximum water level
level at Incora has been derived from system cannot be exceeded. In wet years it is possible to bring in
requirements such as minimum water levels to connect a relatively high discharge for a longer period of time for
the Canal with ciénagas and Embalse Guajaro. Figure 17 the benefit of the dynamics in the ciénagas without
shows the water levels and corresponding discharges at compromising on the maximum acceptable water level at
Incora for the year 2005. The pattern of the controlled Incora.
inflow is defined by 3 flows:
- Base flow (1.83 msnm at Incora): Minimum level 6 Conclusion
necessary for navigation.
- Medium high peak flow (varying water level Numerous natural disasters happening in the Canal
around 5.2-6.0 msnm): Used for filling of the del Dique system in the last couple of years clearly show
large reservoir (Embalse Guajaro). Duration the urgent need for a large scale intervention. Restoration
around 30 days. of the environmental system aims at improving the water
- High peak flow (7.00 msnm at Incora): Used for system with respect to its most important functions:
flushing ciénagas, during a maximum of 50 days. navigation, ecology of ciénagas, water quality in the bays
The possibility of bringing in water from the of Cartagena and Barbacoas and flood protection of the
Magdalena River into the Canal del Dique is largely entire area. The proposed solution reduces the inflow of
determined by the water level in the Rio Magdalena in water and sediment into the Canal del Dique and tops off
front of the entrance with the Canal del Dique. In case the peak discharge to ensure flood safety. It redistributes
water levels at Calamar reach extreme levels (like 2010) water in the system to better facilitate various system
the gates will be opened until the water level at Incora functions.
reaches the limit of 7.0 msnm and then closed, allowing
only environmental flow to pass. The extreme water
levels in the Río Magdalena near Calamar, upstream of
the Canal entrance, will increase by 10 cm because of
closing the Canal.

6
E3S Web of Conferences 7, 12005 (2016) DOI: 10.1051/ e3sconf/2016 0712005
FLOODrisk 2016 - 3rd European Conference on Flood Risk Management

7 Acknowledgements

This article is based on work carried out by Consortio

Dique, a joint venture of Royal HaskoningDHV of the
Netherlands and Gomez Cajiao of Colombia under the
assignment from Fondo Adaptación. Deltares was
involved in the project as a technical advisor and quality
reviewer.

8 References

1. Mogollón Vélez J.C. (2013). El Canal Del Dique -
historia de un desastre ambiental, El Áncora Editores
Bogotá, ISBN 978-958-36-0149-1
2. Consorcio Dique (2015). Recalibración de los
modelos hidráulicos del Canal del Dique,
[Link].00-005
3. [Link]
analysis_monitoring/ensostuff/[Link], accessed
on 22.02.2016
4. Nakaegawa T. and Vergara W. (2010). First Projection
of Climatological Mean River Discharges in the
Magdalena River Basin, Colombia, in a Changing
Climate during the 21st Century. Hydrological Research
Letters, 50-54.
[Link] (2013). Climate Change 2013: The Physical
Science Basis. Contribution of Working Group I to the
Fifth Assessment Report of the Intergovernmental Panel
on Climate Change.
6. Future Water, Sar Vision, UNESCO IHE, Deltares
(2014). Climate adaptation Colombia, Scaling of climate
data for the Magdalena basin.
7. Torres, R. Ricardo; Tsimplis, Michael N. Sea-level
trends and interannual variability in the Caribbean Sea.
Journal of Geophysical Research: Oceans, 118 (6). 2934-
2947
8. Krauss, K. W., McKee, K. L., Lovelock, C. E.,
Cahoon, D. R., Saintilan, N., Reef, R. and Chen, L.
(2014). How mangrove forests adjust to rising sea level.
New Phytologist, 202: 19,34. doi: 10.1111/nph.12605
9. -.*/# 01$7*89:;
de cambio climático, CD ID 113 HIDR INF 00-002-C2
10. Consorcio Dique (2015). Desarrollo, evaluación y
selección de las alternativas, [Link].00-
002