Journal of Environmental Planning and Management

ISSN: 0964-0568 (Print) 1360-0559 (Online) Journal homepage: https://www.tandfonline.com/loi/cjep20

Modelling support to citizen observatories for

strategic Danube Delta planning: Sontea-Fortuna
case study

A. B. Venturini, T.H. Assumpção, I. Popescu, A. Jonoski & D. P. Solomatine

To cite this article: A. B. Venturini, T.H. Assumpção, I. Popescu, A. Jonoski & D. P. Solomatine
(2019) Modelling support to citizen observatories for strategic Danube Delta planning: Sontea-
Fortuna case study, Journal of Environmental Planning and Management, 62:11, 1972-1989, DOI:
10.1080/09640568.2018.1523787

To link to this article: https://doi.org/10.1080/09640568.2018.1523787

Published online: 21 Jan 2019.

Submit your article to this journal

Article views: 65

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=cjep20
Journal of Environmental Planning and Management, 2019
Vol. 62, No. 11, 1972–1989, https://doi.org/10.1080/09640568.2018.1523787

Modelling support to citizen observatories for strategic Danube

Delta planning: Sontea-Fortuna case study
A. B. Venturinia, T.H. Assumpç~aoa, I. Popescua,b
,
A. Jonoskia and D. P. Solomatinea,c
a
IHE Delft Institute for Water Education, Delft, The Netherlands; bFaculty of Civil Engineering,
Politechnica University of Timisoara, Romania; cWater Resources Section, Delft University of
Technology, Delft, The Netherlands

(Received 11 December 2017; final version received 11 September 2018)

Recent research developments indicate that citizens’ observatories, a novel

approach for data collection, management and governance, can provide valuable
contributions to strategic delta planning processes. Most citizen observatories are
limited in spatial coverage during data collection, according to the citizens’
availability and static locations. However, there are times in which citizens
participate as trained volunteers during data collection field campaigns. In this later
scenario, an important aspect in organising and maintaining citizen observatories is
having a clear plan for data gathering activities, including determination of routes
to be followed by these citizens. This article addresses the issue of determining
such routes (called pathways) related to the specific problem of gathering data in
deltaic areas, which are composed of intricate canals and wetlands. Data collection
activities consist of citizens acquiring images and videos with mobile phones at
predetermined locations (Points of Interest) that are only accessible by boats. A
pathway selection approach is presented, supported by a hydrodynamic model,
developed to capture the specific processes of the delta. The aim of the approach is
to define: where to send the citizens to fulfil stakeholders’ interest and planning
goals, how to reach these points and which pathways should be selected (prioritised
based on such interests), considering possible flood patterns. The developed
methodology is part of an ongoing H2020 funded research. The proof of
concept is carried out in the Sontea-Fortuna area, of the Danube Delta, which,
similar to many wetlands, contains a large variety of unique, endemic species.
With the developed approach, pathways to be followed by citizens during data
collection campaigns were generated and scored considering local interests.
Analysis of the boat’s accessibility into the delta during different hydrological
scenarios showed that the wetland is more accessible than expected, hence the
proposed pathway approach was useful in prioritizing some canals over others.
The determined pathways will be applied in field campaigns. The approach can
be applied to other delta regions with different environmental problems, such as
water quality or ecosystem assessment, contributing to the organisation of
effective active citizen data collection.
Keywords: delta planning; citizen observatories; Danube Delta;
hydrodynamic modelling

Corresponding author. E-mail: i.popescu@un-ihe.org

ß 2019 Newcastle University

 Journal of Environmental Planning and Management 1973

1. Introduction
Flooding is a natural phenomenon of river basin systems that can be regarded as either
a hazard or a service to the environment. Deltas are examples of areas with high bio-
diversity that require periodic flooding in order to ensure the health of their ecosystem
(Millennium Ecosystem Assessment 2005). In many regions of the world, past human
interventions have negatively affected deltas’ natural evolution. Direct interventions
include converting natural deltaic zones into agricultural areas, or conversion of natural
wetlands into artificial fishponds. Similarly, interventions upstream of the delta, such
as dam construction and water diversions, can also bring changes in inundation pat-
terns, as well as disruptions in water and sediment supply. Therefore, the comprehen-
sion of water circulation and horizontal mixing is very important for many successful
management decisions, such as restoration or enhancement project designs
(Arega 2013).
To properly identify drivers behind biodiversity degradation and to propose lasting
improvements in such systems, authorities can make use of strategic delta planning.
Strategic planning, in general, aims at balancing the actions to be taken by stakehold-
ers in a sustainable manner and as a long-term process (Albrechts 2004). Seijger et al.
(2017) expanded the urban strategic spatial planning concept for the particular case of
delta areas. This new delta specific defined framework, known as ‘Hourglass frame-
work’1, covers three different dimensions: (1) an agenda-setting for proper identifica-
tion of the problem by ‘actor coalitions’2 (2) innovative solutions to overcome the
drawbacks of conventional delta planning and management (e.g. floating houses), and
(3) use of participatory tools for delta planning and management. One of the require-
ments in delta planning and management is data accessibility for understanding and
properly representing system behaviour, frequently through the setup of inundation,
water quality or ecological models. These models are becoming increasingly complex
and data demanding, while data is a scarce resource in studies on natural ecosystems.
Implementation and maintenance of traditional monitoring services, through the instal-
lation of gauges and field campaigns, or acquisition of multi-level Earth observation,
can be insufficient and financially unsustainable (Giosan et al. 2014). Additionally, the
spatial coverage of these monitoring stations is limited to its static location, environ-
mental variables observed and users’ data needs.
Citizen observatories, where environmental monitoring is undertaken by local citi-
zens, have recently been introduced as an innovative approach to data collection, man-
agement and governance. They are driven by new and low maintenance measurement
equipment that, depending on the case study, can be more robust and low-cost. One of
the advantages of citizen observatories’ involvement in environmental planning is their
complementary way of collecting data, while it allows for indirect involvement of local
citizens in the environmental policy-making process, and, consequently, in the manage-
ment of deltas. However, citizen observatories also face challenges, such as maintain-
ing long-term continuous engagement of local communities and stakeholders and
ensuring minimum data quality standards and organizing campaigns (Gharesifard,
Wehn, and van der Zaag 2017; Zaman et al. 2014).
Although studies applying citizen observatories to water resources do exist, the sci-
entific development is still in its infancy. Reviews of case studies in which citizen
observatories were applied were carried out by Buytaert et al. (2014) in a hydrological
context; McKinley et al. (2017) for the fields of biology conservation, natural resource
management and environmental protection; Poser and Dransch (2010) and Klonner
1974 A. Venturini et al.

et al. 2016 in natural hazard management, and Assumpç~ao et al. (2018) for studies
related to flood modelling. To the best of the authors’ knowledge, the research con-
ducted by Busch et al. (2016) on phytoplankton dynamics in Ebro Delta, is the only
citizen observatories study in delta areas. To date, there is no application of citizen
observatories linked with the modelling of flood processes in delta systems. All
reviews show that citizen observatories are a viable way of gathering complementary
data, in multiple contexts, with some studies obtaining positive results in terms of
employing such a data source. In most water resources data gathering studies, data is
collected by citizens randomly. Mazzoleni et al. (2017) evaluated the spatial distribu-
tion of locations for water level data collection by citizens, in the context of data
assimilation for hydrological modelling. However, a lack of clear methods to identify
and select points of interest for data collection and routes to access these points of
interest, can be identified as a scientific gap within citizen observatories.
Recognizing the importance of citizen involvement in environmental monitoring
and management, the European Commission took the initiative to fund innovative citi-
zen observatories projects. The work presented in this article has been carried out as
part of the EU H2020 funded project, SCENT3, which aims to develop a strategy to
engage citizens in environmental monitoring (Tserstou et al. 2017). The project main-
tains the dual objective of gathering relevant data and facilitating citizens and local
stakeholders’ involvement in the management process, in two pilot areas: Danube
Delta in Romania and Attica region in Greece. The SCENT project started with
involvement of the stakeholders, in a process similar to the agenda-setting dimension
of the Hourglass framework. This initial phase aimed at identifying user requirements
through meetings, workshops and surveys conducted with local stakeholders (author-
ities, environmental monitoring institute, NGOs, and policy makers/public bodies). The
outcome served as a basis to determine purposes and clear locations for collecting data
with raking scores depending on order of importance for the environment and commu-
nity. Data collection will be performed through two campaigns, organized in conjunc-
tion with local decision-makers. They are comprised of field visits in the case study
area, training sessions and workshops. During the campaigns, citizens will board a
boat and follow a predefined route in the delta area. Data collection entails taking pic-
tures and videos in predefined locations, using developed apps and games. The
SCENT project scheduled campaigns for the end of 2018 and the beginning of 2019,
on routes determined based on an approach detailed herein.
Thus, the main research objective of the present study is to enable and improve
the use of citizen observatories for strategic delta planning, by developing an
approach to determine and assess possible pathways for citizen data collection cam-
paigns. More specifically, to fulfil the objective, the following research questions
are posed: where to send the citizens to fulfil stakeholders’ interest and planning
goals?; how to reach these points?; and because of resource availability for travel,
which pathways should be selected (prioritised based on stakeholders interests),
considering different flood patterns. Engagement of citizens in data collection,
along with facilitation of their involvement, is an important aspect of Citizen
Observatories and has been addressed by SCENT project partners located in the
Danube Delta area. Present research does not address this aspect, which is consid-
ered fulfilled, and adds to the citizen observatories effort of determining the paths
to be followed.
 Journal of Environmental Planning and Management 1975

The proposed solution comes as an innovative approach addressing the second

dimension of the Hourglass framework. Citizen observatories also focus on strengthen-
ing the actor coalitions as defined by the first dimension, which allows for a quick
start for the participatory process work (the third dimension of the framework).
The present article is structured in six parts. After this introduction, a description
of the case study is presented in Section 2; and the development of the flood model
followed by the proposed approach for citizen pathway selection are presented in
Sections 3 and 4. Section 5 shows the resultant pathways and further discusses and
analyses them. Finally, Section 6 presents the main conclusions and recommendations
for future research.

2. Case study description

Danube Delta (DD) is the third largest European delta, after those the Rhine and Ebro.
It contains one of the most well-preserved wetland areas in Europe, rich in biodiver-
sity, it is a RAMSAR site and one of the World Natural Heritage sites of UNESCO
(G€uttler, Niculescu, and Gohin 2013).
Danube Delta’s main water supply is the Danube River, which has, on average, a
discharge of 6,300 m3 s1 (Pringle et al. 1995). The entire DD is considered a low flat
plain with a gradient slope of around 0.043%. The Danube River splits into three main
branches (Chilia, Sulina, and St. George, Figure 1) forming the delta. In addition, only
5% of the Danube River flows into the inner network of the delta (Popescu et al.
2015). The lowest water levels are experienced every year from September to
November (Oosterberg et al. 2000).
There are seven naturally structured hydrographic systems in DD that have experi-
enced many changes throughout time (Cioca, Bondar, and Borcia 2010), some of
which are irreversible. Nevertheless, planning and management of the delta needs to
consider flooding patterns of the DD as one of most important phenomena for preserv-
ing the ecosystem. To this end, collection of data, including citizens’ contributions, is
important for the whole DD, however, this research is carried out on the Sontea-
Fortuna sub-area. The research outcomes could then be applied more generally.

2.1. Sontea-Fortuna wetland

The Sontea-Fortuna area is located in the central-west part of the Danube Delta, a link
between several hydrographic systems of the delta. It is the first ecosystem in the DD
to receive the waters from the Danube River, as such being an important hub in the
system. Understanding flow circulation in the Sontea-Fortuna area, as well as locations
of stagnant water in the historically modified delta, is important for proper planning of
measures to control the amount of water to be taken from the Danube River into the
delta. The case study area comprises Sontea-Fortuna hydrographic unit and Sireasa
Polder (Figure 1).
Sontea-Fortuna covers 290 km2, 7% of the whole DD, and together with Sireasa
Polder the whole study region is about 370 km2. The river network of the study area is
composed of a primary river network of 106 km, comprising the main branches of the
Danube Delta with surrounding channels, and, a secondary network of 153 km of
minor canals and brooks, in which water has high residence time (Oosterberg et al.
2000). Apart from the canal network, there are 11 larger natural lake formations in the
1976 A. Venturini et al.

Figure 1. Sontea-Fortuna and Sireasa Polder study area.

area (Bondar and Panin 2000). Vegetation in the area consists mainly of reed swamp,
which contributes to reducing the water velocity.

2.2. Strategic Danube Delta planning

The Danube Delta was recognized as a UNESCO Natural World Heritage Site in
1991. As a consequence, the Danube Delta Biosphere Reserve Administration
(DDBRA) was created and appointed to elaborate and implement a comprehensive and
effective management plan for the Delta (Van Assche et al. 2011). IUCN East
European Programme (1992) revealed that the biggest challenge in the DD is the frag-
mented and uncoordinated actions taken, sometimes by different actors, without a com-
prehensive integrated management approach. At the request of DDBRA, the Danube
Delta National Institute for Research and Development (DDNI) elaborated, in 2005,
the Danube Delta Regional Development Master Plan, which is publicly available
(DDNI and IVL 2006).
This plan contains similar elements to the strategic planning defined by Albrechts
(2004) and Seijger et al. (2017). For example, there is a long-term plan in which one
of the main objectives is the Integrated Monitoring System for all DDBRA districts.
To do that, short-term objectives are assigned in the master plan, one of which is to
have a local, regional and national monitoring program with excellence centres and
technical assistance. Additionally, the use of hydraulic modelling is mentioned as an
important support of the master plan. Modelling objectives are to explore possible
actions towards restoration of agricultural land and fishponds by simulating breach
openings and gate operations under different flood scenarios, such that the water circu-
lation is similar to the natural state before damming the delta.
 Journal of Environmental Planning and Management 1977

The main similarities of the master plan with the Hourglass framework are in the con-
cept of including diverse actors and evaluation of multiple disciplines’ policies. The plan-
ning methodology, however, is different and does not include, for example, participatory
approaches. This corroborates with some of the major challenges identified in the elabor-
ation and implementation of the Master Plan. One of these challenges is the conflict of
interests among stakeholders, such as conflicts between hunting and sport fishing associa-
tions and ecologists for the right to exploit natural delta resources (Ştiuca and Nichersu
2006). The lack of involvement of citizens’ points of view in management was also iden-
tified as an important challenge because of the little trust in local and regional policies,
resulting in a lack of awareness and lack of real opportunities for citizens to participate
(Van Assche et al. 2011). To face these challenges, one of the principles of the Master
Plan is a decentralized management, involving all the stakeholders and equilibrating the
local interests with the interests of all the population. To further support the existing
Master Plan, the Ministry of Regional Development and Public Administration (MRDPA)
of Romania initiated, in 2013, the Danube Delta Integrated Sustainable Development
Strategy for 2030, by signing an agreement with the World Bank (World Bank 2014a).
One of the identified pillars of future planning aims at protecting environmental and nat-
ural resource assets through empowerment of local communities. This should be achieved
by scientifically guided strategic interventions, such as ecological restoration, through
hydrological and hydraulic modelling and strengthening flood protection and monitoring
systems. These would lead to a better understanding of the delta’s hydrodynamics and
achieve the planning objectives. Therefore, in line with the official Master Plan of 2005,
the main priority actions for this pillar include financial and technical support for monitor-
ing, system modelling and implementation of management activities for re-naturalization
and conservation of the ecosystem (World Bank 2014b).
In the realm of research, a Strategic Research and Innovation Agenda (SRIA) for
the Danube – Black Sea Region (DBS), was proposed by Gault et al. (2015), after
consultation with regional stakeholders. The strategic research looks for short term
innovations that will result in climate change and mainstream policy adaptation solu-
tions, in order to achieve the long-term objectives of the strategic delta plan. In this
context, one of the research priorities acknowledged is the need for citizen observato-
ries to promote environmental conservation.
Based on the strategic Danube Delta planning, it is clear that, although challenging,
the implementation of citizen observatories is welcome and can potentially contribute
towards many aspects. The main purpose is to contribute to environmental conserva-
tion of the delta area, as foreseen in the Master Plan, by better understanding the flow
in the delta through monitoring stagnation zones. This is followed by the need for data
to build and maintain the desired flood models, identified as necessary tools for testing
management measures. A third purpose was identified during user-requirements
SCENT workshops: better understanding of flows in the commercially important Canal
Mila 35. These three purposes will be considered in the pathway selection approach
presented in Section 4.

2.3. Past modifications

The whole DD river network has been heavily modified, from 1889 till 1984, aiming
at improving navigation, agriculture and fish production. The creation of artificial
channels, dams and embankments has resulted in increased discharge and changes in
1978 A. Venturini et al.

eutrophication rates (Oosterberg et al. 2000). In parallel with canalization of the area,
several areas were drained and made available for reed cultivation and intensive agri-
cultural production, which changed the way the ecosystem functions (IUCN East
European Programme 1992; Oosterberg et al. 2000; Bondar and Panin 2000). One
such example is the Sireasa Polder.
An important construction in the Sontea Fortuna study area is the Mila 35 canal,
built in 1984 for navigation purposes. Mila 35 is the main water supplier for Sontea-
Fortuna, responsible for modifying the natural hydro-morphological system of the
region due to the process of erosion and sediment accumulation caused by high dis-
charges (Popescu et al. 2015).
Referring to the entire DD, Bondar and Panin (2000) argued that all these modifi-
cations drastically changed the regime of the Danube Delta, the water volume being
renewed now three times a year, as opposed to once a year in the past. However, the
renewal occurs mostly in the main channels.
Pringle et al. (1995) summarized the deterioration of the environmental status of
the ecosystems belonging to the DD in terms of four main causes. Two of these rea-
sons are based on the high nutrient loads entering the Danube River and the loading
coming from agricultural activities inside the delta. The last two causes are both based
on hydraulic changes, such as removal of river floodplains along the Danube River
and canal dredging and impoundments in the Danube Delta. These modifications have
reduced the rivers’ capacity for self-purification and removal of nutrients and pollution,
which has led to the need to develop a plan for sustainable management of the
DD area.

3. Model development
A hydrodynamic model has been developed for the study area in order to obtain inun-
dation and flow patterns for different hydrological regimes (wet, dry and average),
which are being used to identify stagnation points and accessibility to points
of interest.
The detailed hydrodynamic model for the Sontea-Fortuna area was developed using
the USACE modelling software for river analysis, HEC-RAS. The concept of the
model, a one-dimensional representation of the river network, with lakes in the area
being represented as storage, defined with depth-volume relationships, can be seen in
Figure 2. This setup was appropriate for complex systems formed by connected chan-
nels and lakes, such as Sontea-Fortuna. A total of 341 cross-sections were used to rep-
resent 340 km of main delta branches; and 230 km of internal canals. The DD system
is comprised of canals and junctions. Canals are river segments located between two
junctions, the locations where canals intersect/connect. There were 101 canals and 72
junctions defined in the model. Additionally, the storage of the 11 lakes included in
the model was computed based on bathymetries of 5 m  5 m cell size resolution.
Simulations were made for a period of 2 months for three different flow regimes,
by setting different Danube River upstream boundary conditions. These three boundary
conditions were modelled as time series of constant discharge with minimum
(2,300 m3 s1), average (7,000 m3 s1) and maximum (15,800 m3 s1) computed from
historical recorded values corresponding to the three selected flow regimes. The three
main branches of the model were extended to the Black Sea where downstream bound-
ary conditions of 0.6 m above sea level were set for all three main branches, as
 Journal of Environmental Planning and Management 1979

Figure 2. Conceptual representation of the hydrodynamic model.

constant water level, given that the tidal variation at the Black Sea at this location
is negligible.
Model calibration was performed under steady state conditions, given that only
limited single time discharge data were available at several locations, only for the dry
period of 2015. Manning’s roughness coefficient at 11 locations was adjusted. Model
performance was measured by the Root Mean Square Error (RMSE). Model validation
was subsequently performed at six additional locations. The calibration and validation
resulted in a RMSE of 1.4 m3 s1 for a constant inflow discharge of 3,035 m3 s1 and
in a RMSE of 22.6 m3 s1 for a constant inflow of 2,380 m3 s1, respectively. The
highest values of Manning’s coefficient were found in the small canals inside the
Sontea-Fortuna area, because these canals are known for not being maintained and the
grown vegetation reduces the velocity of the flow. It is worth mentioning, that because
of limited data available for calibration and validation, the confidence in the model
results is limited, but it could still be used for demonstrating the methodology for path-
way selection.

4. Pathway selection approach

This approach was developed to support citizen observatories in delta areas, in the
context of delta planning, to define which boat routes to follow when citizens partici-
pate in data collection organised campaigns. The approach is not specific to the
Danube Delta case study, and is therefore presented in generic terms. Its applicability
is tested and demonstrated in the Sontea-Fortuna case study.
The main principle of the proposed approach is to combine the interest of local
stakeholders and the characteristics of the study area, as presented in Figure 3. Based
on analysis of these two aspects, possible pathways can be identified and prioritized.
Then, the data collected by citizens using the selected pathway(s) can be used (e.g. in
flood modelling) and the knowledge gained can indicate new data needs. There is also
experience gained that helps in adjusting pathway parameters (e.g. time needed to
make an observation). Such feedback mechanisms, although indirect, enable the cre-
ation of new campaigns, which in turn can create more sustained citizen engagement
in the process. In other words, we propose to cyclically re-evaluate actor coalition
interests and redesign campaigns, not only to continually gather data but to sustain par-
ticipation. There are other aspects that should be thought of, such as communication of
results and campaign frequency, but these are outside the scope of this article.
Figure 3 shows the application steps. The first step of the approach is to define the
purposes for data collection based on delta planning needs. Based on these purposes, it
1980 A. Venturini et al.

Figure 3. Pathway selection approach concept.

was necessary to identify the location in space (points of interest) where data can be
collected. To prioritise pathways, each group of points of interest related to a specific
purpose (named “location set”) was attributed with a score, depending on its import-
ance. Thus, each point of interest (in this case reach of interest) accumulates a score
(reach score), according to the purposes it serves.
For the Sontea-Fortuna area, investigation of stagnant water was set as a purpose,
hence a “location set” with all canals showing model results with stagnant water was
defined. Canal scores were selected to vary from 0 to 4, from the least stagnant to the
most stagnant canal. The canals that were mostly affected by water stagnation are pre-
sented in Figure 4. For the purpose of improving model accuracy, two locations sets
were defined. One set is composed of the canals located in the downstream section, in
order to enable the model to be limited to the Sontea-Fortuna. To these canals, a score
of 2 was given. The second location set included observations in all canals, also with
a score of 2, because they provide data for model calibration. For the purpose of inves-
tigating engineering modifications, the canal Mila 35 was selected and a score of 2
was attributed to it. All the scores were set according to the ranking, as discussed by
stakeholders during user requirement meetings. A summary of the scoring approach is
presented in Table 1. By summing up scores, the overall score for each canal, accord-
ing to data collection purposes, is presented in Figure 5.
In the following step, parameters for the pathway design are defined and an acces-
sibility analysis is conducted, both depending on the case-study characteristics. The
parameters are: the maximum available time for a boat trip (Tmax ); time for citizens to
collect data (To ) and velocity of the boat (v). The time spent on the observation itself
is dependent on the type of sensor used and the velocity of the boat depends on the
boat, on the flow and the navigation regulations in each canal. Apart from the three
mentioned parameters, the number of observations taken along a pathway is defined as
n and it is comprised of all the observation points along the main pathway, plus a
number of important points located outside the main pathway, but not too far. The
minimum distance from the main pathway to the outside observation point
 Journal of Environmental Planning and Management 1981

Figure 4. Rivers and canals experiencing water stagnation in Sontea-Fortuna. Note: Lakes are
not considered stagnant water.

Table 1. Scoring of selected purposes of data collection in the case study.

Investigate
Investigate Engineering
Purposes water stagnation Improve model accuracy modifications

Location Stagnant canals Downstream All canals (model Canal Mila 35

Sets BC canals calibration)
Scores 0–4 2 2 2

(Doutmin Þ needs to be defined in order to ensure correctness of the collected data. For
example, if velocity data is going to be collected in an observation point that is located
away from the main pathway, the boat needs to travel to a point where velocity is not
affected by the velocity on the main pathway.
For the present case study, a minimum distance of 500 m was defined for Doutmin .
To determine the maximum time possible for one trip, it was assumed that the boat
travels during working hours, from 9:00 AM until 5:00 PM; hence the boat should
return to the starting point of the trip within eight hours (Tmax ). Considering that, in
this case study, observations are performed by taking pictures and videos using a
smartphone, it is assumed that for each individual observation a time frame of
5 minutes is needed (To ). Looking into the DDBR navigation regulations4, it was found
that the maximum allowed navigation speed is 40 km h1 or 15 km h1, depending on
the size of the canal. Canals located near colonies of birds have a maximum speed of
5 km h1. Because the main part of the study area is located in small canals, the
adopted boat speed for calculation of pathways was 15 km h1 (v).
Accessibility analysis is performed to take into account all existing constraints in
the area. The first constraint is the non-accessibility of boats in the canals due to low
water level for navigation. In Sontea-Fortuna, a standard boat size with a maximum
capacity of 35 passengers was considered. The draft for this boat, which is the vertical
distance between the waterline and the bottom of the boat, was 0.5 m. This draft was
1982 A. Venturini et al.

Figure 5. Final scores for data collection and starting points for pathways.

then compared to the water depth map results simulated for the dry, average and wet
scenarios, resulting in different accessibility for these hydrological regimes. A second
constraint took into account strictly protected areas where access is prohibited. Inside
the study area, there is only one such lake, the Nebunu Lake.
The third step of the approach is the actual generation of pathways. A pathway is
generated when a boat departs from a start/end point, and stops for observations to be
taken along the navigated canals. As the overall time varies from one pathway to
another, Equation (1) was proposed to estimate the time covered for each generated
pathway. The time to navigate the pathway was considered as the sum of the overall
time spent with observations plus the boat travel time. The time of each pathway
should be equal to or smaller than the predefined available time, as shown in Equation
(2). The observation distance outside the main pathway should be equal to or greater
than the minimum distance for accurate data collection, as presented in Equation (3).
LP
Tpath ¼ n
T0 þ (1)
v
Tpath  Tmax (2)
Dout  Doutmin (3)
where:Tpath ¼ pathway time [h]; n ¼ number of observations; T0 ¼ time of observation
[h]; LP ¼ pathway length [km]; v ¼ boat velocity [km/h]; Tmax ¼ maximum available
time [h]; Dout ¼ observation distance outside the main pathway; Doutmin ¼ minimum
observation distance outside the main pathway
If the maximum available time constraint is violated, that pathway is rejected and a
new pathway is generated. A pathway score is obtained by summing up the scores of
the canals where observations were carried out (Equation 4).
X
n
Spath ¼ Si (4)
i

where Spath ¼ pathway score; Si ¼ total score for an individual canal where the observa-
tion is carried out.
 Journal of Environmental Planning and Management 1983

Following the above procedure, a set of possible pathways is generated. Then,

pathways can be prioritized by ranking them according to their scores, and field cam-
paigns can be set up for data collection.

5. Results and discussion

To investigate the applicability of the proposed approach, an analysis of the generated
pathways for the three proposed hydrological regimes was performed. The analysis
helped in better understanding the spatial distribution of the pathways (the accessible
ones) and the resulting scores.
For the dry scenario, there were 27 canals with a total length of 56 km (out of
407 km) that were not accessible. Additionally, 7 out of 11 lakes located on the west
side of the study area were not accessible due to non-navigability of the access canals
(Figure 6). On the other hand, in the average and wet scenario, all canals were access-
ible, except for the protected area of Nebunu Lake. The accessibility analysis showed
that there were plenty of pathways for citizen data collection, even during the
dry scenario.
For the boat trip starting and ending in Tulcea city, four pathways were proposed.
Their characteristics are presented in Table 2. These scores were generated to assess
the pathways’ performance according to the distance to the start/end point, number of
observations taken and the considered hydrological regime. In Figure 7, pathways 1
and 2 are shown, where the boat trajectory passed through Canal Mila 35 and sur-
rounding Sontea-Fortuna wetland. The difference between Pathway 1 and Pathway 2
only is that in the later were included observation points that were outside of the main
pathway track. For Pathway 1, the number of observations was lower than Pathway 2,
and so was its score. Conversely, the time needed for the second pathway is larger
than the maximum time imposed for data collection. For that reason, a balance
between the maximum time constraint and the number of observations should be
sought. To carry out observations in nearby canals was a good approach, because it

Figure 6. Accessibility map for dry scenario.

1984 A. Venturini et al.

Table 2. Pathways summary.

Pathway Scenario No. observations T path Spath

1 Dry 21 7:29 52
2 Dry 39 10:17 99.5
3 Dry 32 7:57 78.6
4 Wet 33 7:57 90.7

Figure 7. Pathway 1 (left) and 2 (right) from Tulcea for dry hydrological regime.

allows for collection of more data, however, it might not fit the time constraint
because it increases the time spent in observations and in navigation. Pathway 1 would
increase its score to, at most, 70 by adding three observations (the only ones possible
within the constraint). Another aspect revealed by the analysis was that the scores of
the canals play an important role for prioritizing some pathways when time is limited.
In Figure 8, the third and fourth pathway were proposed for dry, average and wet
regimes. Because the average and wet regimes show the same path, they are noted as
average/wet. By comparing pathways 1 and 2 to pathways 3 and 4, it could be noticed
that when the pathway is short (close to the start/end point) the number of observations
will most probably increase, because, as the distance travelled decreases, the time
spent in travel from one observation point to the other decreases and, consequently,
there is more time to collect data. Comparing Pathway 3 to Pathway 4, the reason for
a higher score of pathways with the same amount of time was the fact that the first
collects data in canals that were not accessible in the dry scenario. Therefore, a path-
way that gets the best score in one scenario might not be the same for
another scenario.
In the interests of having an overall visualization of the pathway distribution over
the canals for different start/end points, a map including one pathway, was proposed,
for each city/village (Figure 9), strategically located in the study area and with a crit-
ical number of citizens living there. All six pathways were determined with the aim of
achieving high scores within the time constraint. It could be observed that the path-
ways with highest scores had their start/end points in Tulcea and Mila 23. Both of
them included 33 observations and the reason for a high score for the Tulcea pathway
was the fact that it passed through canal Mila 35, St. Gheorghe branch (downstream
BC) and four canals which had high stagnation. The Mila 23 pathway also had a high
score because it covered three downstream BC and four canals with high water
 Journal of Environmental Planning and Management 1985

Figure 8. Pathway 3 (left) and 4 (right) for dry and average/wet hydrological regime
respectively.

Figure 9. Overall view of pathways assessed for all start/end points for average/wet
hydrological regime.

stagnation. In addition, Chilia Veche was the worst scoring start/end point because
larger distances needed to be covered to reach more observation points. However,
Chilia Veche is the second most populated city in the area and this variable should be
considered in further analysis.
The already existing local touristic boat routes were also taken into consideration
for the pathway analysis. Assessing the three routes located in the study area showed
that none of them was able to fit the time frame of 8 h. The one with the best score
covers 21% of all canals, with 25 observation points. The pathway selection approach
is a contribution that was missing in citizen observatories research. In other fields,
such as transportation systems, there is the definition of routes to be followed by
vehicles such as cars and there are optimization algorithms to minimize costs and time
(Clarke and Wright 1964). Despite similarities, there are important differences to citi-
zen observatories pathways that hinder direct application, mainly in terms of such
methods, objectives and level of complexity. The objective of the pathway for citizen
observatories is to collect data at selected points, which, in this case study, includes
1986 A. Venturini et al.

more than 110 points, whilst in transportation problems, there are few tar-
geted locations.
Improvements to the proposed pathway selection approach can be made in different
directions. This study considered only an average boat velocity for all canals and one
type of sensor (mobile phone). When the speed limit is available on each canal and
more than one sensor is planned to be used, variable velocities and time for observa-
tion should be considered. The scale ranges of the proposed scores need to be further
tested regarding their effectiveness, depending on the purposes of data collection.
Finally, the pathway selection approach can be posed and solved formally as a math-
ematical optimization problem, in which the pathway score is maximized, constrained
by the accessibility and maximum available time.
The proposed approach responds to the need for collecting data through organised
citizen observatory campaigns at stakeholders’ request; therefore insights on citizens’
opinions on pathways to be followed, or cultural differences, are not yet available.
Local settings were embedded in defining the start/end points of the routes. From a
scientific perspective, it is advisable that citizens participate and/or collaborate in the
design and development of data collection campaigns (Wehn and Evers 2015).
Therefore, research could better consider citizen preferences and/or to reflect on these
modelling outputs’ results for the next workshops.

6. Conclusions
The Danube Delta is a large, complex dynamic ecosystem where decision makers have
considerable interest in creating and adapting policies for environmental protection.
Data gathering and sharing is important for better understanding and management of
the system. This aspect has been identified as a weakness in the Danube Delta
Integrated Sustainable Development Strategy. Citizen observatories bring opportunities
to complement the traditional data collection efforts. Moreover, the involvement of
citizens supports one of the necessities identified in the 2005 Master Plan report
regarding recognition of the importance of the delta environment to local inhabitants.
Engaging citizens in data gathering activities is most effective when carefully
planned. Such plans may include determination of pathways (routes) to be followed,
which enable gathering of sufficient data and sustain citizens’ engagement. To this
end, in this study a hydrodynamic model of the Danube Delta was built, calibrated and
validated with the available data. Based on the model results regarding water stagna-
tion and boat accessibility for different hydrological regimes, a pathway selection
approach was introduced, assuming that data gathering activities were limited to
acquiring images and videos by mobile phones. Another important aspect of the
approach was the consideration of local stakeholders’ interests.
By applying the developed pathway selection approach in the Sontea-Fortuna area
of the Danube Delta, it was found that the wetland is more accessible than expected,
for all scenarios, and that there are many possible pathways for points of interest, even
when only one city is involved in the campaigns.
The proposed pathway approach was robust, attributing distinct scores to different
pathways. The selection of pathways was dependent on the number of observations
collected, which impacts significantly both pathways’ time and score. The analysis
showed that the best pathways are located relatively close to their start/end points and
that they can be associated with different start/end points in different hydrological
 Journal of Environmental Planning and Management 1987

regimes. This information also supports that more than one start/end point needs to be
considered when planning citizens’ observatory campaigns in the case study area. Use
of already existing touristic routes might be a good alternative due to regular accessi-
bility; however, these routes covered a small part of the study area, not including
observations in interconnected and water stagnant canals.
The pathways will later be implemented in organized field experiments, planned
within the SCENT project. Given positive results the same approach can be expanded
to other environmental problems in the region, such as water quality or ecosystem
assessment, contributing to wider strategic delta planning.
The approach can be applied in other citizen observatories, having the advantages
that it makes clear why and where to collect the data; how to get there, considering
different hydrological regimes; and which pathways to prioritize to maximize gains.
Specifically, the approach is replicable in deltas, wetland areas and other regions with
an intricate network of channels, where it is challenging to know where to go.
Moreover, we envisioned that the approach can be adapted to terrestrial pathways, e.g.
hiking routes, to collect, for example, soil moisture data from low-cost sensors.

Notes
1
The term ‘Hourglass framework’ can be found the literature with various meanings. In this
study, when the term was used, it refers specifically to the version of Seijger et al. (2017).
2
The term ‘actor coalitions’ is used in this study, exactly as defined by Seijger et al. (2017), to
refer to ‘groups of individuals and organizations that unite in a planning process’.
3
https://scent-project.eu.
4
http://gov.ro/en/news/new-traffic-rules-on-danube-delta-inland-channels-and-lakes.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
The research leading to these results has received funding from Horizon 2020, the European
Union’s Framework Programme for Research and Innovation (H2020/2014-2020) under grant
agreement no 688930 - SCENT project.

ORCID
I. Popescu http://orcid.org/0000-0002-1009-5424
A. Jonoski http://orcid.org/0000-0002-0183-4168
D. P. Solomatine http://orcid.org/0000-0003-2031-9871

References
Albrechts, L. 2004. “Strategic (Spatial) Planning Re-examined.” Environment and Planning B:
Planning and Design 31 (5): 743–758. doi:10.1068/b3065.
Arega, F. 2013. “Hydrodynamic Modeling and Characterizing of Lagrangian Flows in the West
Scott Creek Wetlands System, South Carolina.” Journal of Hydro-Environment Research 7
(1): 50–60. doi:10.1016/j.jher.2012.11.001.
Assumpç~ao, T. H., I. Popescu, A. Jonoski, and D. P. Solomatine. 2018. “Citizen Observations
Contributing to Flood Modelling: Opportunities and Challenges.” Hydrology and Earth
System Sciences 22 (2): 1473–1489. doi:10.5194/hess-22-1473-2018.
1988 A. Venturini et al.

Bondar, C., and N. Panin. 2000. “The Danube Delta Hydrologic Database and Modelling.” Geo-
Eco-Marina 5 (6): 5–52.
Busch, J. A., I. Price, E. Jeansou, O. Zielinski, and H. J. van der Woerd. 2016. “Citizens and
Satellites: Assessment of Phytoplankton Dynamics in a NW Mediterranean Aquaculture
Zone.” International Journal of Applied Earth Observation and Geoinformation 47: 40–49.
doi:10.1016/j.jag.2015.11.017.
Buytaert, W., Z. Zulkafli, S. Grainger, L. Acosta, T. C. Alemie, J. Bastiaensen, B. De Bievre.,
et al. 2014. “Citizen Science in Hydrology and Water Resources: Opportunities for
Knowledge Generation, Ecosystem Service Management, and Sustainable Development.”
Frontiers in Earth Science 2: 1–21. doi:10.3389/feart.2014.00026.
Cioca, E., C. Bondar, and C. Borcia. 2010. “Hydrographical Network of the Danube Delta
Biosphere Reserve: Modelling the Morphological Dynamics.” Paper presented at the 38th IAD
Conference, Dresden, Germany, June 22–25. https://www.researchgate.net/publication/
318722483_Hydrographical_network_of_the_Danube_Delta_Biosphere_reserve_-_Modelling_
the_Morphological_Dynamics.
Clarke, G., and J. W. Wright. 1964. “Scheduling of Vehicles from a Central Depot to a Number
of Delivery Points.” Operations Research 12 (4): 568–581. doi:10.1287/opre.12.4.568.
DDNI and IVL. 2006. Master Plan: Support for Sustainable Development in Danube Delta
Biosphere Reserve/Tulcea County (Romania) Logical Framework Analyse (LFA),
Stockholm: Swedish Environmental Research Institute. Accessed November 29, 2017, http://
breiling.org/kidd/upload/development_plan_danuba_delta.pdf
Gault, J., C. Bradley, A. Tyler, A. Stanica, V. Papathanassiou, and G. Gettel. 2015. Towards the
Integrated Management of the Danube River – Danube Delta – Black Sea System: Proposal
for a Strategic Research and Innovation Agenda. Bucharest: GeoEcoMar.
Gharesifard, M., U. Wehn, and P. van der Zaag. 2017. “Towards Benchmarking Citizen
Observatories: Features and Functioning of Online Amateur Weather Networks.” Journal of
Environmental Management 193: 381–393. doi:10.1016/j.jenvman.2017.02.003.
Giosan, L., J. Syvitski, S. Constantinescu, and J. Day. 2014. “Climate Change: Protect the
World’s Deltas.” Nature 516 (7529): 31–33.
G€uttler, F. N., S. Niculescu, and F. Gohin. 2013. “Turbidity Retrieval and Monitoring of
Danube Delta Waters Using Multi-Sensor Optical Remote Sensing Data: An Integrated
View from the Delta Plain Lakes to the Western-Northwestern Black Sea Coastal Zone.”
Remote Sensing of Environment 132: 86–101. doi:10.1016/j.rse.2013.01.009.
IUCN East European Programme. 1992. Environmental Status Reports: Conservation Status of
the Danube Delta. Cambridge: IUCN Publications Unit.
Klonner, C., S. Marx, T. Uson, J. P. de Albuquerque, and B. H€ofle. 2016. “Volunteered
Geographic Information in Natural Hazard Analysis: A Systematic Literature Review of
Current Approaches with a Focus on Preparedness and Mitigation.” ISPRS International
Journal of Geo-Information 5 (7): 103. doi:10.3390/ijgi5070103.
Mazzoleni, M., M. Verlaan, L. Alfonso, M. Monego, D. Norbiato, M. Ferri, and D. P.
Solomatine. 2017. “Can Assimilation of Crowdsourced Data in Hydrological Modelling
Improve Flood Prediction?” Hydrology and Earth System Sciences 21 (2): 839–861. doi:
10.5194/hess-21-839-2017.
McKinley, D. C., A. J. Miller-Rushing, H. L. Ballard, R. Bonney, H. Brown, S. C. Cook-Patton,
D. M. Evans., et al. 2017. “Citizen Science Can Improve Conservation Science, Natural
Resource Management, and Environmental Protection.” Biological Conservation 208: 15–28.
doi:10.1016/j.biocon.2016.05.015.
Millennium Ecosystem Assessment. 2005. Ecosystems and Human Well-Being: Wetlands and
Water – Synthesis. Regions and Cohesion. Washington, DC: Island Press.
Oosterberg, W., M. Staras, L. Bogdan, A. D. Buijse, A. Constantinescu, H. Coops, J. Hanganu,
et al. 2000. Ecological Gradients in the Danube Delta Lakes: Present State and Man-Induced
Changes. RIZA rapport, 2000.015. IAPN 90.369.5309x. Lelystad: Institute for Inland Water
Management and Waste Water Treatment RIZA.
Popescu, I., E. Cioaca, Q. Pan, A. Jonoski, and J. Hanganu. 2015. “Use of Hydrodynamic
Models for the Management of the Danube Delta Wetlands: The Case Study of Sontea-
Fortuna Ecosystem.” Environmental Science and Policy 46: 48–56. doi:10.1016/
j.envsci.2014.01.012.
 Journal of Environmental Planning and Management 1989

Poser, K., and D. Dransch. 2010. “Volunteered Geographic Information for Disaster
Management with Application to Rapid Flood Damage Estimation.” Geomatica 64: 89–98.
Pringle, C., G. Vellidis, F. Heliotis, D. Bandacu, and S. Cristofor. 1995. “Environmental
Problems of the Danube Delta.” Ekistics 62: 144–340.
Seijger, C., W. Douven, G. van Halsema, L. Hermans, J. Evers, H. L. Phi, M. F. Khan., et al.
2017. “An Analytical Framework for Strategic Delta Planning: Negotiating Consent for
Long-Term Sustainable Delta Development.” Journal of Environmental Planning and
Management 60 (8): 1485–1509. doi:10.1080/09640568.2016.1231667.
Ştiuca, R., and I. Nichersu. 2006. “Master Plan: Support for Sustainable Development in
Danube Delta Biosphere Reserve/Tulcea County (Romania) Logical Framework Analyse
(LFA).” Paper presented at the 36th IAD conference, Vienna, September 4–8.
Tserstou, A., A. Jonoski, I. Popescu, T. Herman Asumpcao, G. Athanasiou, A. Kallioras, and I.
Nichersu. 2017. “SCENT: Citizen Sourced Data in Support of Environmental Monitoring.”
Paper presented at the 21st International Conference on Control Systems and Computer
Science, CSCS 2017, Bucharest, Romania, May 29–31. doi:10.1109/CSCS.2017.93.
Van Assche, K., M. Duineveld, R. Beunen, and P. Teampau. 2011. “Delineating Locals:
Transformations of Knowledge/Power and the Governance of the Danube Delta.” Journal of
Environmental Policy and Planning 13 (1): 1–21. doi:10.1080/1523908X.2011.559087.
Wehn, U., and J. Evers. 2015. “The Social Innovation Potential of ICT-Enabled Citizen
Observatories to Increase eParticipation in Local Flood Risk Management.” Technology in
Society 42: 187–198. doi:10.1016/j.envsci.2014.12.017.
World Bank. 2014a. A Vision Statement for the Danube Delta Region (2030): Danube Delta
Integrated Sustainable Development Strategy. Washington, DC: World Bank Group. http://
www.mdrap.ro/userfiles/delta_dunarii/rezultate_proiecte/2_Raport_Viziune_en.pdf
World Bank. 2014b. Report 2.2 Draft Danube Delta Integrated Sustainable Development
Strategy (2030). Washington, DC: World Bank Group.
Zaman, J., E. D'Hondt, E. G. Boix, E. Philips, K. Kambona, and W. De Meuter. 2014. “Citizen-
Friendly Participatory Campaign Support.” Paper presented at the IEEE International
Conference on Pervasive Computing and Communication Workshops (PERCOM
WORKSHOPS), Budapest, Hungary, March 24–28. doi:10.1109/PerComW.2014.6815208.