Knowledge Driven Community Self-reliance and Flood Resilience
Study of the Communities in the Lower Sava Valley, Slovenia
Jernej Agrež and Nadja Damij
Faculty of Information Studies in Novo mesto, Ulica Talcev 3, Novo mesto, Slovenia
Keywords: Experiential Learning, Community Learning, Floods, Loosely Coupled Systems.
Abstract: In this position paper, we focus on the learning issues of the flood-endangered communities, situated in the
Lower Sava Valley. The main issue and position expose characteristics of the loosely coupled system in the
context of learning and knowledge management. We detected learning anomalies within the assessed flood
response system, which we selected as a research example of the loosely coupled system. Therefore, we
conducted a statistical analysis of the flood data and designed a fuzzy knowledge evaluation system, to be
able to grade community learning. Finally, we designed learning improvement mechanism that would
significantly contribute to the flood response system effectiveness and higher community self-reliance and
flood resilience.
1 INTRODUCTION
Floods in Slovenia became a constant threat during
the autumn and spring periods in the past few years.
Even though devastating floods used to affect the
Lower Sava Valley only ones per 100 years,
communities in the region faced four floods between
2010 and 2014. Comparing the four floods by the
ferocity and hydrological data, two of four floods
could be titled as the “hundred years waters” – very
devastating floods.
The flood response system consists of
professional and voluntary response units including
civil protection, fire brigades, military, police and
other technical services. The system emerges ad hoc,
when the flood forecast announces alarming flood
possibilities. Overall system, which is directly
influenced by the floods, exists of entities in need for
protection and entities that provide necessary support
in order to minimize possible life loss and damage to
property.
Within the system, there is limited information
flow and no traceable knowledge flow that would
include all entities` fractions of the system. On the
first hand, flood responding entities possess a wide
range of explicit knowledge how to react during the
distress situation, how to use the equipment, how to
organize the work, etc. Such knowledge is a part of
several formal and tight subsystems, which are
capable of autonomous operating, even though during
the floods, they merge in the overall response system.
On the other hand, entities in need of support, that
form communities, possess mostly tacit knowledge,
gained through the participation in the events during
one or more floods.
The overall flood protection system contains a lot
of explicit knowledge, but without open access to all
entities included. It contains also constantly updated
tacit knowledge, but with no proper mechanism to
gather and codify it. Consequently the overall system
produces flood response process, which is not capable
of reaching its optimum, due to lack of efficient
knowledge management approach.
2 ISSUE AND POSITION
Formal and successful organizational system, as
described by Ionita (2011), is strongly dependant on
its business process architecture. Business process
architecture enables execution of its daily activities
and consequently makes possible to reach the set
outputs. Even though such a system is tight and
stable, it must be capable of flexible adaptation when
necessary. Contrary to formal and modern
organizational system, a system that emerges during
floods is loosely coupled and faces an unstable
process architecture. Process architecture as a
descriptor of the system`s structure includes elements
like inputs, outputs, entities, activities, procedures,
Agrež, J. and Damij, N..
Knowledge Driven Community Self-reliance and Flood Resilience - Study of the Communities in the Lower Sava Valley, Slovenia.
In Proceedings of the 7th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2015) - Volume 3: KMIS, pages 201-206
ISBN: 978-989-758-158-8
Copyright
c
 2015 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
201

information flows, etc. Within the loosely coupled
system, process architecture is subjected to dynamic
change, due to frequent and spontaneous interaction
among entities, information flows, and process
patterns. Consequently, there is no guarantee to
foretell, how the final process model will settle and
make process execution possible. Taking in the
consideration haste of the system`s emergence and its
short-term operation, such system evolves with a
tendency to reach a desired process output and
afterwards disintegrates into few tight organizational
systems.
Niu (2010) advocates the significance and the
influence in the relation between the knowledge and
the system`s adaptation. According to Martinez-Leon
and Martinez-Garcia (2011) less formal and less
centralized organizational systems enhances the
organizational learning process. Open, less rigid,
loosely coupled organizational system on the one
hand creates an open environment that encourages
organizational learning, but on the other hand creates
also a need to assess more complex, less transparent
and harder to follow learning process. Tennant and
Fernie (2013) found learning within the loosely
coupled system similar to the underdeveloped
knowledge management approach in the industrial
enterprises. In both cases, learning adapted to the
process and changed with activity flows in a
reactionary and interventionary manner. Firestone
and McElroy (2004) argue that rapid change in the
process architecture not only boosts new variants of
work processes, but also learning processes and
processes for managing knowledge.
Even though several authors detected and
described learning processes within the loosely
coupled system, we found no tangible and wide
knowledge interaction within the assessed flood
response system. We detected two different learning
processes with no traceable interaction. First learning
process occurs within formal and tight subsystems in
the loosely coupled system. Learning takes place
within the scope of the subsystem before its
integration into the overall flood response system.
The knowledge gained through such a learning
process is explicit, specific and differs on the kind of
the learning subsystem. No traceable interchange
among entities with such knowledge or other entities
is detected. Second learning process occurs when the
flood endangered communities face direct flood
threat. They are subjected to the experiential learning,
emerging tacit knowledge about one or several
floods. There is no traceable knowledge interchange
among flood-endangered communities and among
other entities, as well.
Even though the Resolution of national security
strategy of the Republic of Slovenia (2010) and the
Resolution of the national program of protection
against natural and other disasters from year 2009 to
year 2015 (2009), contain guidelines which would
practically establish knowledge interchange between
both groups, no such attempt has been recorded yet.
Implementation of a knowledge interchange
mechanism would on the first hand enabled the
smooth transfer of the knowledge among different
entities within the system, and on the other hand, it
would significantly optimize the flood response
process, executed by the loosely coupled system.
3 BACKGROUND
INFORMATION
To be able to understand how communities in the
flood response system perceived floods and how they
learned from them and about them, we collected
general information about the flood threatened area,
together with hydrological and meteorological data,
describing all 4 floods.
3.1 General Information
A good part of a Lower Sava region occupies Krško-
Brežice field, which is a valley, surrounded by
Gorjanci Hills on the southern side and Posavje hills
on the northern side. Two bigger municipalities
(Krško and Brežice) are situated in the valley and one
smaller (Kostanjevica na Krki). There are five flood
sources in the valley. In addition to the Sava and the
Krka, as two major rivers, the streams that carry water
from the hills quickly grow into torrents with a
threatening power within few days of continuous rain.
The rain itself can cause considerable problems when
meteoric water starts to overwhelm the low
positioned planes with impermeable soil layers. The
communities and the infrastructure located low and
near the river can experience groundwater flooding,
which usually affects the underground parts of
buildings, such as basements, engine rooms, garages,
workshops, etc.
According to the Department for hydrological
prognosis of the Slovenian environmental agency
(2012), flooding of the Krka River in the
communities, in the municipality of Brežice, which
are located within the 8 km area before the confluence
with the Sava River, is highly dependent on the Sava
River and its flow rate. High flow rates of the Krka
River alone represent a threat to the western
KMIS 2015 - 7th International Conference on Knowledge Management and Information Sharing
202

communities, such as the town of Kostanjevica na
Krki in the municipality of Kostanjevica na Krki.
However, eastern communities from Cerklje ob Krki
to Krška vas may face high water levels but with no
severe consequences. There are two major reasons
behind such hydrological dynamics. Firstly, the town
of Kostanjevica na Krki is built on an island which
acts as a natural barrier against the Krka River flow.
At the same time, there is a large primeval forest to
the northeast of the Kostanjevica island, which acts as
a retention area during floods and prevents water
from draining out of the area. The second reason lies
in the ratio between the flow rates of the Sava River
and the Krka River once they exceed the average rate.
The power of the Sava’s flow starts to block the
Krka’s flow, drastically increasing the drainage
capacity of the latter. As a result, the Krka River
floods the communities that are located near its bed
and close to the confluence with the Sava River.
3.2 Flood Comparison
We compared hydrological and meteorological data
of four floods (period of four days) through the
multivariate analysis of variance and by the simple
quantitative comparison of variables. Table 1
Table 1: MANOVA, river flow data.
Jesenice
station,
Podbočje
station,
Širje station,
Hrastnik
station,
flow rate flow rate flow rate flow rate
Year
p
Wilk`s λ
F
Wilk`s λ
F
Wilk`s λ
F
Wilk`s λ
F
2010
0.05
0.463 0.594
0.5322 0.4262 0.4612 0.5991 0.4392 0.6626
2012 2010
0.1406 26.741 0.6564 0.2135 0.1752 22.127
0.05837
48.703
2013 2010
0.4884 0.5275 0.3651 0.9218 0.4267 0.7012
0.14
2.684
2014 2012
0.2091 18.656 0.7397 0.1184
0.357
0.9553
0.07379
42.283
2013 2012
0.9823
5,00E-04
0.009884
11.312 0.9175 0.0114 0.2932 12.656
2014 2013
0.2491 15.448
0.03464
64.582 0.4425 0.6528 0.4221 0.7159
2014
Table 2: MANOVA, rainfall data.
Bizeljsko Sromlje Brege Smednik
Year
p
Wilk`s λ
F-value
Wilk`s λ
F-value
Wilk`s λ
F-value
Wilk`s λ
F-value
2010
0.05
0.2164 1.8009 0.1809 2.1483 0.2095 1.8623 0.1953 1.9973
2012 2010
0.3229
1.11
0.3925 0.8169 0.3217 1.1155 0.3583 0.9499
2013 2010
0.3715 0.8962 0.5047 0.4877 0.6111
0.28
0.6233
0.261
2014 2012
0.4814 0.5452 0.2351 1.6483 0.4438 0.6488 0.2283 1.7019
2013 2012
0.5559 0.3777
0.239
1.6184 0.3442 1.0108 0.3301
1.075
2014 2013
0.9379 0.0065 0.7901 0.0758 0.5736 0.3441 0.6519 0.2195
2014
Kostanjevica Planina Cerklje ob Krki
year
p
Wilk`s λ
F-value
Wilk`s λ
F-value
Wilk`s λ
F-value
2010
0.05
0.2242 1.7355 0.2576 1.4859 0.2397 1.6131
2012 2010
0.5038
0.49
0.637
0.2406 0.4857 0.5342
2013 2010
0.6476 0.2255 0.7688 0.0925 0.8246 0.0525
2014 2012
0.2728 1.3869 0.2708 1.3994 0.3252 1.0987
2013 2012
0.3538 0.9687
0.337
1.0429
0.378
0.8711
2014 2013
0.8443 0.0412 0.8694 0.0288 0.6869 0.1747
2014
Knowledge Driven Community Self-reliance and Flood Resilience - Study of the Communities in the Lower Sava Valley, Slovenia
203

compares the Wilk`s Lambda values from flow rate
MANOVA testing. The results of the comparison of
flood events in the year 2012 vs. 2014 and 2013 vs.
2014 show significant difference only in the Podbočje
station flow rate. Furthermore, nearly significant
difference was identified with the regard to the
Hrastnik station flow rates in the year 2010 vs. 2013
and 2012 vs. 2013. There was no significant
difference in the comparison between 2010 vs. 2012
and 2010 vs. 2014.
We used also rainfall data, analyzed with the
multivariate analysis of variance and by quantitative
comparison. MANOVA (Table 2) revealed no
significant difference in the variance among 4 day
rainfall measurements during floods. Therefore, we
determined similarities, based on quantitative
comparison, which revealed similarities between the
events in 2010 vs. 2014 and 2013 vs. 2014.
3.3 Community Learning Evaluation
We integrated the results of the flood similarity, into
the fuzzy system, implemented in the R programming
environment. To be able to create an evaluation data
set, we gathered the emergency response data from
the database, governed by the Slovene Administration
for civil protection and disaster relief. We extracted
the data, which included flood response in
communities, during the floods, previously detected
as similar.
We divided the data into community sets. The
structure of every set was following: S = {A, B, C, D,
E}.
 Subset A represents the distress source,
 Subset B represents the number of distress cases
during the first flood,
 subset C represents the number of distress cases
during the second flood,
 subset D represents the primary response
activities during both flood events,
 subset E represents the secondary response
activities during second flood event.
The following stage of the evaluation included the
categorical evaluation in which learning performance
of the communities was measured using the condition
of the same distress source during both flood events,
and following five criteria:
 x1 ∈ B · x2 ∈ C ⇒ L1 = {1}, distress is detected
during both flood events;
 x1 ∈ B ⇒ L2 = {1}, distress is detected only
during the first flood event;
 x2 ∈ C ⇒ L3 = {1}, distress is detected only
during the second flood event;
 x3 ∈ D`· x3 ∈ D`` ⇒ L4 = {1}, primary response
activities are the same during both flood events
 x4 ∈ E`· x4∈ E`` ⇒ L5 = {1}, secondary response
activities are the same during both flood events.
The subset X = {x1, x2, x3, x4} represents data
typical for a single entity (communities are built out
of entities, which are in reality flood endangered
households), while the subset L = {L1, L2, L3, L4,
L5} represents its learning grades. To be able to place
the grades of the entities into the community
perspective, we summed them into community
subsets CS
n
= {
∑
1


,
∑
2


,
∑
3


,
∑
4


,
∑
5


} ⇒ CS
n
={CL1n, CL2n, CL3n, CL4n, CL5n}
and further weighted them with the weighting rules
presented in Table 3.
Table 3: Weighting criteria.
CL1n CL2n CL3n CL4n CL5n
Value ranks
Value
Weight
Value
Weight
Value
Weight
Value
Weight
Value
Weight
High rank
3 8
<=16
8
<=16
8
>10
15
>1
15
Middle rank
2 4
< 6
4
< 6
4
10 10
1
10
Low rank
1 1 1 1 1 1
<10
10
- -
The community learning analysis provided us with
the insight, how poor is the overall learning process
of communities about floods in order to protect
themselves and gain higher resilience. Out of 59
communities, only 10 were graded with the grade
higher than poor. Most communities, even though
they faced two similar floods in a period of four years,
have not used the first experience to learn about how
and when the flood would threaten them and how to
protect their property against rising water.
Those communities, which demonstrated learning
process, graded between good and excellent, gain
their knowledge based only on the experiential
learning. No other, more formal knowledge source or
learning process that would provide them with
knowledge that is more explicit, took place.
KMIS 2015 - 7th International Conference on Knowledge Management and Information Sharing
204

4 PROPOSED SOLUTIONS
REGARDING THE ISSUE
We propose a solution to the identified state, which
consists of the learning mechanism integration, based
on the guidelines, extracted from the national security
and natural disaster protection strategic documents of
the Republic of Slovenia.
4.1 Solution Methodology
We designed a simple simulation model, which acts
as the learning mechanism. It provides communities
with the knowledge interchange, that enables learning
and in response to the floods with higher self-reliance
and flood resilience.
We designed an optimization algorithm using R
programming environment. The algorithm uses for
the input, data, gathered from the Administration for
civil protection and disaster relief. Further, it divides
the data in 5 threat source categories, predicts whether
geolocation of the endangered entity faces flood risk
and calculates whether levels of river flows are below
or above the critical flooding level.
We integrated insights and guidelines from the
strategic documents through the discriminatory
criteria in the algorithm. The algorithm used the
discriminatory criteria to exclude from the emergency
response those entities, which would on the premise
of the learning and education capacity be capable to
provide self-reliance until the critical flood levels.
4.2 Results
With the simulated learning mechanism integration,
we reduced: communication time to 66,67%,
responders` travel distance to 43,28%, the number of
process architectures to 61,6%, the number of process
patterns to 68,33%, the number of activities for
66,13%, the number of entities in distress in 60,73,
total number of executing standard operating
procedures to 55,81% and the number of different
standard operating procedures to 63,74%.
T-test (Table 4) of the data revealed significant
differences between the state with the integrated
learning mechanism and without it. Therefore, it
confirms the importance of additional community
knowledge and importance of better, more systematic
approach to the flood response learning, to be able to
provide more effective flood response system and
gain higher community self-reliance and flood
resilience.
Table 4: t-test of the flood response process with and
without learning mechanism.
Paired t-test
Field of inquiry
AS-IS
process
state
TO-BE
process
state
t p Interp.
communication
time
Mean
32,80 21,86
4,99
0.0001
Significant
SD
27,94 23,85
responders` travel
extend
Mean
7,46 3,23
1,51 0.14
Not significant
SD
21,62
5,44
number of process
architectures
Mean
1,92 1,17
5,87
0.0001
Significant
SD
1,53 1,16
number of process
patterns
Mean
9,53 6,51
5,629
0.0001
Significant
SD
4,12 5,45
number of activities
Mean
37,63 24,88
3,824
0.00001
Significant
SD
40,28 36,13
number of entities
in distress
Mean
3,24 1,97
4,369
0.0001
Significant
SD
4,31 3,52
total number of
executed standard
operating procedures
Mean
5,10 2,85
4,407
0.00001
Significant
SD
6,63 4,79
number of different standard operating procedures
Mean
2,90 1,85
6,923
0.00001
Significant
SD
2,35 1,94
5 CONCLUSION
Overview of the as-is state referring to flood
preparedness of the communities in the Lower Sava
Valles in Slovenia reveals us low self-reliance during
flood events, which consequently leads to the lower
Knowledge Driven Community Self-reliance and Flood Resilience - Study of the Communities in the Lower Sava Valley, Slovenia
205

flood resilience. Communities dispose with the
situational knowledge, obtained through the
experiential learning. Learning evaluation clearly
revealed that such knowledge is insufficient in order
to provide solid flood resilience. Therefore,
optimization of the as-is state in the form of
simulation-applied strategic guidelines, revealed how
significantly better natural disaster educational
measures would improve flood self-reliance. Through
better education, communities would have the
capacity to establish own flood protection, based on
internal cooperation, using simple and easy available
material and technical resources. Consequently, the
official flood responding force could divide its
resources in a more optimal manner and relieve their
workload. The vast majority of the responding force
represents firefighters - volunteers. Even though,
their main mission is to intervene in fire incidents,
they became an operative force of local civil
protection establishments. Better educated and flood
prepared communities would create the possibility to
become a civil protection auxiliary response force in
the general natural disaster protection system.
ACKNOWLEDGEMENTS
This research has been financed by the Slovenian
Research Agency ARRS, Young researcher program,
ARRS-SP-2784/13
REFERENCES
Administration for Civil Protection and Disaster Relief,
2012. Velika Nevarnost Poplav! (5.11.2012).
Available: http://www.sos112.si/slo/clanek.php?
catid=27&id=5858 [2015, July 9]
Firestone, M. J. & McElroy M. W., 2004. Organizational
learning and knowledge management: the relationship.
The Learning Organization. 11(2/3): 177-184.
Ionita, C., 2011. The BWS Open Business Enterprise
System Architecture. Informatica Economică. 15(4):
155-166.
Martinez-Leon, I. & Martinez-Garcia, J., 2011. The
influence of organizational structure on organizational
learning. International Journal of Manpower. 32(5/6):
537-566.
National Assembly, Republic of Slovenia, 2010. Resolution
of national security strategy of Republic of Slovenia.
Uradni List RS. 92(7).
National Assembly, Republic of Slovenia, 2009.
Resolution of national program of protection against
natural and other disasters from year 2009 to year 2015.
Uradni List RS. 51(6).
Niu, K., 2010. Knowledge management practices and
organizational adaptation, Evidences from high
technology companies in China. Journal of Strategy
and Management. 3(4): 325-343.
Tennant, S. & Fernie, S., 2013. Organizational learning in
construction supply chains. Engineering, Construction
and Architectural Management. 20(1): 83-98.
KMIS 2015 - 7th International Conference on Knowledge Management and Information Sharing
206