Toward a resilient organization: The management of unexpected hazard

on the polar traverse

Aude Villemain

⇑

, Patrice Godon

University of Reims, Research Center on Work and Development (CRTD), Laboratory of Ergonomics, CNAM, Paris, France

Paul Emile Victor Institute (IPEV), France

article info

Article history:

Received 21 October 2015

Received in revised form 31 January 2016

Accepted 8 March 2016

Available online xxxx

Keywords:

Organizational resilience

Safety management

Polar conditions

Unexpected events

Proactive–reactive adjustments

abstract

The aim of this research is to understand the organizational resilience through the safety management

when unexpected events occurred, on an atypical transport environment, the polar traverse. Three polar

traverses were studied, one of which being a detailed case study. Thus, ethnological observations over

3 year periods from 2012 to 2015 (to understand the traverse logic, functioning through unexpected

event) and all-day interviews during a traverse (to understand actions and strategies of organizational

resilience to cope unforeseen events) were collected. The main results, from quantitative and qualitative

analysis, indicated (1) mechanical, organizational and both interventions allowed to face unexpected

incidents on the traverse, (2) great possibilities to take actions on the convoy organization enabled to

develop a pro-active management of the safety in alternation with reactive adjustments; (3) the impor-

tance was to preserve the machines functionality even if operators have to face environment hostility to

repair; and (4) the variation of the convoy organization was permanent in its whole even if the incidents

concerned only one road track. The strategies of organizational resilience building will be discussed in

this article, around the proactive–reactive management, the organizational dynamic, the risk evaluation,

and the risk taken to preserve the room manoeuver.

1. Introduction

The term « resilience » can sometimes, in certain cases, be

reserved for the management of unexpected disturbances « which

exceed the anticipated areas of adaptation » (Lundberg and

Johansson, 2006, 2007; Woods, 2006, 2009). A system is resilient

if workers adapt themselves by understanding the context in

which adaptation takes place. Adjustments are thus constantly

made by individuals and organization, even if they are more often

approximate rather than exact (Hollnagel, 2012). Every organiza-

tion is stretched to operate at its full capacity and, to be resilient,

a system needs to be able to anticipate whatever may happen,

monitor what is going on, respond effectively when something

happens, and learn from past experiences (Hollnagel, 2009;

Woods and Cook, 2002; Woods and Hollnagel, 2006). Conse-

quently, the organizational resilience strategies are questioned to

understand how a system could adjust itself to disturbances or

unexpected hazard.

One of the main objective of researches focused on the strate-

gies of organizational resilience is to understand the organizational

preconditions conducive to a safe performance (Pidgeon and

O’Leary, 2000). Some studies emphasized the need for proactive

measures in safety management, while proactive manner invested

in safety and resource allocations to safety improvement are key

factors in ensuring a resilient organization (Dekker et al., 2008).

Reactive adjustments are, by far, the most common ones. Short

terms responses are not enough to guarantee a system’s safety

and survivability. One reason for this is that the system can only

be prepared to respond to a limited set of events or conditions

and over a limited period of time. The reactive approach quickly

appears too restrictive (Daniellou et al., 2009; Hale and Heijer,

2006), as this new way of conceiving safety is of little interest if

it only reacts to events and does not anticipate them (Dekker,

2006; Hollnagel and Woods, 2006; Westrum, 2006). Consequently,

the proactive vision of resilience is therefore essential when aimed

at the prevention and adaptation of a system to changing condi-

tions prior to the occurrence of undesirable events (Hollnagel,

2006, 2008, 2009; Leveson et al., 2006; Morel et al., 2008;

Westrum, 2006; Woods and Hollnagel, 2006). Proactive adjust-

ments, however, mean that the system can change from a state

of normal operation to a state of readiness before something

http://dx.doi.org/10.1016/j.ssci.2016.03.008

⇑

Corresponding author at: University of Reims Champagne-Ardenne, UFR STAPS,

Campus Moulin de la Housse, Bât. 25, Chemin des Rouliers, BP 1036 - 51 687 REIMS

Cedex 2, France.

E-mail address: aude.villemain@univ-reims.fr (A. Villemain).

Safety Science xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Safety Science

journal homepage: www.elsevier.com/locate/ssci

Please cite this article in press as: Villemain, A., Godon, P. Toward a resilient organization: The management of unexpected hazard on the polar traverse.

Safety Sci. (2016), http://dx.doi.org/10.1016/j.ssci.2016.03.008

happens. In this case, resources are allocated to match the require-

ments of the expected event and special functions may be acti-

vated. The Safety Management System (SMS) is interested in this

problematic, considering the fact that organizational and manage-

ment factors have to be taken into account to understand human

contribution to major accidents (Hale and Hovden, 1998). SMS is

the most efﬁcient way of allocative resources for safety regarding

which organization plays a major role. Supplementary resources

are mobilized and local strategies are deployed to cope with distur-

bances. This is « opportunistic bricolage » which is a way to offset

the disturbances and maintain the functioning of the system at

the lowest possible level of risk. It is a safety measure deployed

by human expertise, as well as the use of speciﬁc individual and

collective skills in real time, as previous studies in polar context

have already shown (Villemain and Lémonie, 2014; Villemain

and Godon, 2015).

Studies focused on extreme situation at work are poor in the

ergonomics approach, and even more so in polar context. Few

researches in polar conditions have been led in a logistics thematic

in arctic (Lièvre, 2007). At the present time, only our researches on

working conditions in the Antarctic context working are being car-

ried out with an ergonomics approach (Villemain and Lémonie,

2014; Villemain and Godon, 2015). To conduct research in such

conditions requires a speciﬁc methodology to collect data because

of the harsh conditions. Subsequently, the ethnographical method

is the most appropriate to the ground constraints (Rix-Lièvre and

Lièvre, 2010). These authors used this kind of method in order to

study ski polar expeditions in arctic through ethnological observa-

tions and interviews.

The context and motivations of this research are particular and

require to be speciﬁed. The French polar traverse, initiated in 1993,

was followed by the creation of the scientiﬁc French–Italian station

on the Antarctic continent called Concordia (situated to 1150 km

from the French station Dumont D’Urville (DDU) and opened in

2005). At that time, the objective was to design a freight and mate-

rial transport mode, in order to build the station, from DDU (car-

ried by boat until this station) to Concordia. Thus, the issue then

was a technical, material, technological, logistical and economical

challenge. The traverse has been explored more from a technical

point of view than a safety aspect, only guided by an experiential

or empirical observation. In Antarctica, the environment hostility

does not facilitate human activities, more particularly when those

take place outside such as mechanical task to repair machines.

During the traverse, temperatures can be below 50 °C in Febru-

ary. Thus, in this context, the smallest incident, if not managed

immediately, can bear heavy consequences and become dramatic

due to a limited medical assistance and the isolation. The vital

prognostic is quickly engaged.

The polar traverse could be deﬁned as a set of vehicles in move-

ment in the polar continent, with a total autonomy. Eleven days are

necessary to reach Concordia from DDU. Three return-trips DDU–

Concordia are organized during the austral summer (between

November and February). The convoy is composed with about

ten persons, with a minimum of seven mechanics and one doctor,

three (snow trains), seven machines, three levelling machines, and

loads are consisting of fuel tanks and containers (see Diagram 1).

The logistical traverse has to carry freight to Concordia with both

quantitative and qualitative criteria, as quickly as possible and con-

suming as little fuel as possible. Today, twenty-two years after the

ﬁrst traverse, no human loss has been reported since the traverse

was set up.

During about twenty-three days for a return trip, traversers will

cross the white-ice desert living in a caravan and driving eleven

hours per day. The traverse will be punctuated with mechanical

incidents considered as unexpected events. All raiders know that

incidents will happen during the traverse. Thus, in that way, we

can consider that such situations are not unforeseen. It is however

impossible to determine which kind of incident, when (in bad

weather conditions), neither how and which consequence such sit-

uations will entail (pieces to repair or not? Know-how or experi-

ence to face the event?) nor where exactly in the convoy, which

material, etc. In this regard, such events can be considered as

unforeseen... Everybody knows that this will have incidents, but

nobody knows exactly which ones. The real risk rests more in

the incident conditions per se than concerning the unforeseen

event in its current form.

Thus, in this context, the goal of the research deals with the sys-

tem capabilities to withstand shocks or unexpected events, and to

face harsh conditions every year, in order to answer the question

regarding strategies of organizational resilience used to ensure

the safety in a productive system. In the case of the traverse, ‘‘un-

foreseen” can be considered as a risky situation, jeopardizing the

traverse group during a limited period of time, thus calling on

the resilience abilities of this system. What is the nature of the

unforeseen situations on the traverse? What solutions can be

found? What strategies of organizational resilience can be

deployed? To investigate these questions, we rely on the operative

logic to manage risks during the traverses when unexpected events

occurred. An explanatory study was conducted over three com-

plete traverses (return trips) from which we gathered in situ data

in real and dynamic situation to understand how the unforeseen

events were managed, ensuring the organizational resilience of

the system. Firstly, an analysis of the unforeseen events will be

presented, as well as the solutions offered. Secondly, a single tra-

verse will be used to present a case study to understand the role

of the operators and the organization to act in the management

of unexpected events and strategies deployed.

2. Method

2.1. Tools and procedure

The methodology used was, ﬁrstly, aimed at describing and

characterizing the unforeseen events on the polar traverse and,

secondly, at conducting ergonomic analysis of operators’ activities

when the unforeseen situation occurred in particular. The method-

ology chosen was to analyze immediately operators’ activity in

unforeseen and natural situation during the traverse and their

interventions by means of observations in so far as, in such situa-

tions, the event is not necessarily known and neither are the tech-

nicians who are likely to intervene. As mentioned by De La Garza

(2000), troubleshooting activities, by nature unannounced, make

it impossible to deﬁne accurate observation conditions. Conse-

quently, the ergonomic analysis of operators’ interventions only

concerned the three polar traverses studied. Nevertheless, particu-

lar attention was paid, whenever possible, to the potential transfer

of this methodology to other risky environments.

Using both à quantitative and a qualitative approach, data were

gathered over a 3-year period from 2012 to 2015. Participating eth-

nological observations were carried out from immersion work

(note taking, ﬁlms, photos). On board the outward and return trips

of the convoys, the goal was to experience the traverse from the

inside, (a) to understand exactly what unforeseen event referred

to (the nature of unforeseen events), (b) to access proposed solu-

tions without hindering the work in progress and (c) to understand

the various actions undertaken during the traverse which could

impact the management of unforeseen events and strategies of

organizational resilience employed by operators. The methods

used were the following: (1) ﬁrstly, objective and quantitative data

relating to the unforeseen events encountered were identiﬁed, in

order to proceed to a categorization of events (Amalberti, 1996).

2 A. Villemain, P. Godon / Safety Science xxx (2016) xxx–xxx

Please cite this article in press as: Villemain, A., Godon, P. Toward a resilient organization: The management of unexpected hazard on the polar traverse.

Safety Sci. (2016), http://dx.doi.org/10.1016/j.ssci.2016.03.008

Considered as unforeseen events were those, which required a

complete and unforeseen halt of the convoy or a delayed departure

with regard to the starting point. It was studied as a stop sign of the

normal process (Hollnagel, 2010). To begin with, ground data were

collected when the unforeseen events occurred, such as the date

and the time of the unforeseen events, the length of time the con-

voy was stopped, the GPS location, the nature of the unforeseen

event, and the nature of the solutions put forward. The goal was

to understand the conditions of the occurrence of the unforeseen

events as well as the intervention logic of the traverse members

in the management of the unforeseen events, but also to compared

by experience the difference between ‘‘normal situations” and ‘‘un-

foreseen situations”; (2) secondly, equipped observations (by

notes, audio–visual recordings) of the raiders’ activities when an

unforeseen events occurred to keep an update of the strategies

deployed by the system and operators to deal with the distur-

bances; Monitoring these operators’ activities would allow imme-

diate interventions to be observed in real world setting; and (3)

thirdly, interviews were collected in situ, each day throughout

the traverses, with the leader of the convoy involved. Thus, the

chronological order of events on the day (hour by hour) was

respected during interviews in order to trigger the memory. The

aim was to try to understand each action carried out and the

strategic operational actions to respond to unexpected incidents.

2.2. Analysis

In a ﬁrst time, all data from each traverse were transcribed and

compiled so as to give meaning to the management of the unfore-

seen events in the form of matrices to document each unforeseen

event produced and each action carried out during the traverse.

In a second time, analyses of interventions on the three traverses

were led from a temporal reconstruction of the raiders’ course on

the basis of all the data. This reconstruction combined a temporal

structure (history of the intervention on the unforeseen event) and

a functional structure (real work activities of the operators). Vari-

ous elements contributing to this history were then investigated

(in verbalizations and equipped observations) in order to better

understand the global organizational logic. In a third time, verba-

tim and all usable data were categorized using a thematic analysis

(Corbin and Strauss, 2008) with frequencies of categories for each

unforeseen situation. Moreover, the position of the unforeseen

events was converted and then transcribed onto a traverse route

map (Diagram 1). The traces collected during the traverse were

documented (Tables 1 and 2) highlighting the day, the observa-

tions, the interventions, and the length of the intervention.

2.3. Reliability

In this study, the data was validated in four steps: (1) the notes

taken during the three traverses were transcribed, sorted and orga-

nized according to three ideas: all the actions carried out during

the traverse commanded by the convoy leader, the unforeseen

events per se, and the solutions applied. We therefore retained

all the information consistent with these 3 preoccupations. Two

investigators analyzed material following the procedures recom-

mended by Miles and Huberman (1994). Each investigator read

the transcripts and individually encoded them following this pro-

cedure; the reliability between judges was veriﬁed and the agree-

ment rate was 100%; (2) the notes from the informal interviews

(for the traverse 50, the case study) were transcribed, presented

and discussed with the convoy leader of the traverse when there

was disagreement until it was clariﬁed; (3) the ﬁnal results

obtained, for the three traverses and including the traverse 50

and notably concerning the nature of the unforeseen events as well

as the actions implemented, were presented to the designer of the

traverse who validated them, according to his great experience;

and (4) thematic units concerning operators’ interventions to cope

unforeseen events were categorized. The agreement rate was 95%

and discrepancies were discussed until agreed upon.

3. Results

The results are presented in 2 parts. In the ﬁrst part, objective

and quantiﬁed dimension with ﬁgures and regrouping the data

from all three traverses are presented in order to understand the

organizational resilience with a global perspective. They aim to

show all the interventions carried out on the traverse, consecutive

or not to breakdowns. The objective of this approach is to under-

stand the nature (a) of the unforeseen events; (b) the interventions

proposed; (c) the actions carried out during the traverses linked to

the management of the unforeseen events. In the second part, a

single traverse will be presented as a case study in order to attempt

to extract a strategic operational understanding from it with a

speciﬁc point of view born of action and activity. For this, an expla-

nation of the composition of the convoy during traverse 50 is

described, as well as cartography of the unforeseen events. Each

unforeseen incident, the intervention proposed as well as the

Diagram 1. Composition of convoy at the start of the traverse.

A. Villemain, P. Godon / Safety Science xxx (2016) xxx–xxx

Please cite this article in press as: Villemain, A., Godon, P. Toward a resilient organization: The management of unexpected hazard on the polar traverse.

Safety Sci. (2016), http://dx.doi.org/10.1016/j.ssci.2016.03.008

Table 1

Types of incidents, interventions and length of interventions of the traverse 50.

Days Observations Interventions Length

Traverse outward journey

1 Speed of ST1 too slow Attachment of grading machine 8 to aid with traction 10 min

2 Hose (2) Repair 1 h 15 min

Alternator (2) Repair 20 min

3T°C high exhaust level (10) Modiﬁcation of convoy 20 min

Broken tank hook Abandon – modiﬁcation of convoy 10 min

Hydraulic motor (K1) Repair 2 h 15 min

Command ventilator (K4) Repair 1 h

6 Off-track, tank stuck in snow Fuel transfer – convoy modiﬁcation 3 h

7 Broken tank hook Dumping tank – convoy modiﬁcation 20 min

8 Injectors (2) Irreparable – machine loaded onto 2 h

Sledge – convoy modiﬁcation

9 Control blade (K4) Repair 1 h

Alternator belt (K1) Repair 30 min

10 Starter (10) Repair 6 h

Torn sheet metal container Convoy modiﬁcation 30 min

2 off-track incidents Intervention (X2) 2 h

Total: 20 h 40 min

Traverse return

1 Valve blade (K2) Repair 40 min

2 Headlights (K3) Repair 20 min

Broken caravan hook Repair 1 h

Leak in power generator Repair 10 min

Inﬁltration of exhaust fumes in cabin (6) Repair 20 min

4 Headlights (11) Convoy modiﬁcation 20 min

Hydraulic motor ventilation (K2) Irreparable – machine loaded onto sledge – convoy modiﬁcation 2 h

5 Leak of radiator cooling liquid (9) Repair 20 min

Starting (9) Repair 10 min

6 Wheel unscrewed when Caterpillar leaving (9) Irreparable – machine loaded onto sledge – convoy modiﬁcation 1 h

Lack of power ST1 Convoy modiﬁcation 20 min

7 Door handle broken (10) Repair 30 min

Drawbar broken Repair + convoy modiﬁcation 2 h

8 Flexible + command blade (K1) Repair (2) + convoy modiﬁcation 1 h

Lack of power (10) Convoy modiﬁcation 20 min

Pistons (K1) Repair + convoy modiﬁcation 20 min

Total: 10 h 50 min

Table 2

The convoy re-organization to repair and anticipate the future.

Observations ST1 Interventions

Speed too slow – Hitching of grading machine 8 to tow (day)

Hose (2) – Mechanical intervention

Alternator (2) – Mechanical intervention

Off-track, tank stuck in snow – Organizational intervention

Broken tank hook ? reparable but not carried out 1. Abandon tank

2. Recuperation of a tank from ST2 ? relief

3. Machine mounted on transport ski in ST2

4. Modiﬁcation of convoy (8 tractor on ST1)

5. Recuperation of a tank from ST2

Injectors (2) – Irreparable

Observations ST2 Interventions

T°C high exhaust level (10) 1. ST2 tank in ST1 following broken hook

2. Pumping of ST2 tank

3. Tank in RT1 following machine on ski

Tank hook broken (end of day) ? reparable but not carried out 1. Pumping tank ST2

2. Abandon tank ST2

Command ventilator (K4) – Mechanical intervention

Command blade (K4) – Mechanical intervention

Starter (10) 1. Mechanical intervention (6h)

2. Pumping tank ST2

3. Abandon tank ST2

2 off-track incidents – Organizational intervention

Observations ST3 Interventions

Hydraulic motor (K1) – Hydraulic motor (K1)

Sheet metal container torn => irreparable 1. Modiﬁcation of convoy in 4 trailers with the container hitched to 8 (ST4)

2. Modiﬁcation of position of machines between STl and ST3

3. Modiﬁcation of position of grading machines: all in front of ST1

Alternator belt (K1) – Mechanical intervention

Observations ST4 Interventions

Off-track incident – Organizational intervention

4 A. Villemain, P. Godon / Safety Science xxx (2016) xxx–xxx

Please cite this article in press as: Villemain, A., Godon, P. Toward a resilient organization: The management of unexpected hazard on the polar traverse.

Safety Sci. (2016), http://dx.doi.org/10.1016/j.ssci.2016.03.008

length of the latter, will be analyzed in order to better understand

the strategies of organizational resilience developed.

3.1. Objective and quantiﬁed dimension to understand the resilience

On the three traverses carried out, we identiﬁed 95 unforeseen

events, that is to say 95 regulatory actions following the occur-

rence of a hazard.

3.1.1. Five kinds of unforeseen events

The results indicate the existence of ﬁve types of unforeseen

events provoking a temporary stop of the convoy or its delayed

departure (Fig. 1):

– Stalling of the tractors (15%) arising either from increased trac-

tive effort following the sinking in snow of the loads or because

the loads stayed stuck to the ground at morning startup, or from

a lack of grip of the tractor itself because of the surface being too

icy.

– Breakdowns (58%), which concern mainly the machines, the

generator set.

– Breakages (16%), which concern the hitches for items being

towed.

– Bad weather (8%), which only allows reduced visibility in the

best of cases and a circulation with special headlights to

increase the contrast.

– Off-track incidents (3%), linked to inattention or to falling asleep

at the wheel, following which the loads sink into the loose snow

because the skis left the compacted track.

3.1.2. Mechanical, organizational solutions, or both

The results indicate that these unforeseen events can be regu-

lated in three ways (Fig. 2): by means of a mechanical intervention

(53%), by an intervention on the organization of the convoy (orga-

nizational) (40%), by a mechanical and an organizational or com-

bined intervention (7%). Although mechanical interventions are

in the majority, actions on the organization of the convoy are an

undeniable lever in the management of the unforeseen events. Sev-

eral interventions combine both to ensure the repair and also pre-

vention to relieve trailer hitches. An action on the organization of

the convoy is most of the time to act as a support for the repair,

for the mechanical intervention performed. Modiﬁcations to the

convoy can also be carried out instead of a repair, if the targeted

part is irreparable and does not exist in the stock of spare parts.

When we look at the nature of the interventions according to

the type of unforeseen events, it can be observed that the break-

downs and breakages are managed by three types of intervention

(Fig. 3). Organizational interventions are carried out whatever

the type of unforeseen event.

3.1.3. Other organizational actions carried out on the traverses to

anticipate incidents

In total, we identiﬁed 143 actions performed during the three

traverses for 95 unforeseen events. These results suggest that other

actions were performed independently, or at least indirectly, to the

occurrence of the unforeseen events. The actions identiﬁed corre-

spond to the organizational interventions: if 44% of the actions

were performed following the unforeseen events, 56% of these

actions were not consecutive to the occurrence of the events

(Fig. 4). It is very likely that these peripheral actions play a funda-

mental role in the management of unforeseen events.

3.1.4. The organizational game

The organizational modiﬁcations during the traverse and non-

consecutive to the occurrence of an unforeseen event meet two

objectives with regard to the lever of action chosen: ﬁrst, an objec-

tive of breakdown prevention by acting on the position of the

machines in the convoy (Fig. 5). It is possible to « play » with the

58%

16%

15%

Breakdowns

Breakages

Trenching

Oﬀ-track

Bad weather

Fig. 1. Nature of unforeseen events on the traverses.

53%

40%

Mechanical

intervenons

Organizaonal

intervenons

Combined

intervenons

Fig. 2. Nature of interventions following the appearance of unforeseen events.

Mechanical

intervenons

Organizaonal

intervenons

Combined

intervenons

Fig. 3. Nature of intervention for each type of unforeseen event on all 3 traverses.

44%

56%

Organizaonal

intervenons

following

unforeseen

events

Other

organizaonal

intervenons

Fig. 4. Distribution of organizational interventions on the traverses.

30%

55%

11%

Organizaon of

the machines

Organizaon of

tanks

Organizaon of

work

Organizaon of

road trucks

Fig. 5. Possible variations in the organization.

A. Villemain, P. Godon / Safety Science xxx (2016) xxx–xxx

Please cite this article in press as: Villemain, A., Godon, P. Toward a resilient organization: The management of unexpected hazard on the polar traverse.

Safety Sci. (2016), http://dx.doi.org/10.1016/j.ssci.2016.03.008

location of the machines inside a train depending on the patholo-

gies they present (30%, breakdown prevention according to the

tractor symptoms). Next, an objective to improve production and

output will lead to modiﬁcations in the positions of fuel tanks

(55%, the best compromise between fuel consumption and the

homogenization of the speed of movement of the trains in the con-

voy). To do this, a game of musical chairs will take place. In fact, the

convoy leader calculates and decides which fuel tanks will be left

on the road with the fuel of the return trip. In parallel he must also

choose which tank will be pumped at the end of the day to to ﬁll

the reservoirs of the machines.

The work can also be organized otherwise depending notably

on climatic conditions: ﬁlling the reservoirs tanks can be shifted

to the morning of the next day to shorten the exposure to the

low temperatures of the operators in charge of this task (it is colder

at night than during the day). If the weather conditions are too

deteriorated (white out), the driving time can be adapted: when

it is darker, the landscape shows better with the headlights than

during peak daytime hours (11%). Finally, actions on the disposi-

tion of the loads inside a train or on the position of the tractors

can be implemented (4%). For example, if the conditions reduce

visibility too much and it becomes necessary to move with head-

lights, the equipped tractors could change place in a train or in

the convoy because not all the tractors are equipped. Finally, as

some loads can also be associated with certain tractors, it will be

necessary to reshufﬂe the sledge order of a train and sometimes

the whole convoy.

3.2. Case study: the example of traverse 50

3.2.1. The convoy composition logic to anticipate incidents

The liaison by surface convoy must satisfy both technical and

economic constraints. The freight must be delivered in good condi-

tion, on schedule and with a transport cost as low as possible. The

ﬁnal goal being to reduce the unitary tractive effort. Traverse 50

was a convoy composed of 3 snow trains that we will name ST1,

ST2, and ST3. Snow train 1 carried the caravans dedicated to per-

sonnel, the one which produced electric energy (heating), the one

with the dormitory, the refectory and the kitchen (life), the one

with the food supplies and the one with the spare parts. The ﬁrst

train therefore determines the speed of the convoy. Between

brackets is shown the number of involved machines. Seven Cater-

pillar tractor type machines to tow the loads (machines 2, 10, 11, 6,

9, 5, 3), 3 levelling machines to facilitate the passage of the convoy

(K1, K2, K3). The fuel tanks with the fuel for the return will be left

along the route on the outward journey to lighten the convoy

(Diagram 1).

3.2.2. Repair at any cost ...

The traverse took 19 days to make the 2300 km return trip.

From observations and traces taken during the traverse, 32 inci-

dents were identiﬁed (Fig. 6): 18 on the outward journey, 14 on

the return (Table 1). The results show a total of 31.30 h of interven-

tions out of 209 h of movement of the convoy.

Fig. 6. Mapping of incidents during the traverse 50.

6 A. Villemain, P. Godon / Safety Science xxx (2016) xxx–xxx

Please cite this article in press as: Villemain, A., Godon, P. Toward a resilient organization: The management of unexpected hazard on the polar traverse.

Safety Sci. (2016), http://dx.doi.org/10.1016/j.ssci.2016.03.008

The results which follow were supported by the content of

interviews carried out with the convoy leader in situ according

to the time of the unforeseen events. The results of Table 1 show

that the length of the mechanical intervention is not a criteria in

the decision to repair or not. Only the availability of spare parts

on board the traverse shop is decisive. As a result, we can observe

a repair of 6 h (example of the starter, in bold in the Table 1), with

cutting the bodywork using a grinder etc. The operators are there-

fore exposed to cold temperatures during this time. The impor-

tance seems to be to preserve the machine as long as possible

despite the hostility of the environment and the hard conditions

of the repair.

3.2.3. Modiﬁcations of the convoy during the traverse

The outward journey of the traverses has many more con-

straints than the return journey: there is a strong ascending eleva-

tion, the convoy is loaded to the maximum at departure, no load

can be abandoned since the objective is to supply Concordia. There

is also the question of dumping the tanks on the track to store fuel

for the return trip. However there is the same number of incidents

than during the return.

We have therefore only studied all the variations of the convoy

in detail on the outward journey of the traverse (Table 2). The

results of this case study highlight several things. First, with regard

to mechanical interventions following breakdowns or breakages,

two types of situation are possible: either it is reparable or it is

not. On this outward trip, in 4 out of 11 situations of breakdown

or breakage, repair was not possible. As a consequence, the results

show an organizational intervention on several levels, having

repercussions on other snow trains, each time on at least 2 snow

trains (ST), with 2–3 parallel actions, and this, uniquely in the case

of non-reparation. Thus organizational intervention is carried out

on the whole of the convoy depending on the information of the

moment. In certain cases (such as the breakages of the tank hooks),

the breakage was irreparable. But the arrival of this unforeseen

event at that particular moment (aided by the proximity of arrival

in Concordia for example) was a decisive factor in the dumping of

the tank to waste as little time as possible and to deliver the goods

to the station as early as possible. Thus the option chosen can

sometimes be to postpone the repair until the return trip, even

to the return to base.

4. Discussion

4.1. Synthesis

Resilience focused not only on creating resilient system but also

maintaining and managing system resilience. Framework contri-

butions proposed by Rankin et al. (2014) is intended to support

both perspective and prospective safety management activities.

Prospective analysis is supported by the analysis of adaptation

enablers, objectives and situational conditions and supports mak-

ing prediction on how changes may affect the system. How practi-

tioners and their management recognize, communicate and

perceive work practice is critical for managing and maintaining

system resilience (informal exchange between raiders during the

fuel problem). The system should support and enable people to

be adaptative to increase the system resilience potential thanks

to a systematic analysis of strategies.

The results highlighted strategies of organizational resilience to

cope unforeseen events which resided in reactive and proactive

action alternation during the traverse, acting on (1) the dynamic

convoy organization : the logic convoy composition to distribute

loads (according to engines and their power and symptoms) and

the organizational game with sledges and machines during the

traverse before the unforeseen event; and (2) the competences

and the risk evaluation: a speciﬁc work organization to repair at

any coast in a hostile context, with mechanical and organizational

competences.

4.1.1. A dynamic organization of the convoy to manage safety

The results emphasize that almost half the interventions on the

traverse concern preventive actions or attempts to optimize out-

put, notably with regard to the ratio between the consumption of

fuel and the speed of the convoy. Even if the maintenance opera-

tions carried out each evening were preventive, this proactive per-

spective plays an essential role in the management of unforeseen

events on safety (in accordance with Safety Management System)

and on the appearance of future breakdowns.

4.1.1.1. Autonomy. Moreover, breakdowns or mechanical problems

on the traverse created opportunities to build the proactive safety,

thanks to precedent reactive mode and operators’ autonomy. These

opportunities from mechanical incidents allowed modifying the

convoy organization to ensure the future safety. This logic is the

same when traversers have to move tanks through the caravan’s

fuel station for example: it is a costly task in time, but also the

opportunity to bring a new organization of the convoy in order

to try to anticipate future incidents. The proactive management

is representative of the autonomy allowed by the organization on

the traverse, an observation made in previous studies (Villemain

and Godon, 2015). Thus, proactive management is only possible

on condition that the organization allows the operators enough

freedom to be autonomous and due to reactive work. It is thanks

to this autonomy that the system can be kept in perpetual move-

ment. The possibility of acting so much on the organization of

the snow trains, the position of the machines in the convoy, the

positioning of the tanks, the choice of fuel tanks to be emptied or

even the hours of work all represent the leader of the convoy’s

margin for maneuverability, allowed by the organization. More-

over, the objective is thus not so much to arrive in Concordia in

the best time possible but to keep the machines functioning to

the maximum, in short, to keep the margin of maneuver existing

for as long as possible, opening the ﬁeld of possibilities in order

to have a supplementary option in the management of future

breakdowns. This can be equated with a strategy of management

of uncertainty and of the organizational resilience. This constitutes

the main guarantee of organizational reliability and of safety. It is a

question of permanent negotiation between advancing and pre-

serving the equipment, the redeﬁned goal ﬁnally being to advance

with a convoy, which has the most possible functional resources to

ensure future safety.

4.1.1.2. Perpetual modiﬁcations. In the ﬁeld of resilience, perfor-

mance variability is seen as essential to ensure an organizational

resilience (Hollnagel et al., 2010). Dynamics is maintained by the

perpetual modiﬁcations of the convoy, thus enabling the preserva-

tion of ﬂexibility to half a day, thanks to operators’ autonomy and

mechanical/organizational competences. Proactive management

therefore exists in the multiple efforts and attempts at changes

to allow the system to keep in perpetual movement and not to

become frozen. As if risk emanated more easily from a static posi-

tion, equated with a non-control of the situation, with a risk suf-

fered, with a rupture of rhythm. Furthermore, the results reveal a

complexity in organizational activity: the organizational modiﬁca-

tions are not limited only to a speciﬁc modiﬁcation to the snow

trains on which the unforeseen event occurs, but affect all the

trains and therefore the convoy, with consequences in the form

of ramiﬁcations, affecting several elements of action. Resilience is

not immediate. It is not an organizational reﬂex but there is a tem-

poral density necessary to its building.

A. Villemain, P. Godon / Safety Science xxx (2016) xxx–xxx

Please cite this article in press as: Villemain, A., Godon, P. Toward a resilient organization: The management of unexpected hazard on the polar traverse.

Safety Sci. (2016), http://dx.doi.org/10.1016/j.ssci.2016.03.008

4.1.2. Competences and risk evaluation

4.1.2.1. Speciﬁc work organization. Even when exposed to low tem-

peratures, the objective is to repair the machine to preserve it for

as long as possible despite the hostility of the environment and

the extreme conditions of repair. This strategy seems to be a deci-

sive factor in the construction of safety in this type of environ-

ment: future safety is ensured by being in danger temporarily. As

much as possible, and whatever the length of time taken for the

repair of the breakdown, the priority is to repair the machine; even

if it breaks down repeatedly, the purpose is to keep it functional for

as long as possible. These results demand attention, as it has been

shown previously that the work on the traverse is mainly orga-

nized around reducing the time exposed to cold temperatures, a

risk factor (Villemain and Godon, 2015).

4.1.2.2. Combined competences. And it is without doubt here that

the skills of the operators speciﬁc to the polar environment come

into play: beyond the mechanical know-how of the raiders, their

understanding and knowledge of the polar environment and the

constraints (especially for the convoy leader who is the decision

maker on the traverse) enable modiﬁcations and transformations

to be made to the convoy a priori of unforeseen events. Resilience

lies in the end on the skills of the operators, the strategies and

compromises developed on a daily basis so that the system works

(Woods and Cook, 2002), even if mechanical skill is followed very

closely by the development of organizational skills as well as a

general understanding of the environment and of its constraints,

adaptation skills. In addition, the results of the current study

seemed to indicate the existence of an alternation between reac-

tive and proactive modes in the traverse organization.

4.1.2.3. Risk evaluation and the risk acceptability. The expression

safety-performance is perhaps only possible in such conditions,

with a calculation of the acceptability of the risk taking. Previous

researches on polar environment in Antarctica showed the work

in extreme conditions (hostile) could develop new competences

and know-how for operators, considering thus the polar environ-

ment as an enabling environment (Villemain and Lémonie, 2014),

as the constructive ergonomics indicated it (Falzon, 2014). A recent

research on the polar traverse indicated that the speciﬁc organiza-

tion of the traverse was enabling, thus the traverse was an enabling

organization (Villemain and Godon, 2015). According to these con-

siderations, risky environments are favorable to the competences

development, but in a certain limit: Because of the habitude, peo-

ple minimized the risk evaluation, because of the acquisition of

these competences in the time. Thus, the risk acceptability and

the risk evaluation, depend on the operators’ competences. And

the danger appears when the operator used to be in hostile

environment.

4.1.2.4. Opportunistic bricolage or inventive solutions. Furthermore,

great lengths are taken to repair, with the implementation of dar-

ing intervention strategies. Cutting the bodywork of a tractor with

a power-cutting tool because the mechanics could not access the

starter is a good example of this: apart from war, these processes

would not be used on another continent. The organizational relia-

bility is based above all on having ﬂexibility, notably in the cases of

critical situations, where « do-it-yourself » improvisation is recog-

nized as a source of resilience (Weick, 1993). While organizational

resilience, in this case, certainly consists in the ability of the orga-

nization to resist shocks, it also consists in avoiding them (Roux-

Dufort and Vidaillet, 2003). In the case of the traverse, safety does

not lie in « set » formalisms, but on strategies, on showing initia-

tive, on ﬁnding opportunistic solutions for breakdowns, bricolage,

an improvisation and resourcefulness shown by the operators in

real-life situations. It is thus built by human expertise and by

putting into action in real time both individual and collective skills.

All these things reinforce the dynamic vision of safety.

4.2. Limitations

These results have limitations that require further studying.

Firstly, they are derived from a study of only 3 traverses, due to

the difﬁculty to participate and to collect data on a large number

of traverses. The data collections are costly because they are led

in immersion. No studies have been done in this area. It is thus

an explanatory study which needs to investigate the better way

to understand the safety dimension as a whole. Finally, it would

have been better to complete our analysis by gathering the collec-

tive activity analysis of the work and not only the task; that is to

say focusing on the traversers behaviors (called ‘‘ the doing”) and

on verbal communication exchanged within the team (called ‘‘the

understanding”) (

Cuvelier and Falzon, 2015). This dimension did

not appear in this article. But a reﬂexive and methodological work

is necessary in order to grasp the system resilience through tra-

versers’ actions in a natural world setting, in a speciﬁc context such

as the polar environment.

4.3. Implications

From these results, some key-lessons could be proposed regard-

ing the traverse management. First, organizational actions will

allow to be in a complexity reading of the situation (convoy as a

whole) and not in an analytic reading in order to solve problems

one by one on an isolated basis. Secondly, the know-how devel-

oped during the traverse does not only reside on mechanical com-

petences but also organizational ones to negotiate the risk

acceptability with regards to repairing and to ensuring the future

safety. Thirdly, the organizational strategies development con-

tributes to a permanent movement and prevents a set and static

convoy composition. This movement also prevents being hindered

by uncertainty which is difﬁcult and impossible to manage as a

whole.

The results of this study address high-risk environments, speci-

ﬁc and conﬁned, like in space; the analogy can particularly be

made with long-duration space ﬂights, like to Mars, where there

will be a group of humans, left to their own resources, autonomous

and without possible assistance, who will have to learn to cope

with unforeseen events.

Finally, it is important to emphasize that the notion of an

unforeseen event remains very vague and « is always relative to a

concrete subject and to cognitive processes situated in a context

of action, social relationships, a physical context and limited by

the knowledge and know-how of the operator » (Perrenoud,

1999, p. 123). The question regarding unforeseen for whom

remains to be seen. One has to take into account that the tra-

versers, regulars on the traverse now for a good twenty years,

are no longer surprised by anything. It is noted that working for

long periods in polar conditions has led to habituation (Villemain

and Godon, 2015). The situations met, even though remaining con-

sidered as undesirable events, are not really unforeseen. As a

result, the situations chosen represent mainly what Hollnagel

(2004) called possible situations, i.e., thought about unforeseen

events. It is normal that on the traverse there should be a type of

unforeseen events described throughout this article. Ultimately,

it is a normal abnormality! It therefore becomes interesting to note

what happens to the system as soon as this balance is broken by an

unusual unforeseen event and to understand how the management

of an unthought-of or unforeseen event is built (Hollnagel, 2004)

on the traverse, like in 2013, when the fuel started to freeze on

the return journey of the traverse.

8 A. Villemain, P. Godon / Safety Science xxx (2016) xxx–xxx

Please cite this article in press as: Villemain, A., Godon, P. Toward a resilient organization: The management of unexpected hazard on the polar traverse.

Safety Sci. (2016), http://dx.doi.org/10.1016/j.ssci.2016.03.008

References

Amalberti, R., 1996. La conduite des systèmes à risques. PUF, Coll. Le travail humain,

Paris

Cuvelier, C., Falzon, P., 2015. The collective construction of safety: a trade-off

between ‘‘understanding” and ‘‘doing” in managing dynamic situations. Appl.

Ergonomics 47, 117–126

Corbin, J.A., Strauss, A., 2008. Basics of Qualitative Research. Sage, Thousand Oaks,

Daniellou, F., Simard, M., Boissières, I., 2009. Facteurs humains et organisationnels

de la sécurité industrielle: un état de l’art. FonCSI, Toulouse

Dekker, S., 2006. Resilience engineering: chronicling the emergence of confused

consensus. In: Hollnagel, E., Woods, D., Leveson, N. (Eds.), Resilience

Engineering: Concepts and Precepts. Ashgate, Aldershot, UK, pp. 77–92

Dekker, S., Dahlstrom, N., van Winsen, R., Nyce, J., 2008. Crew resilience and

simulator training in aviation. In: Hollnagel, E., Nemeth, C., Dekker, S. (Eds.),

Resilience Engineering Perspectives: Remaining Sensitive to the Possibility of

Failure, 1. Ashgate Studies in Resilience Engineering, Aldershot, UK, pp. 119–

126

De La Garza, C., 2000. L’activité de diagnostic dans un système dynamique: le cas du

dépannage d’installations de signalisation ferroviaire. Rev. Eur. Psychol. Appl.

50 (1), 39–49

Falzon, P., 2014. Constructive Ergonomics. Taylor & Francis, CRC Press.

Hale, A.R., Heijer, T., 2006. Deﬁning resilience. In: Hollnagel, E., Woods, D.D.,

Leveson, N. (Eds.), Resilience Engineering: Concepts and Precepts. Ashgate,

Aldershot, UK, pp. 35–40

Hale, A.R., Hovden, J., 1998. Management and culture: the third age of safety. A

review of approaches to organizational aspects of safety, health and

environment. In: Feyer, A.M., Williamson, A. (Eds.), Occupational Injury. Risk

Prevention and Intervention. Taylor & Francis, London

Hollnagel, E., 2004. Barriers and Accident Prevention. Ashgate, Aldershot, UK.

Hollnagel, E., 2006. Resilience: the challenge of the unstable. In: Hollnagel, E.,

Woods, D.D., Leveson, N. (Eds.), Resilience Engineering: Concepts and Precepts.

Ashgate, Aldershot, UK, pp. 9–19

Hollnagel, E., 2008. Risk + barriers = safety? Saf. Sci. 46 (2), 221–229.

Hollnagel, E., 2009. The four cornerstones of resilience engineering. In: Nemeth, C.,

Hollnagel, E., Dekker, S. (Eds.), Resilience Engineering Perspectives, Vol 2.

Preparation and Restoration. Ashgate, Farnham, UK, pp. 117–133

Hollnagel, E., 2010. Prologue: the scope of resilience engineering. In: Hollnagel, E.,

Pariès, J., Woods, D., Wreathall, J. (Eds.), Resilience Engineering in Practice: A

Guidebook. Ashgate: Studies in Resilience Engineering

Hollnagel, E., 2012. FRAM. The Functional Resonance Analysis Method for Modelling

Complex Socio-technicals Systems. Ashgate, Farnham, UK

Hollnagel, E., Woods, D., 2006. Epilogue: resilience engineering precept. In:

Hollnagel, E., Woods, D., Leveson, N. (Eds.), Resilience Engineering: Concepts

and Precepts. Ashgate, Aldershot, UK, pp. 347–358

Hollnagel, E., Pariès, J., Woods, D., Wreathall, J., 2010. Resilience Engineering in

Practice: A Guidebook. Ashgate Studies in Resilience Engineering

Leveson, N., Dulac, N., Zipkin, D., Cutcher-Gershenfeld, J., Carroll, J., Barrett, B., 2006.

Engineering resilience into safety critical systems. In: Hollnagel, E., Woods, D.,

Leveson, N. (Eds.), Resilience Engineering: Concepts and Precepts. Ashgate,

Aldershot, UK, pp. 95–123

Lièvre, P., 2007. La logistique. Editions La découverte. Collection Repères, Paris.

Lundberg, J., Johansson, B., 2006. Resilience, stability and requisite interpretation in

accident investigations. In: Hollnagel, E., Rigaud, E. (Eds.), 2nd International

Symposium on Resilience Engineering, 8–10 November, Juan-les-Pins, France,

pp. 191–198.

Lundberg, J., Johansson, B., 2007. Pragmatic resilience. In: Woltjer, R., Lundberg, J.,

Johansson, B., (Eds.), Proceedings of the Resilience Engineering Workshop, 25–

27 June, Vadstena, Sweden. <https://www.ep.liu.se/ecp/023/, > (Last visited

March 17, 2008).

Miles, M., Huberman, A.M., 1994. Qualitative Data Analysis: An Expanded

Sourcebook. Sage, Thousand Oaks, CA

Morel, G., Amalberti, R., Chauvin, C., 2008. Articulating the differences between

safety and resilience: the decision-making process of professional sea-ﬁshing

skippers. Hum. Factors 50 (1), 1–16. http://dx.doi.org/10.1518/

001872008X250683.

Perrenoud, P., 1999. Gestion de l’imprévu, analyse de l’action et construction de

compétences. Education Permanente 140 (3), 123–144

Pidgeon, N., O’Leary, M., 2000. Man-made disasters: why technology and

organizations (sometimes) fail. Saf. Sci. 34 (1), 15–30

Rankin, A., Lundberg, J., Woltjer, R., Rollenhagen, C., Hollnagel, E., 2014. Resilience in

everyday opérations: a framework for analyzing adaptations in high-risk work.

J. Cognitive Eng. Decision Making 8 (1), 78–97. http://dx.doi.org/10.1177/

1555343413498753.

Rix-Lièvre, G., Lièvre, P., 2010. An innovative observatory of polar expedition

projects: an investigation of organizing. Project Manage. J. 41 (3), 91–98

Roux-Dufort, C., Vidaillet, B., 2003. The difﬁculties of improvising in a crisis

situation – a case study. Int. Stud. Manage. Organization 33 (1), 86–115

Villemain, A., Godon, P., 2015. Construction de la ﬁabilité organisationnelle en

environnement extrême à partir de la sécurité réglée et gérée: étude de cas du

raid Concordia. (Organization reliability building in extreme environment from

the regulated and managed safety: the Concordia traverse case study).

Perspectives Interdisciplinaires sur le Travail et la Santé, 1–17

Villemain, A., Lémonie, Y., 2014. Environnement capacitant et engagement des

opérateurs: une mise en débat à partir de l’activité des techniciens de la base

polaire Dumont D’Urville. (Enabling environment and operators engagement: A

debate from the activity of technicians of the polar station Dumont D’Urville).

Activités 11, 26–43

Weick, K., 1993. The collapse of sensemaking in organizations: the Mann Gulch

disaster. Adm. Sci. Q. 38 (4), 628–652

Westrum, R., 2006. A typology of resilience situations. In: Hollnagel, E., Woods, D.,

Leveson, N. (Eds.), Resilience Engineering: Concepts and Precepts. Ashgate,

Aldershot, UK, pp. 55–65

Woods, D., 2006. Essential characteristics of resilience. In: Hollnagel, E., Woods, D.

D., Leveson, N. (Eds.), Resilience Engineering: Concepts and Precepts. Ashgate,

Aldershot, UK, pp. 21–33

Woods, D., 2009. Escaping failures of foresight. Saf. Sci. 47 (4), 498–501.

Woods, D., Cook, R., 2002. Nine steps to move forward from error. Cogn. Technol.

Work 4 (2), 137–144

Woods, D., Hollnagel, E., 2006. Prologue: resilience engineering concepts. In:

Hollnagel, E., Woods, D., Leveson, N. (Eds.), Resilience Engineering: Concepts

and Precepts. Ashgate, Aldershot, UK, pp. 1–6

A. Villemain, P. Godon / Safety Science xxx (2016) xxx–xxx

Please cite this article in press as: Villemain, A., Godon, P. Toward a resilient organization: The management of unexpected hazard on the polar traverse.

Safety Sci. (2016), http://dx.doi.org/10.1016/j.ssci.2016.03.008