Reconstituting past flood events: the contribution of citizen science

Abstract. Information gathered on past flood events is essential
for understanding and assessing flood hazards. In this study, we present how
citizen science can help to retrieve this information, particularly in areas
with scarce or no authoritative measurements of past events. The case study
is located in Yeumbeul North (YN), Senegal, where flood impacts represent a
growing concern for the local community. This area lacks authoritative
records on flood extent and water depth as well as information on the chain
of causative factors. We developed a framework using two techniques to
retrieve information on past flood events by involving two groups of
citizens who were present during the floods. The first technique targeted
the part of the citizens' memory that records information on events,
recalled through narratives, whereas the second technique focused on scaling
past flood event intensities using different parts of the witnesses' bodies.
These techniques were used for three events that occurred in 2005, 2009 and
2012. They proved complementary by providing quantitative information on
flood extents and water depths and by revealing factors that may have
contributed to all three flood events.


in-person interview was forecast to last between 45 to 60 minutes but it depended on the history told and no time limit was 95 imposed. At the end, interviews lasted from 30 to 60 minutes. In some cases it was possible to register the narrative. The information obtained from the narrative allowed identifying which neighbourhoods were flooded. Then the chiefs of flooded neighbourhoods were involved in a participatory mapping in the house and in the field, together with hand/GIS mapping for the latter case. The purpose of this second step was to formalize and express the chiefs' memories of the floods (as witnesses or victims) in explicit form in order to obtain past information useable for flood hazard assessment, such as flood extent and 100 water depth. Tools such as land-use paper maps of the area with footprint of houses and different land-use categories (see figure 1), fixing pins, handled GPS and mobile GIS were used.

Stage 1: Investigation on past flood information in neighbourhood chief's house
The methodology of this stage was derived from techniques used in police investigation (Fisher, 2010, Perfect et al 2008. Compared to other forms of interviews, it allows the witness (here the neighbourhood chief) to play a more active role, by 105 expressing freely his history without being interrupted or influenced by questions, which could distort the memory (Loftus and Palmer, 1974). First, neighbourhood chiefs were put into a mental condition allowing them to focus their thoughts, cognitive and emotional state by closing their eyes (Perfect, 2008) and avoiding physical and psychological distraction (e.g. telephone calls) during this phase, as it requires intense concentration (Fisher, 2010). Some neighbourhood chiefs felt uncomfortable when closing their eyes. In such cases, they were told to focus on a blank surface, like a table or the floor. Once ready, they 110 expressed their memories of the event in the form of descriptive stories, as they came to their mind, using their own words and language, Wolof in that case, in order to avoid misunderstanding. They were instructed to describe in detail anything that may be related to the event, such as a) processes that accompanied the flood (e.g. the rupture of a water drainage pipe, manmade obstacles ); b) important political or public events as time indicators (e.g. proximity to a presidential election, football game); c) peculiar flood-related measures taken by the authorities enabling the event to be dated; d) spatial indicators such as place 115 names, street names allowing to reconstruct the flooded areas, and, e) the event itself, with information making it possible to deduce the height of water reached (e.g. "water reached our knees").
Following the narrative, only chiefs who indicated having been confronted with floods went through participatory mapping using scale mapping (62 chiefs out of 82, see supplementary material 2). This phase required a training on how to read and use a map. Therefore, the concerned neighbourhood chiefs were first familiarized with a land-use map of their neighbourhood 120 locating their house and details about their area including main and secondary roads as well as houses. After this introductory explanation, the neighbourhood chiefs used the map to describe their spatial perception of the different flood events, using a distinctive colour pencil to draw the flood contours of each year. Coloured pins were used for indicating the water depth at different locations on the map; red for high level of water, green for medium and yellow for low. This method allows obtaining a qualitative indication of the water depth as well as its spatial distribution. 125 The objective of stage 2 was to consolidate the response by confronting the story leading to the first map to the on-site mapping.

Stage 2: Investigation on past flood information with neighbourhood chiefs in the field
To do this, neighbourhood chiefs brought us to the places they previously described. This is important because memory retrieval is facilitated when the context of the event is recreated, and neighbourhood chiefs can also use their other senses (sight, hearing, smell) to better remember the event. We drew the polygon of the spatial extension using a mobile GIS, with 130 GPS receiver automatically recording the site location. Furthermore, we measured 64 sites for the water level indicated by 49 neighbourhood chiefs with a graduated ruler (supplementary material 3, 4 and 5) and took the GPS coordinates. Post-processing treatments include merging the contours of flooded areas obtained on paper map with the ones obtained in the field as well as checking the correspondence between qualitative water levels obtained with the coloured pins to the water level measurements.

Local representatives: participatory mapping on flood extent and water level of past flood events 135
The second group involved in investigating past flood events was composed of 182 local representatives selected by associations well implemented locally (e.g. "Réseau d'Information d'Education de Communication", "Association des Relais Communautaires de Yeumbeul") dealing with development of the neighbourhood and awareness on health issues. The aim for involving local representatives is to confront their information with the one provided by the neighbourhood chiefs in order to check the consistency between the two sources. 2 or 3 local representatives were selected per flooded neighbourhood, 140 accounting for 130 out of 182 representatives, in order for them to confront their memories and reach a common agreement (Swanson et al., 2016) before providing information on flood extent and water depths for the different flood events. Data on flood extent were retrieved by participatory mapping using hands-on techniques. For this, representatives were trained the same way as the chiefs. These maps were then digitized. Regarding water level, information was retrieved on the same 64 sites indicated by the neighbourhood chiefs using the different parts of the human body, e.g., ankle, knee or shoulder. This strategy 145 was proposed to provide local representatives with a visual resource to describe the water level more easily. Then, the predefined tags were converted into quantitative data by using average body segment lengths expressed as a fraction of body height defined in the field of physical anthropometry (Winter, 2009;Drillis and Contini, 1966). The bottom-up dimensionless coefficients applied for each anthropometric segment (supplementary material 6) are (Winter, 2009;Contini, 1972): ankle (0.039), knee (0. 285), wrist (0.485), elbow (0.63), chest (0.72), shoulder (0.818), chin (0.870). Finally, the water depth was 150 obtained by multiplying the value of the appropriate coefficient by the contributor's (local representative) height, measured on site using tape measure (supplementary material 3, 4, 5 and 6).
As we used two different approaches to obtain the same information, we needed to assess the level of agreement instead of the correlation between the two datasets. We used the Bland-Altman method (1986) (Bland and Altman, 1986), which allows determining the level of agreement between data acquired with two different techniques, even if there is no information about 155 the "true" values (Bland and Altman, 1986). In our case, we assessed depth values that could not be measured instrumentally during the flood events under study. The Bland-Altman method calculates the differences between the results obtained with two different approaches and plot them against the average of the two approaches.

Remote sensing analysis
In order to assess the reliability of flooded areas provided by the two citizens groups, we used remote sensing analysis. Our 160 constrains were: 1) the availability of images for the years considered, 2) the availability of free access, 3) a resolution sufficient for the size of our study area (9km 2 ), and 4) a minimum cloud cover. Radar images such as TerraSAR-X, Radarsat-2 or COSMO-SkyMed, can provide information with high resolution (Schubert et al., 2012)  The multispectral Spot images of 10 m resolution were merged with a Spot panchromatic image with a spatial resolution of 170 2.5 m to increase their spatial accuracy. In order to detect water-covered areas, we applied the normalized difference water

Flood extent mapping
Flood extents for the 2005, 2009, and 2012 events were obtained from the two citizen groups using the methodologies described in Figure 2 and then compared to the results derived from remote sensing analysis (Fig. 3). 195 As a first result, the citizen science revealed that the 2005 event was the most widespread whereas the 2012 event was the smallest (table 3). Flooded areas provided by local representatives are slightly smaller than those indicated by neighbourhood chiefs (table 3), showing variations from 1.8% in to 2005 to 0.6% in 2012 (table 3). In terms of mapping, slight differences appear on the extents identified between the two citizen groups (Fig. 3), but areas overlap reasonably well (Fig. 4).
Remote sensing analysis confirms the main flooded areas in the centre part of the study area (Fig. 3), but some discrepancies 200 occur on the edges. The total surface area is smaller than the one provided by citizen science for all years (table 3).

Water depth information
Water depth is one of the key parameters considered in describing flood intensity and mapping hazard (Van Alphen et al., 2007), but difficult to record during flood events. Therefore, retrieving flood depths from past events is of prime interest.

Discussion and conclusion
In this study, we used citizen science to retrieve information on three past flood events having impacted the region of Dakar 215 during the past 10 years. Our approach provides quantitative information on water depth, helps retrieving the flood extents and provides insights on aggravating factors of the intensity of floods.
Our methodology consisted of a set of techniques to gather the most complete spectrum of information. Those techniques are unusual in the field of flood hazard assessment and we had to resolve some challenges related to the time elapsed and the understanding of maps. One of the techniques is based on people's episodic memory. We used face-to-face interviews with 220 neighbourhood chiefs, applying specific tools in order to limit external influence and distortion of memory. Moreover, the procedure was completed with a scene visit with each neighbourhood chief involved in order to consolidate the information provided verbally. The scene visit is very important because the time elapsed between the oldest event and the date of the study is about 12 years. As time goes by, memories can become vague (Lacy and Stark, 2013). However, people having faced traumatic and stressful events, like floods in our case, tend to keep a more accurate, detailed, and time persistent memory of 225 the event (Sotgiu and Galati, 2007).
Another technique involved participatory hands-on mapping. Mapping can represent a challenging task for laypersons (Handmer, 1985;Żyszkowska, 2015, 2017 as they may have difficulties to understand and to locate themselves on a map. Moreover, maps are usually composed of a graphic semiology applying standard rules and recommendations that rarely take into account the culture or knowledge of the citizen (Fuchs et al., 2009). Therefore, if a citizen has no experience in reading or 230 producing maps, information can be incorrectly reported. To overcome this problem, we trained people to ensure they understood the map; we explained what they should be doing and how to do it.
Quality and reliability of citizen science data is a growing research field (Flanagin and Metzger, 2008;Crall et al., 2011;Silvertown et al., 2015). In our study, we developed different strategies in order to improve these two aspects. We decided to work with two different target groups according to the context and the purpose of the study. The objective was to check the 235 consistency of information between the two groups. If the same flooded area is mentioned by the two groups, there is a good chance that the area was effectively flooded. Due to the social organisation of the Dakar region, one way to limit issues about source credibility (Flanagin and Metzger, 2008) was to involve neighbourhood chiefs. Indeed, these chiefs are appointed by local citizens, based on the trust placed on them and on their long-lasting presence in the area. Usually, they have a good memory and good verbal abilities. Moreover, as a witness or sometimes as a victim, they were at the forefront of the flood 240 scene, therefore representing a valuable source of information to describe the chain of events. The second group was composed of local representatives. The selection was made with the support of local and well-implemented associations.
Identifying the chain of processes generating flooding is very important for flood hazard assessment (DAEC, 2016) as it enables considering more realistic flood scenarios. Citizens living in flood affected areas are not often included in post-event or flood hazard assessments, although they could provide accurate insights as they have a good understanding of their surroundings 245 (Tran et al., 2009), and an in-depth knowledge on the specific realities. Our study demonstrates this as neighbourhood chiefs allowed identifying additional natural and man-made factors that contributed to flooding, such as the rise of ground water, the Warouwaye lake overflow, and the emptying of septic tanks.
In terms of flooded areas, the results obtained per event with the two groups of citizens are similar. Some spatial differences can be observed regarding the extent. Reasons for the differences could be related to a) a more in-deep knowledge of the 250 neighbourhood and their surroundings by the chiefs, as they have the confidence (Tall, 1998) of the inhabitants, therefore having access to more detailed information; b) the techniques used in mapping the areas: with neighbourhood chiefs, we used a two-stage procedure to retrieve the flood extent, involving hands-on mapping and GIS mapping in the field, whilst the local representatives only produced hands-on maps that were then digitized.
When compared to remote sensing analysis, spatial differences can also be noticed which can be explained by a) the resolution 255 of the selected images varies from 0.5 to 2.5 m, probably not accurate enough to capture all flooded areas (Grimaldi et al., 2016); b) the acquisition time as the post-event images were taken up to one year after the flood; c) the technical limitation as satellites cannot capture floods inside a structure that are solid like buildings or under vegetation canopies (Wilson et al., 2007; Mallinis et al 2013);d) the efficiency of the algorithm used for the remote sensing analysis (Bates et al., 1997;Schumann et al., 2009). 260 Concerning water depth retrieval, one of the techniques was inspired from studies expressing flood hazard levels on maps using body scale (e.g. EXCIMAP, 2007;Luke et al., 2018). Therefore, quantitative data on water depth were retrieved using a proportion of the size of the human body borrowed from physiology field (Winter, 2009). These values represent an average (Drillis and Contini, 1966), since the length of human body segments depend on body structure (Contini, 1972), gender and racial origin, and therefore could be a source of uncertainties. However, when comparing the two approaches used for water 265 depth investigation, we find a fairly good agreement, with an average differences below 0.3m, which is in the range of other comparisons done for example between observed and simulated methods (Kutija et al., 2014).
Both implication and motivation from citizens are necessary for the success of citizen science projects (Rotman et al., 2012).
To ensure people implication in our project, we first created a Facebook page (http://www.facebook.com/xeexmeude.com) in order to interact with the local citizens. Second, we designed and presented the project in a way to convince contributors that 270 their contribution will be beneficial for them and their neighbours. Third, we worked with community leaders (Bénit-Gbaffou and Katsaura, 2014) and local associations to ensure a better acceptance of the project.
Citizen science requires involvement and time, compared to remote sensing analysis, which could now also take advantage of the free availability of radar images such as sentinel (Malenovský et al., 2012). However, at the scale we worked, these images don't offer the required spatial resolution (Twele et al., 2016), and they don't provide information on depth of flood, which is 275 one of the critical information for flood hazard assessment that we were able to obtain with citizen science.
In conclusion, our study shows the potential of citizen science in retrieving quantitative and reliable information on past flood events, especially in areas where no or few records of past events are available. Our investigation strategy, by involving two different groups of citizens, increases the reliability on the data obtained. Provided the functioning of the society subject to floods is well understood, such approach can be replicated in other parts of the world. Moreover, considering the fact that 280 citizens have been involved in the various steps of this project, they have developed skills in flood data acquisition and understanding of flood processes. They can thus better integrate a decision-making process around flood risk.