^{1}

^{*}

^{2}

^{3}

Cotonou, like West African cities, experiences recurrent flooding. This work aimed to study the effect of climate change on the risk of flooding in Cotonou’s zoos on the periphery Nokoué Lake. We analysed climate data (rainfall and temperature) in Cotonou over the period 1953 to 2013, the flow of the Ouémé River at Bonou, and the variation in the height (elevation) of Lake Nokoué. The results of the analysis of temperatures over the period 1953-2013 show an upward break from 1987 onwards (+0.7°C between 1987 and 2013), while for rainfall, no break was observed over the same period. None of the 13 rainfall indices calculated is statistically significant, but they show a relatively stable or declining trend except for the average daily rainfall, the maximum number of consecutive dry days and the 95th percentile, which all show a slightly upward trend. We have shown that the rainfall recorded in Cotonou does not reveal any abnormal or exceptional character over the last ten years, on the one hand, and the flooding of Ouémé River influences that of Nokoué Lake, on the other hand. This study shows that the floods observed in this area are more related to the effect of climate change.

Floods account for more than a quarter of major natural disasters worldwide. These floods affect all regions of the globe (Ozer, 2008). The countries of West Africa have been affected by many natural disasters for a long time. In recent decades, there has been a significant increase in their frequency and intensity attributable to global warming (IPCC, 2012). Floods, droughts, disruption of rainy seasons and heat waves are the most tangible extreme weather events that affect West African populations (DARA, 2013).

The United Nations Environment Program has identified nineteen climatic “hot spots” in West Africa (UNEP, 2011). Benin is one of these hot spots and has been under serious threat in recent decades by flooding especially the periphery of coastal areas. Cotonou is one of the cities directly exposed to the risk of flooding. Recurrent flooding is linked not only to its geographic location, but also has grown in magnitude over time. Despite the damage caused by these floods, there is an increase in habitats in risk areas, especially around Lake Nokoué. These disasters will not be without consequences for the residents. In Benin, in 2009, displacement due to floods affected at least 20,000 people and, in September 2010, 150,000 people lost their homes (Ferris & Stark, 2012). So it is urgent to assess the risks of this phenomenon by the effects of climate change.

Cotonou area is located on the coastal strip that stretches between Nokoué Lake and the Atlantic Ocean, made up of alluvial sands with a maximum height of about five meters. It is the only commune of the Littoral department and is bordered to the north by the commune of Sô-Ava and Lake Nokoué, to the south by the Atlantic Ocean, to the east by the commune of Sèmè-Kpodji and to the west by that of Abomey-Calavi. It covers an area of 79 km^{2}, 70% of which is located west of the channel (Akomagni, 2006).

Nokoué lake is located south-east of the Beninese lagoon network (6˚25'N, 2˚36'N), ^{2} and its average length is 20 km in its East-West direction with a width of 11 km in its North-South direction (Laleye, 1995).

To conduct this study, we have used R, XLSTAT, EXCEL, STATISTICA, Sphinx and ArcGis 9.3.

The climate data are those collected at Cotonou Airport station. The hydrometric data (the flow rates of Ouémé at Bonou and temporal variation of Nokoué Lake), the annual rainfall maxima in 24 h, 48 h and 72 h as well as the days when Nokoué Lake and the Ouémé at Bonou overflowed are data that we were able to obtain from the General Water Direction of Benin.

The analysis of Cotonou climate data (rainfall and temperature) was performed with R, XLSTAT, EXCEL and STATISTICA software. The analysis of the climate data allowed to assess the links between the evolution of flood risks and the climate change observed over the last decades. For this purpose, from the annual and daily rainfall data over the period 1953 to 2013, we have calculated using R and EXCEL software, the annual total rainfall (PTOT), the number of rainy days (JP), the total number of wet days with rainfall of more than 10 mm, 20 mm and 25 mm (Rx10, Rx20 and Rx25), the average rainfall per wet day (SDII), the maximum rainfall obtained in 1 and 5 consecutive days (Rx1 and Rx5). We also studied the maximum number of consecutive dry days (CDD), the maximum number of consecutive rainy days (CWD) and also the number of days with rainfall above the 95th, 99th and 99.5th percentiles (R95p, R99p and R99.5p) which correspond to intense, extreme and very extreme rainfall events respectively.

We have also plotted the curves of temperature variation and reduced centred annual rainfall and their smoothed average over 7 years. The smoothed 7-year average centred on the fourth value allowed us to appreciate the large multi-year fluctuations.

In order to have a general view of our study, we carried out a homogeneity test which is nothing other than the Pettitt (1979) test to see if there is a break.

We wanted to understand whether the floods in Cotonou are caused by extreme events. For do this reason, we studied extreme rainfall events over the period 1953 to 2013, 1953 to 1982, 1963 to 1992, 1973 to 2002 and 1983 to 2012. To characterize the return periods, we used Gumbel’s law (1958). Note that according to Ozer and Sene (2002), the rainfall values corresponding to a 6-year return period are considered abnormal, those of 10 years very abnormal, those of 30 years exceptional and very exceptional for rainfall with a return period of 100 years.

We also analysed the evolution of the frequency of annual rainfall maxima in 24 h, 48 h and 72 h in Cotonou.

The study of the variations in the heights (dimensions) of Nokoué Lake and the Ouémé flow rates at Bonou was also carried out with R and XLSTAT software.

At the first time, we analysed the homogeneity of the flows by applying the test of PETTIT (1979) to see the stability of the flows from the Ouémé to Bonou during the period 1965-2010. We then established correlations between the Ouémé flows at Bonou, the shoreline of Nokoué Lake and the rainfall in Cotonou to understand the influence of the Ouémé River on Nokoué Lake.

1) Centred annual temperatures reduced and smoothed average over 7 years

Anomalies corresponding to temperature deficits between 1953 and 1957 on one hand and between 1960 and 1962 on other hand can be observed. Subsequently, deficits were also recorded in 1964, 1965, 1974, 1976 and 1986. On other

hand, temperatures were in surplus in 1973, 1987, 1995 and 1998. The same pattern of surpluses is observed from 2000 to 2002, in 2006, from 2008 to 2011 and finally in 2013.

2) Test of homogeneity of the average annual temperature

Test Interpretation:

p-value (bilateral) = 0.0001.

HO: The data is homogeneous.

Ha: There is a date from which there is a change in the data.

Since the calculated p-value is below the significance level threshold alpha = 0.05 the null hypothesis HO must be rejected and the alternative hypothesis Ha must be retained. The risk of rejecting HO when it is true is 0.01%.

As we have a p-value less than 5%, we can say that the series presents a significant break (Ago et al., 2005).

3) Reduced centred rainfall

Analysis of

4) Test of rainfall homogeneity

Test Interpretation:

p-value (bilateral) = 0.415.

HO: The data is homogeneous.

Ha: There is a date from which there is a change in the data.

5) Studies of some rainfall indices

On the different rainfall indices calculated, we note that no index is statistically significant, except for the maximum number of rainy days (CWD) (

6) Study of extreme rainfall events in Cotonou and its relation to flooding

We note that the rainfall corresponding to the return periods of 6, 10, 30 and 100 years are 143.2 mm, 159.3 mm, 192.5 mm and 228.2 mm respectively (

The rainfalls corresponding to the return periods of 6, 10, 30 and 100 years are 131.9 mm, 144.7 mm, 171.2 mm and 199.5 mm respectively (

7) Trends in the frequency of annual maximum rainfall in 24, 48 and 72 hours in Cotonou

The analysis of

8) Analysis of the relationship between the flooding of the Ouémé River in Bonou and the floods in Cotonou

p-value (bilateral) = 0.053.

HO: The data is homogeneous.

Ha: There is a date from which there is a change in the data.

9) Analysis of the variation in the maximum coastline at Bonou and the number of overflow days of Ouémé at Bonou

After analysing

10) Correlation between the peak high water level at Bonou and the number of days of overtopping

Analysis of ^{2} = 0.70).

11) Correlation between the overflow of Nokoué Lake and that of Ouémé at Bonou

The analysis of

12) Analysis of the evolution of rainfall and water level in Nokoué Lake

Analysis of

13) Analysis of the evolution of the flow at Bonou and the water level of Nokoué Lake

A break in the series of temperature data recorded in Cotonou over the period 1953 to 2013 was observed in 1987, splitting two homogeneous periods: one between 1953 and 1986 (27.1˚C) and the other relatively warmer by +0.7˚C between 1987 and 2013 (27.8˚C). This recent increase in temperature is global (IPCC, 2013) and is observed in the same order of magnitude elsewhere in West Africa. Cotonou is therefore definitely affected by global warming.

Moreover, the study of rainfall in Cotonou shows us that there is no break, and therefore no obvious change over the period 1953 to 2013. We carried out an analysis of the rainfall data by calculating 13 rainfall indices. None of these indices shows a statistically significant trend. Except for the index of the maximum number of consecutive rainy days (CWD) which is close (p-value = 0.084)

with a decreasing trend. The numbers of days with rainfall ≥ 10 mm, ≥20 mm and ≥25 mm (Rx10, Rx20 and Rx25), the daily maximum rainfall and the 5 consecutive days (Rx1 and Rx5) show a relatively decreasing trend. The 99th and 99.5th percentiles for extreme and very extreme rainfall events are stable. When it is known that the frequency of flood risk is increasing and that all these rainfall indices have shown a relatively stable or decreasing trend, it can be deduced that climate change does not have too much influence on flooding in Cotonou.

The graphs analysis characterising the return periods of abnormal (6 years), very abnormal (10 years), exceptional (30 years) and very exceptional (100 years) rainfall over the periods 1953-2013, 1953-1982, 1963-1992, 1973-2002 and 1983-2012 enabled us to understand that the different thresholds of torrential rainfall decrease over time and therefore, a priori, that extreme rainfall decreases. Taking the last 30 years over the period 1953 to 2013, only two abnormal rains (1997 and 2003) have been recorded and none in the last ten years. However, we found that floods are becoming more and more frequent in the study area. It is true that heavy rains are still observed today, but both their frequency and their return period are less important now than in the past. Nevertheless, while it is plausible from this analysis that the floods are not caused by the extreme daily rains recorded in Cotonou, it can be deduced that the floods in the study zone are caused by the accumulation of low intensity rains. However, it should be noted that Cotonou has experienced non-abnormal rainfall, which has nevertheless caused flooding in the study zone. The city experienced flooding in June 2007 and the maximum rainfall recorded during the month was 95.3 mm on 28 June 2007. Even if we take the cumulative rainfall for five consecutive days, i.e. from 24 to 28 June 2007, we obtain a rainfall of 129.1 mm, which corresponds to a return period of 4 years, and is therefore quite normal. Then let us also take the example of the recent flooding in Cotonou on 2 June 2015. The entire city and its surroundings were flooded, making the roads impassable. The maximum rainfall recorded that day was 118.4 mm (http://www.ogimet.com/), which is quite normal if we refer to the return period calculated over the period 1983-2012, the threshold of which was 131.9 mm. This rain corresponds to a return period of 4 years. The Beninese press in its entirety has relayed this flooding. As an example, the daily newspaper “24 h au Bénin” of June 3, 2015, states the following: “Would nature be against the current municipal council of Cotonou? There are many reasons to answer in the affirmative. Indeed, the floods that are often decried in Cotonou have resumed ... Almost all the streets of Cotonou were under water. The neighbourhoods of Akpakpa, Zogbo, Agla, Gbégamey and Mènontin suffered the fury of the waters. It took almost five (05) hours of rain to make Cotonou, the beloved but also hated city, unliveable for its inhabitants... Many point an accusing finger at those in charge of Cotonou’s town hall. For them, the city’s sanitation policy is a bitter failure ... In any case, water has driven many citizens from their dormitories”. However, the rain recorded in Cotonou on 02 June 2015 in 6 hours (00 h to 6 h) and 12 hours (00 h to 12 h) was respectively 35 mm and 105 mm (http://www.ogimet.com/). From then on, it is obvious that rain that is not abnormal creates floods in the study area. These low rainfall events associated with the nature of Cotonou’s soil, which has a sub-surface water table, can lead to flooding. In fact, the intensity of extreme rainfall is nowadays decreasing, but the frequency of flooding is increasing.

The various analyses and correlations that have been carried out with the flows of Ouémé River at Bonou, the shoreline of Nokoué Lake and the rainfall in Cotonou have enabled us to understand that the Ouémé River has a strong influence on the rise in the level of Nokoué Lake. And since we know that the Bonou station receives all the water from the large Ouémé watershed, we can expect a strong flow of river water at the Bonou station that directly influences the lake’s rise in level. This is confirmed by the variation curve of rainfall and the height of Nokoué Lake. Since Nokoué Lake reaches its maximum peak between September and October, when rainfall in Cotonou is lower, this corresponds to the great (and only) rainy season in the central and northern parts of the country, from which Ouémé receives most of the water. Thus, the lake's flooding is strongly influenced by the waters of Ouémé at Bonou. In fact, if Nokoué Lake, which is the last receptacle of water from Ouémé, overflows, it will occupy the entire flood plain which is already occupied by populations. It should also be pointed out that the lake is increasingly filled by the acadja system (branches deposited in the lake to serve as spawning grounds for fish), which could accentuate the overflow of Nokoué Lake. Cledjo and Ogouwale (2006) have shown that “acadjas decompose and are deposited in the bottom of Nokoué Lake” They also pointed out that “the accumulation of water hyacinth piles, household waste and the decomposition of the acadjas contribute enormously to the silting of the lake”. Thus, it is clear that the depth of Nokoué Lake has decreased sharply in recent decades. The filling of the lake will have a strong influence on flooding since the lake will no longer be able to contain the water from Ouémé River that feeds it.

In view of the work above, the study of climate indices shows that none of the indices is statistically significant. In fact, it can be seen that the indices in general have shown a trend that is either stable or declining. Nevertheless, there is a slight upward trend in indices such as the number of dry days, the average rainfall per rainy day and the 95th percentile characterising intense rainfall. The study of extreme rainfall led to the conclusion that floods are not related to local extreme rainfall events. The analysis of Ouémé river flows in Bonou and the variation of the coast of Nokoué Lake coupled with the rainfall in Cotonou allowed us to understand that the flooding of Ouémé has a strong influence on the flooding of the lake and directly on the study area.

The authors declare no conflicts of interest regarding the publication of this paper.

Chokpon, A. E., Ozer, P., & Lawin, E. (2021). Assessment of Flood Risk in Cotonou Areas Surrounding Nokoué Lake Due to the Effect of Climate Change. Journal of Geoscience and Environment Protection, 9, 262-279. https://doi.org/10.4236/gep.2021.93016