Extreme Temperature Trends in the Beninese Niger River Basin (Benin)

In the context of climate change, the study of the variability of the climatic extremes in several regions of the world is of capital importance. This study has as main objective to analyze the variability of extreme temperature events in the Beninese basin of the Niger River for the recent and the near future. To achieve this objective, seven (07) extreme temperature indices based on historical daily temperature observations (1976 to 2019) and REMO RCM simulation outputs of RCP4.5 and RCP8.5 scenarios (2021-2050) were calculated. The obtained results were represented by calculating the means for each index and analyzing the trends and their significance by the Mann-Kendall method. The results show that the indices of extreme temperature intensity (TNn, TXx, and DTR), and those related to the frequency of warm sequences (WSDI, TN90p and TX90p) have experienced a significant increase in the past. This increase will continue until 2050. In contrast, the cold sequence frequency index (CSDI) decrease over the historical period as well as over the future period. These indices show much more change with the RCP8.5 scenario than with the RCP4.5 of the REMO climate model. Only the TXx and CSDI indices show statistically significant changes at all stations.

Indeed, deforestation and species extinction are relatively more important and living conditions have become more precarious especially in tropical Africa (Bush & Flenley, 2007). This area, whose economy is essentially based on agriculture, is not without sensitivity (IPCC, 2007).
West Africa is a region where people are always faced with high climate variability. In addition to this variability, climate change resulting from anthropogenic greenhouse gas emissions adds to these challenging conditions (Badara & Camara, 2017). The impact of these gases on climate change is now being assessed through a series of climate simulations over past and future periods (Obada, 2017). For example, we have the future climate simulation scenarios A1B, A2 and B1 which are the most used in impact studies (Haidu, 2009). Thus, the study of climate appears to be of capital interest for the understanding of events but also because of the very negative consequences that a change in variables could have on societies (for example sea level rise).
In fact, explaining and modeling these climate variations at different scales in time and space have become a challenge for climate scientists (AMMA ISSC, 2005;Janicot et al., 2008). This has led the IPCC to introduce emission scenarios to study climate since the 1990s (Leggett et al., 1992). The main objective of this research is to analyze past and future extreme temperatures in the Beninese Niger River basin.

Study Area and Data
Located in the extreme north of Benin ( Figure 1) with an area of approximately 48,000 km 2 , the Beninese Niger River basin occupies 42% of the total area of Benin (114,763 km 2 ). It is located between latitudes 10˚ and 12˚30' North and longitudes 1˚32' and 3˚50' East, and includes the Mékrou, Alibori and Sota sub-basins. This study required the collection of daily maximum and minimum temperature data recorded from 1976 to 2019 at three (03) synoptic stations installed around the Benin River basin (Figure 1). These data are collected from National Agency of Meteorology (Météo-Bénin). These observed temperatures are complemented by the temperatures of the simulation outputs of the REMO Regional Climate Model. The latter is one of the regional climate models that best responds over the study basin (Badou, 2016

Methods Used
Extreme climate indices are defined from daily data. It is thus important to proceed to a good constitution of the daily series.
The quality control test called "Data QC" (Zhang & Yang, 2004) was employed. This performance test follows two procedures: 1) replace all missing values (by code −99.9) according to the international format; 2) replace all unreasonable values by NA. These values include cases where daily maximum temperatures are lower than daily minimum temperatures (Zhang & Yang, 2004).
A bias correction is generally performed on the outputs of climate models for the majority of climate change impact studies. This correction is generally American Journal of Climate Change uni-varied and corrects each variable of interest independently of the others.
There are a large number of bias correction methods. The bias correction method used in this research is called "Delta change" (DC).
The DC method is the simplest and most widely used of the bias correction methods (Graham et al., 2007;Moore et al., 2008;Sperna Weiland et al., 2010) and consists of scaling the observations to obtain the corrected simulations. It is a modest method in which the parameters are usually corrected with an addictive factor for temperature (Equation (1)) (Lafon et al., 2013). Change 100 where p x is the average of the index over the projected period and r x its average over the reference period.
The Student t-test was used to determine whether there is a significant difference between p x and r x .

Observed Trends of Extreme Temperature Indices
The            reference period. There are deviations of 4, 11, 4 and 2 days respectively for CSDI, WSDI, TX90p and TN90p under RCP4.5 compared to the baseline period ( Figure 6) while under RCP8.5 these differences are respectively 4, 5, 3 and 2 days compared to the reference period (Figure 6). At Natitingou, TX90p and WSDI will increase under both RCP4.5 and RCP8.5 scenarios. There are respectively an increase of 2 and 4 days under RCP4.5 and 3 and 4 days under RCP8.5 compared to the baseline period (Figure 6). At this station, TN90p and CSDI could decrease under both RCP4.5 and RCP8.5 compared to the baseline period.

Changes in Extreme Temperature Indices
The decrease could be of 5 and 1 days for RCP4.5 and 6 and 0.4 days for RCP8.5  and 5 days respectively under RCP4.5 and RCP8.5 compared to the baseline period ( Figure 6). Also for this index group, a warming of the basin is indicated for future years as warm sequences increase more than cold sequences. From

Discussion
In summary, there are increasing variations on the extreme temperature intensity index for all the stations considered over the historical period. These results are confirmed by Badou (2016) who showed that temperatures in the basin have increased. Gnanglè et al. (2011) also found the same results for the three climatic zones of Benin. New et al. (2006)  basin. This will most certainly lead to the basin warming more than it has in the past. It is therefore important that scientists, local, national and international decision makers at various levels take optimal and collegial resolutions to counteract the adverse effects of this phenomenon. It is also important to note that not all these indices show significant changes. Only the TXx and CSDI indices change significantly. This reflects that these indices vary more than the other one.

Conclusion
The results obtained on intensity and frequency extreme temperature index for the Beninese basin of the Niger River, confirm the same climatic changes observed in Benin and in West Africa in general in previous works. Almost all the American Journal of Climate Change calculated indices present worrying situations for the future of the basin. Indeed, the indices of extreme temperature intensity (TNn, TXx, and DTR), those related to the frequency of hot sequences (WSDI and TX90p) tend to increase. In contrast, the cold sequence frequency index (CSDI) is decreasing. Most disturbing is that both regional climate model scenarios used in this study indicate that the deteriorated climate conditions of the recent past will continue for the next 30 years. Statistically significant changes are noted for the TXx and CSDI. It is therefore up to each party to play an effective role in mitigating the consequences.
It should also be noted that the change is more pronounced with the RCP8.5 scenario than with RCP4.5 in the model used.