Variabilité climatique, dynamique des feux de végétation et perceptions locales dans le bassin versant du N'Zi (Centre de la Côte d'Ivoire)

Abstract : The N'Zi River Watershed is a sensitive environment to wildfire risks due to its climate, terrain, ecosystems and socio-economic issues. Unfortunately, knowledge about wildfire regime is under-documented in this area. The present work aimed to evaluate of the pyro-climatic regime and its determinants in the prospect of predictive modelling in relation to local perception in the centre of Côte d'Ivoire. In order to achieve this goal, this study used meteorological, satellite (pyrological and biophysical) and socio-economic (household survey) data that were analysed using descriptive and forecasting statistical approaches. These approaches include multivariate analysis, Spearman correlation, generalized additive models, Kruskal-Wallis test, hierarchical clustering on principal components, cross-correlation, and Box-Jenkins methodology. This thesis is organised into five major axes. The first one, devoted to the study of the climatic variability over the period 1973-2016, shows that the thermometric and hygrometric anomalies have increased respectively by +0.36 °C and +1.24 % and the rainfall anomalies of the watershed slightly decreased by -6.04 mm. The second, based on the characterization of the wildfire regime, revealed that the number of wildfires and burned areas show a downward trend. During the period 2001-2016, 19,156 wildfires were recorded, burning 71,979.7 km² of vegetation. The third part, dealing with the impact of environmental variables on the fire regime, showed that climatic and environmental factors significantly impact wildfire activity. The most important climatic factors of wildfire activity are visibility, relative humidity, precipitation, Angstrom index, dew point, thermal amplitude, vapour-pressure deficit and Lowveld fire danger index. Also, the spatial distribution of wildfires is conditioned by environmental factors. The fourth axis, devoted to predictive modelling of the wildfire regime using Box-Jenkins method, showed the effectiveness of the SARIMA models to simulate and forecast wildfire regime. The validation phase combining Ljung-Box statistics, residue and autocorrelation analysis coupled with comparative analysis of simulated and observed data showed satisfactory results and allowed their use for forecasting over a 48-month horizon. Finally, the fifth part, sharing the local perceptions of the pyro-climatic regime, notes that the local populations perceive climate change and the evolution of wildfires, which tend to be in a deterioration trend and have developed strategies to adapt. These findings are consistent with the analysis of meteorological and satellite data. The findings of this thesis will help decision-makers and managers to make decisions to reduce the vulnerability of local populations to current and future pyro-climatic hazards.
Document type :
Complete list of metadatas

Cited literature [468 references]  Display  Hide  Download
Contributor : Jean-Luc Kouassi <>
Submitted on : Sunday, July 14, 2019 - 9:47:08 PM
Last modification on : Wednesday, July 17, 2019 - 1:23:27 AM


PhD Thesis - Jean-Luc Kouassi ...
Files produced by the author(s)


  • HAL Id : tel-02183044, version 1



Jean-Luc Kouassi. Variabilité climatique, dynamique des feux de végétation et perceptions locales dans le bassin versant du N'Zi (Centre de la Côte d'Ivoire). Sciences de l'environnement. Institut National Polytechnique Felix Houphoët Boigny - Yamoussoukro, 2019. Français. ⟨tel-02183044⟩



Record views


Files downloads