Statistical Modeling of Rainfall and Drought in North West Bangladesh

Abstract

Drought is a region specific phenomenon, its characteristics vary from one climate regime to another. Though Bangladesh is a riverine country, the northwestern part of the country is vulnerable to drought. Thus, drought characterization should receive much attention. Since, the temporal shift in rainfall is the dominant factor in the climatic variations, in order to better understand and predict the possibility of a drought disaster and plan necessary water management activities it is important to carry out precipitation conveyance and modelling. The general conception that droughts occur twice in a decade in Bangladesh is found not applicable through analysis of climate data for the last two decades. However, it is often argued that the recent submission of droughts is a temporal phenomenon which does not indicate a reduction of drought frequency in Bangladesh. The study attempted to assess variability, trends, anomaly, volatility and transition in rainfall in the drought-prone northwest region of Bangladesh to have a better insight on the causes of its occurrence, the possibility of the persistence of present trends in droughts and scope of its predictability. The Innovative Trend Analysis results revealed a fluctuation of trends in different parts of the study period. Analysis of rainfall anomaly also revealed a rapid shift in rainfall regime from wet to dry or dry to wet in a short period. The results indicate that the present trend of declining droughts may not persist. The rainfall region can rapidly shift from present wet longer wet or near wet phase to a severe dry phase and trigger droughts in the region. It was observed that Rain Anomaly fluctuate significantly over time like the rainfall. However, there is no consistency or apparent pattern in variation. The results indicate a sudden occurrence of droughts in the region or the droughts in the region is less predictable. The seasonal Rainfall Anomaly Index values were also calculated, and similar random fluctuations were observed. Therefore, it can be remarked that high variability of rainfall has made the droughts less predictable in northwest Bangladesh. The results indicate that analysis of variability and trends are not enough for a complete understanding of the cause of change in droughts. Besides, such analyses are not enough to understand whether the decreasing trend in droughts will persist in the near future or there will be a rainfall regime shift which would alter drought trends. The novelty of the study is the application of GARCH-Jumps and Markov chain for analysis of predictivity of droughts. The results can indicate whether the recent submission of droughts in Bangladesh is a reduction of droughts or it is a temporary submission and the country may be affected by droughts again in the near future. Another novelty of the study is the application of Normalized Difference Vegetation Index (NDVI) to relate meteorological drought and hydrological drought with land cover changes. It is not necessary the changes in the vegetation was solely influenced by the meteorological and hydrological drought but there could be other possible reasons such as changes in the vegetation pattern, changes in the irrigation system, use of fertilizers , attack from pest or due to some other natural hazards or anthropogenic activities. This indicates that detailed investigations should be carried out to conclude. To verify the less predictivity of droughts revealed in the analysis, volatility and jumping behaviour of rainfall was estimated using GARCH-jump model. Application of GARCH-jump model suggests that the rainfall data are characterized by both volatility dynamics and time-varying jumps. These outcomes indicate that recent rainfall intensity are not stable and are difficult to project. Such findings signal that in future, there might be a volatile level of rainfall and therefore, amount of rainfall could either decrease or increase. Finally, to verify the less predictability of drought scenario a Markov Chain model was developed to evaluate transition behaviour of rainfall and develop drought proneness index. Higher-order transition probabilities (up to 10th step) were calculated which were finally used for the calculation of DI. The results revealed that DI values change significantly within a short period. From application of Markov Chain, eratic behaviour of rainfall pattern observed, indicating drought proneness scenario fluctuate, which form severe, moderate and mild drought process occurs. Hence it is evident that due to the eratic behaviour of rainfall pattern, drought proneness scenario fluctuate in north-west region of Bangladesh. Thus instability of rainfall distorts the predictibility of future variation in drought. The study indicates that droughts in the study area is highly random having large volatility and rapid shifting nature. Large volatility and rapid shifting nature of rainfall have made the predictability of droughts high uncertain in the region. The rainfall region can rapidly shift from present wet longer wet or near wet phase to a severe dry phase and trigger droughts in the region. Analysis can be performed to identify whether there is any specific drought movement path/locus over the country, i.e., to identify the case where drought is initiated in a station in a specific month and then it moves to another station in the next month. The trend of drought movement path can be used as important information for drought warning system in the country. The finding of the study can help disaster risk mitigation policy planning. The procedure used in this study for systematic analysis of rainfall data with robust statistical methods can be replicated for analysis of predictivity of any other climatic phenomena in Bangladesh and any other regions.