@article { author = {Roy, Binata and Islam, G M Tarekul and Islam, A.K.M. Saiful and Bhattacharya, Biswa and Khan, Md. Jamal Uddin}, title = {Trend analysis of pre-monsoon flash floods for the northeast Haor region of Bangladesh to assess the impact on Boro crop productivity}, journal = {Water Productivity Journal}, volume = {2}, number = {2}, pages = {5-36}, year = {2022}, publisher = {Veresk Ettesal}, issn = {2717-3062}, eissn = {2821-1103}, doi = {10.22034/wpj.2021.272369.1030}, abstract = {Introduction: Pre-monsoon (March-May) flash floods observed in the northeast of Bangladesh, known as the "Haor" (saucer or bowl-shaped large tectonic depression) region, have drawn much attention in recent years due to early onset, high frequency, large magnitude, and destructive nature. The Boro crop, which is the primary agricultural production of this region, is adversely damaged by the flash floods. In this study, the trend of the flash floods of the northeast Haor region has been studied to understand past changes and future occurrences and to assess the overall impact on Boro crop productivity.    Material and methods: The trend analysis was carried out on the observed 3 - hourly water level data of 13 hydrological gage stations and daily rainfall data of 2 meteorological gage stations of the Haor region collected from the Bangladesh Water Development Board (BWDB) and Bangladesh Meteorological Department (BMD), respectively. All these stations were located near the Bangladesh-India border, where the flash flood water comes first from the surrounding hilly areas of India; otherwise, the flashy nature may not be discernible in the time series data. The data were processed from 1st March to 15th May, considering the start of the flash flood season and the end of the Boro crop harvesting period, respectively. The statistical Mann-Kendall (MK) test has been used to analyze the trend at a 5% significance level, and Sen's slope has been used to compute the magnitude of the trend at the same significance level. Lag-1 autocorrelation has been determined before statistical trend analysis. The MK test was directly applied to the data with no autocorrelation, whereas the data showing significant autocorrelation were analyzed through the Trend-Free Pre-Whitening Mann-Kendall (TFPWMK) test. TFPWMK was selected because it can effectively detect trends in significantly serially correlated hydrological data. The trend of early onset of flash floods has also been assessed. Based on the trend analysis results, the impact of flash floods on the key varieties of Boro crops produced in the Haor region was evaluated. Results: The trend analysis results showed that the trends of maximum water level and relative water level varied from station to station. The study revealed that 6 and 9 stations out of the 13 hydrological stations, respectively, showed an increasing trend in maximum water level and relative water level, among which only 2 and 3, respectively, showed a statistically significant increasing trend at a 5% significance level. More increasing trends were found for the relative water level, which meant that though the peak of flash floods was not increasing, the relative water level in the pre-monsoon season was increasing in recent years compared to past ones. The rainfall trend was also increasing, though it was not statistically significant. Overall, the vulnerability of the Boro crop was increasing. The trend analysis on peak periods of flash floods showed a decreasing trend, which revealed an early onset of flash floods in recent years. The peak of the flash floods was found to be arriving early in late March-early April (instead of late April-early May), coinciding with the harvesting period of the Boro crop. This early onset of the flash flood warrants catastrophic damage to the Boro crop in future flash floods. The study further showed that the current Boro varieties- BRRI dhan28, BRRI36, BRRI dhan69, BRRI dhan88 were safer in 'normal flash floods' (late April-early May) but not anymore safer in 'early flash floods' (late March-early April) experienced in recent years. Conclusion: The study found an increasing trend in the water level, and rainfall in the Haor region. Though the trend hasn’t crossed the limit of a 5% significance level in most stations, this could increase exponentially in the future due to climate change impact. Also, changing timelines and the early onset of the flash flood can destroy Boro cultivation in the Haor area and imbalance the food security of the local people. Hence, the changing trend of flash floods in magnitude, pattern, duration, and early onset must be considered for flood risk management and sustainable agricultural production in the Haor region. For example, more varieties of Boro should be introduced, which can be harvested by the last week of March before the flash flood hits the Haor region. Also, varieties with a shorter growth duration and cold tolerance at the reproductive phase should be prioritized.}, keywords = { Agriculture,Auto-correlation,BRRI dhan28,Crop Productivity,Mann-Kendall,Onset,Sen’s slope,Trend Free Pre-Whitening}, url = {https://www.waterproductivity.net/article_157580.html}, eprint = {https://www.waterproductivity.net/article_157580_b104bd00ca8db923827d456dbe1081af.pdf} }