Trend analysis of Global Climate Model EdGCM downscaled data for forecasting Boro rice yield in north-eastern region of Bangladesh

Abstract

The inherent relationship between global climate change extremes and crop production is used in this research to explore key management techniques to minimize the losses of agricultural production in the flash flood vulnerable areas in Bangladesh. An investigation is conducted at Sylhet Agricultural University to forecast Boro rice yield through the trend analysis of downscaled climatic data and observed data for North-eastern district, Bangladesh. Normalized difference vegetation index has been used to identify the vulnerability of the water body, build-up area and vegetation from 2015 to 2018 due to flash flood. Due to the lack of adequate weather data, Educational Global Climate Model was simulated to perform downscaling through a proper validation technique. As downscaled data has more satisfactory response than local data, it was adopted and simulated with rice yield data from Bangladesh Rice Research Institute, Bangladesh and Bureau of Statistics to reveal the relationship. To explore the climatic scenario, the trend analysis of climatic parameters from 1968 to 2020 was done. Mann–Kendall trend testing and Sen's slope estimation are also performed. Changes in temperature are significant for both downscaled and local result of p and the value less than α = 0.05. As the computed p value is greater than the significance level α = 0.05, both rainfall data as well as local relative humidity showed non-significant result. This study revealed that the monthly average temperature is decreasing from 1968 to 2007 and considerably increasing during 2008–2020. The trend of monthly average rainfall is increasing from 1968 to 2020. Furthermore, the trend of monthly average relative humidity is increasing from 1968 to 1987 whereas decreasing during 1988–2007. In this regard, relative humidity shows an increasing trend at the period of 2008–2020. Analyzing the yield with downscaled and local data for each year, the appropriate or suitable range of rainfall, temperature and relative humidity for high and low yield is determined. The range of rainfall, temperature, and relative humidity are 10–12 mm, 19–22 °C and 52–60%, respectively, for high yield, whereas rainfall, temperature and relative humidity range 7–9.5 mm, 25–27 °C and 60–65% are considered for low yield. Findings of the study evaluates that HYV of Boro will produce high yield in 2018 based on the above range of rainfall, temperature, and relative humidity. A similar condition is expected in the case of 2020 and yield will have expected to be the highest. Furthermore, HYV Boro will produce satisfactory yield in 2019 due to the change of climatic variations.