Drought prediction is the most effective way to mitigate drought impacts. The current study examined the ability of three renowned machine learning models, namely additive regression (AR), random subspace (RSS), and M5P tree, and their hybridized versions (AR-RSS, AR-M5P, RSS-M5P, and AR-RSS-M5P) in predicting the standardized precipitation evapotranspiration index (SPEI) in multiple time scales. The SPEIs were calculated using monthly rainfall and temperature data over 39 years (1980–2018). The best subset regression model and sensitivity analysis were used to determine the most appropriate input variables from a series of input combinations involving up to eight SPEI lags. The models were built at Rajshahi station and validated at four other sites (Mymensingh, Rangpur, Bogra, and Khulna) in drought-prone northern Bangladesh. The findings indicated that the proposed models can accurately forecast droughts at the Rajshahi station. The M5P model predicted the SPEIs better than the other models, with the lowest mean absolute error (27.89–62.92%), relative absolute error (0.39–0.67), mean absolute error (0.208–0.49), root mean square error (0.39–0.67) and highest correlation coefficient (0.75–0.98). Moreover, the M5P model could accurately forecast droughts with different time scales at validation locations. The prediction accuracy was better for droughts with longer periods.

Countries depending on small-scale agriculture, such as Bangladesh, are susceptible to climate change and variability. Changes in the frequency and intensity of drought are a crucial aspect of this issue and the focus of this research. The goal of this work is to use SPI (standardized precipitation index) and SPEI (standardized precipitation evapotranspiration index) to investigate the differences in drought characteristics across different physiognomy types in Bangladesh and to highlight how drought characteristics change over time and spatial scales when considering different geomorphologies. This study used monthly precipitation and temperature data from 29 metrological stations for 39 years (1980–2018) for calculating SPI and SPEI values. To determine the significance of drought characteristic trends over different temporal and spatial scales, the modified Mann–Kendall trend test and multivariable linear regression (MLR) techniques were used. The results are as follows: (1) Overall, decreasing dry trend was found in Eastern hill regions, whereas an increasing drought trends were found in the in the rest of the regions in all time scaless (range is from − 0.08 decade−1 to − 0.15 decade−1 for 3-month time scale). However, except for the one-month time scale, the statistically significant trend was identified mostly in the north-central and northeast regions, indicating that drought patterns migrate from the northwest to the center region. (2) SPEI is anticipated to be better at capturing dry/wet cycles in more complex regions than SPI. (3) According to the MLR, longitude and maximum temperature can both influence precipitation. (4) Drought intensity increased gradually from the southern to the northern regions (1.26–1.56), and drought events occurred predominantly in the northwestern regions (27–30 times), indicating that drought meteorological hotspots were primarily concentrated in the Barind Tract and Tista River basin over time. Findings can be used to improve drought evaluation, hazard management, and application policymaking in Bangladesh. This has implications for agricultural catastrophe prevention and mitigation.

About 1.0 million ha coastal lands in Bangladesh are mono-cropped suffer from varying degree of soil salinity, waterlogging and climate vulnerability. Low yielding, traditional T. Aman rice is grown only in wet season. Growing non-rice crop after late harvested T. Aman rice is not profitable. This study was aimed to introduce high yielding, short duration T. Aman rice varieties for advancing its harvesting time and to make the avenue for timely establishment of dry season crops. Varietal trials were made at Dacope and Amtali under ACIAR funded project during 2016-2018 and compared with local cultivars. Among tested varieties BRRI dhan76 followed by BRRI dhan77 and BRRI dhan54 in Dacope and BRRI dhan77 followed by BRRI dhan76 and BRRI dhan54 in Amtali were preferred for 0.5–1.0 tha−1 yield advantage and 15–25 days earliness compared to traditional varieties. Early harvesting of T. Aman created the avenue of timely establishment of rice and non-rice crops depending on availability of fresh water and thus crop intensification and land productivity was improved. The new cropping system increased annual rice yield and farmers’ profits by 1.5- to 2-folds compared with traditional system without environmental degradation. This technique can be replicated in similar coastal zones of Bangladesh.

No abstract

Selection of suitable transplanting window is essential for getting desired crop yield and optimizing irrigation water. This study was conducted in four different locations of Bangladesh (Gazipur, Mymensingh, Cumilla and Bogura districts) to investigate the effect of transplanting period on irrigation water productivity during irrigated rice (Boro) cultivation. Ceres-rice model incorporated in DSSAT was used to estimate rice grain yield and agronomic parameters for Boro 2016-17. Daily weather data and soil data were collected from Bangladesh Meteorological Department (BMD) and Soil Resource Development Institute (SRDI). The estimated irrigation scheduling using CROPWAT- 8.0 model was used as input to the DSSAT model. Rainfall distribution showed only about 22% (2% in winter and 20% in pre-monsoon) of annual rainfall occurred in irrigated rice growing period. Delay transplanting after 15 December, the cultivar BRRI dhan28 faced higher mean daily temperature resulted shorter life span. The increased seasonal mean temperature by 2.8ºC in Gazipur and Bogura and 2.6ºC in Mymensingh and Cumilla from 15 December to 01 March reduced growth duration by 24 days in Gazipur, Mymensingh and Bogura and by 26 days in Cumilla district. Cumilla received the maximum rainfall, however Gazipur experienced the lowest among the four study locations. The received rainfall amount increased with the advancement of transplanting date from 15 December. The increased rainfall reduced the irrigation demand of the cultivar. On the contrary, reduced growth duration due to delay transplanting decreased the grain yield. Transplanting up to 1 February produced almost similar grain yield, while irrigation demand decreased from 15 December transplanting. Water productivity showed increasing trend for late transplanting. Considering grain yield and irrigation water productivity, 15 January to 1 February transplanting were found suitable transplanting period for the study locations. Rice crop establishment within the recommended period could be optimized the grain yield and irrigation water productivity in the selected study locations. Thus, maximum coverage of rainfall can be reduced the irrigation demand. Consequently, it may help to optimize groundwater use and to arrest the groundwater mining. Bangladesh Rice J. 25 (2) : 21-30, 2021