Bangladesh is one in all the foremost climate vulnerable countries of the world. In recent years, climate change studies over the country get plenty of attention by the researchers and policy makers. A substantial quantity of global climate change studies over the country use climate models to estimate future projections and uncertainties. Maximum temperature, precipitation and their potential future changes are evaluated in an ensemble of the 5th Phase Coupled Model Inter-comparison Project (CMIP5) within the Intergovernmental Panel on Climate Change (IPCC) diagnostic exercise for the Fifth Assessment Report (AR5) and the available historical data collected by the Bangladesh Meteorological Department (BMD) during the period 1981-2008 in the north-western region of Bangladesh and also the comparison between these two values. It has been found that average maximum temperature shows a positive trend of increase at a rate of 0.29 °C and 5.3 °C per century respectively, for BMD data and MPI-ESM-LR (CMIP5) model data. But the rainfall is decreasing at a rate of 8.8 mm and 40.1 mm per century respectively for BMD data and MPI-ESM-LR (CMIP5) model data. It is seen that July was the maximum monsoon rainfall month and January was the lowest rainfall month. The peak frequency is slightly smaller than 12 months, which indicates that the major events are occurring before ending a year compared to the previous year. According to MPI-ESM-LR (CMIP5) model data, future normal temperature on north-western region will be increased at a rate of 1.62 °C during the period 2040-2100.
The Inter-Governmental Panel on Climate Change (IPCC) defines climate change as a change in the state of the climate that can be identified (e.g. using statistical tests) by changes in the mean and/or the variability of its properties and that persists for an extended period, typically decades or longer [
Annual mean maximum temperature will increase to 0.4˚C and 0.73˚C by the year of 2050 and 2100 severally [
The north-western part of Bangladesh denotes the Rajshahi Division and Rangpur Division. Generally, it is the area lying west of Jamuna River and north of Padma River, and includes the Barind Tract. There are six meteorological stations are situated in this region which are Bogra, Rangpur, Dinajpur, Ishurdi, Rajshahi and Sydpur.
Daily maximum temperature and normal daily rainfall data were collected from Bangladesh Meteorological Department (BMD) for the period 1981-2008. But
Station Name | Station ID | Latitude | Longitude |
---|---|---|---|
Dinajpur | 10,120 | 25.65 | 88.67 |
Rangpur | 10,208 | 25.73 | 89.27 |
Rajshahi | 10,320 | 24.37 | 88.70 |
Bogra | 10,408 | 24.87 | 89.35 |
Ishurdi | 10,910 | 24.13 | 89.50 |
Sydpur | 41,858 | 25.78 | 88.89 |
the data during the period 1981-1990 for the station of Sydpur was not available so that only for this station the value is calculated from 1991-2008. The missing data of temperature have been filled by inverse distance weighting method (IDW). Missing data of rainfall has been filled by data of the neighbor station.
On the other hand, to get and prepare MPI-ESM-LR (CMIP5) model data for calculation is complex and takes several processes. First of all, downloaded the Netcdf file of the data and then crack the NC file by using ArcMap software. Daily maximum temperature and normal daily rainfall data were collected for north-western region of Bangladesh for the period 1981-2008. But the rainfall data was available for the period 1981-2005. The MPI-ESM-LR (CMIP5) model data gives the same value for Rangpur, Dinajpur and Sydpur due to very near longitude and latitude and also same for the station Rajshahi and Ishurdi.
Trend analysis is the prediction of future outcome by using historical result. Increasing or decreasing trend of all the independent weather parameters (e.g. annual and seasonal temperature, rainfalls, sunshine etc.) were statistically examined in two phases. First one is the using of non-parametric Mann-Kendall test and second one is the nonparametric Sens slope estimator. The increasing or decreasing trend was tested based on normalized test statistics (Z) value. When Z is positive, trend is said to be increasing and when Z is negative, it is said to be decreasing. The trend’s slope gives the annual rate and direction of change [
The Mann-Kendall trend test is a non-parametric way for identifying trends in data collected over time series. Mann-Kendall Statistic (S) is given by,
S = ∑ ∑ s i g n ( X i − X j ) (1)
here, i = 2 , 3 , ⋯ , n ; j = 1 , 2 , ⋯ , i − 1 and
s i g n ( X i − X j ) = { 1 , if X i − X j > 0 0 , if X i − X j = 0 − 1 , if X i − X j < 0 (2)
For a sample size > 10, a normal approximation to the Mann-Kendall test may be used. For this, variance of S is obtained as,
V ( s ) = n ( n − 1 ) ( 2 n + 1 ) − ∑ t p ( t p − 1 ) ( 2 t p + 5 ) 18 (3)
here, p = 1 , 2 , ⋯ , q
where tp is the number of ties for the p th value and q is the number of tied values.
Then standardized statistical test is computed by:
s i g n ( X i − X j ) = { S − 1 V ( s ) , if S > 0 0 , if S = 0 S + 1 V ( s ) , if S < 0 (4)
The magnitude of the trend is estimated by Sens slope method [
Q ′ = x t ′ − x t t ′ − t (5)
where, Q ′ is the slope between data points x t ′ and x t , x t ′ is the data measurement at time t ′ and x t is the data measurement at time t.
Sens slope estimator is simply given by the median slope,
B = { Q ′ N + 1 2 , N is odd 1 2 ( Q ′ N 2 + Q ′ N + 2 2 ) , N is even (6)
where, N is the number of calculated slopes. A positive value of B indicates an increasing trend and a negative value indicates a decreasing trend in the time series.
In this study to represent the confidence level ***, **, * and + signs have been used to represent 100%, 99%, 95% and 90% level of confidence respectively.
Some researchers recommend the Shapiro-Wilk test as the best choice for testing the normality of data [
W = ( ∑ i = 1 n a i x i ) 2 ∑ x = 1 n ( x i − x ¯ ) 2 (7)
where, xi is the ith order statistic, i.e., the ith-smallest number in the sample mean, ai is the constant. The correlation coefficient determines the strength of linear relationship between two variables. It always takes a value between 1 and +1, with 1 or 1 indicating a perfect correlation. A correlation coefficient close to or equal to zero indicates no relation-ship between the variables. A positive correlation coefficient indicates a positive (upward) relationship and a negative correlation coefficient indicates a negative (downward) relationship between the variables.
In this chapter outcomes, daily maximum temperature and rainfall records of 28 years (1981-2008) have been analyzed based on temperature and rainfall of BMD data and MPI-ESM-LR (CMIP5) model data for the North-Western (Bogra, Rangpur, Dinajpur, Ishurdi, Rajshahi and Sydpur) region of Bangladesh. It is seen from
Maximum Temperature (˚C) | Rainfall (mm) | |||||||
---|---|---|---|---|---|---|---|---|
Mean | SD | Mean | SD | |||||
Season | BMD | MPI-ESM-LR | BMD | MPI-ESM-LR | BMD | MPI-ESM-LR | BMD | MPI-ESM-LR |
Winter | 25.43 | 31.49 | 1.85 | 2.99 | 10.57 | 3.45 | 15.87 | 6.73 |
Pre-monsoon | 32.82 | 44.66 | 1.74 | 3.18 | 107.03 | 35.93 | 119.85 | 49.18 |
Monsoon | 32.32 | 41.50 | 0.82 | 2.36 | 349.65 | 137.65 | 166.86 | 89.72 |
Post-monsoon | 30.27 | 35.63 | 1.29 | 2.73 | 82.72 | 38.73 | 123.30 | 47.52 |
respectively in conformity with BMD data and MPI-ESM-LR (CMIP5) model data while post-monsoon average maximum temperature was 30.27˚C and 35.63˚C with SD 1.29˚C and 2.73˚C respectively according to BMD data and MPI-ESM-LR (CMIP5) model data. Highest 349.65 mm and 137.65 mm with SD 166.86 mm and 89.72 mm was observed in monsoon season while post-monsoon rainfall was 82.72 mm and 38.73 mm with SD 123.30 mm and 47.52 respectively according to BMD data and MPI-ESM-LR (CMIP5) model data. The temperature may be increasing due to rapid industrialization, greenhouses gases such as carbon dioxide, methane, nitrous oxides, chlorofluorocarbon (CFC).
It is seen from
Seasonal Mann-Kendall trend and Sens slope are shown in
Season | BMD (%) | MPI-ESM-LR (%) |
---|---|---|
Winter | 1.92 | 1.60 |
Pre-monsoon | 19.46 | 16.65 |
Monsoon | 63.58 | 63.80 |
Post-monsoon | 15.04 | 17.95 |
Maximum Temperature(˚C) | Rainfall (mm) | |||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Linear Slope | P-Value | P.N.T | Linear Slope | P-Value | P.N.T | |||||||
Season | BMD | MPI | BMD | MPI | BMD | MPI | BMD | MPI | BMD | MPI | BMD | MPI |
Winter | −0.011 | 0.049 | 0.352 | 0.248 | Yes | Yes | −0.386 | −0.413 | 0.0001 | 0.008 | No | No |
Pre-monsoon | 0.015 | 0.057 | 0.0001 | 0.0001 | No | No | −0.349 | −0.284 | 0.625 | 0.005 | Yes | No |
Monsoon | 0.024 | 0.037 | 0.856 | 0.207 | Yes | Yes | −0.345 | −1.629 | 0.996 | 0.219 | Yes | Yes |
Post-monsoon | 0.008 | 0.059 | 0.167 | 0.762 | Yes | Yes | 1.378 | 1.464 | 0.0004 | 0.770 | No | Yes |
Maximum Temperature (˚C) | Rainfall (mm) | |||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Z-test | Significance | Sens Slope | Z-test | Significance | Sens Slope | |||||||
Season | BMD | MPI | BMD | MPI | BMD | MPI | BMD | MPI | BMD | MPI | BMD | MPI |
Winter | −0.61 | 3.30 | *** | −0.009 | 0.125 | −1.28 | −1.05 | −0.255 | −0.080 | |||
Pre-monsoon | −0.38 | 0.81 | −0.009 | 0.029 | 0.35 | −0.54 | 0.446 | −0.350 | ||||
Monsoon | 2.75 | 0.85 | ** | 0.024 | 0.023 | 0.68 | −1.24 | 1.307 | −1.504 | |||
Post-monsoon | 1.24 | 1.52 | 0.011 | 0.067 | 0.96 | 2.41 | * | 1.795 | 1.478 |
Positively changed for monsoon and post-monsoon for BMD data while all seasons are positively changed for MPI-ESM-LR (CMIP5) model data. Besides that, during the taken period, only winter rainfall has decreased 1.28 mm per year for BMD data and for MPI-ESM-LR (CMIP5) model data only post-monsoon rainfall has increased 2.41 mm which is significant.
Significance of the trend is assessed using a Z value, where negative and positive scores of Z denote downward and upward trends respectively.
Average of the temporal variation of the 336 months average maximum temperature
record is 30.4˚C and 38.8˚C while rainfall record is 159.73 mm and 62.19 mm respectively according to BMD data and MPI-ESM-LR (CMIP5) model data. It seems from the time series that there is an extreme event above the mean value almost every year. The spectrum of monthly maximum temperature and rainfall variation both for BMD data and MPI-ESM-LR (CMIP5) model data is shown in
For average maximum temperature, the periodogram shows three prominent peaks according to BMD data and two peaks in accordance to MPI-ESM-LR (CMIP5) model data while for rainfall two prominent peaks according to BMD data and one peak in accordance to MPI-ESM-LR (CMIP5) model data. It may be seen from the periodogram that, as a rough approximation the first two periodicity can be taken as significant and the others may be neglected for average maximum temperature while only first peak can be identified as significant for both data and the others may be neglected for rainfall. The period (in Months) corresponding to the first peak of the value of 0.084 is compound by 1/0.084 results in periodicity of 12 months and the peak at 0.168 results in periodicity of 6
months both for average maximum temperature and rainfall according to BMD data but there is no such peak in MPI-ESM-LR (CMIP5) model data for six months periodicity for rainfall. According to BMD data the third peak at 0.251 results in a periodicity of four months but there is no such peak in MPI-ESM-LR (CMIP5) model data. However, the peak frequency is slightly smaller than one year which indicates that the major events are occurring before ending a year compared to the previous year. The reason for shifting the higher temperature and rainfall events may be climate change; though it is not logical to conclude that without more analysis.
Monthly average maximum temperature and rainfall variation was filtered through a band pass filter (0.080 to 0.086 per months) compared with the original monthly variation and displayed in
The climatology of maximum temperature and rainfall based of BMD and MPI-ESM-LR (CMIP5) model datasets is presented in
Correlation bet. max. temp. and rainfall | |||
---|---|---|---|
BMD | MPI-ESM-LR | ||
Pearson | r | −0.064 | −0.055 |
P-value | 0.761 | 0.796 | |
Spearman | ρ | −0.093 | −0.117 |
P-value | 0.659 | 0.577 | |
Kendall | τ | −0.042 | −0.210 |
P-value | 0.767 | 0.123 |
Changes of the mean annual temperature from 2040 till 2100 are shown in
An examination on air temperature and precipitation behavior is important for short-term planning and the prediction of future climate conditions. Trends in precipitation and temperature at annual, seasonal and monthly time scales for the periods of 1981-2008 have been analyzed using BMD data and MPI-ESM-LR (CMIP5) model data. Also, the results herein form a good basis of future studies
on temperature variability. Considering all seasons (winter, pre-monsoon, monsoon and post-monsoon), maximum temperature has increased significantly in all seasons except winter which is insignificant over the whole study area for BMD data but for MPI-ESM-LR (CMIP5) model data maximum temperature is on increase in the region. On average, temperature over the entire region increased by 0.29˚C and 5.3˚C per century respectively for BMD data and MPI-ESM-LR (CMIP5) model data. It has clearly found that maximum temperature has been increased dramatically during the period of 1981-2008. In future, the normal temperature will be beyond 31˚C. The highest future normal temperature has occurred in Bogra at 8˚C per century. For north-western region, the highest 63.58% and 63.80% of total rainfall have occurred in monsoon respectively in accordance with BMD data and MPI-ESM-LR (CMIP5) model data. In Dinajpur, the highest percent of rainfall occurred in monsoon for both BMD data and MPI-ESM-LR (CMIP5) model data. Winter rainfall has decreased 1.28 mm per year for BMD data and for MPI-ESM-LR (CMIP5) model data only post-monsoon rainfall has increased 2.41 mm which is significant. Only 1.92% rainfall has occurred in winter in conformity with BMD data and only 1.60% rainfall has occurred in winter in conformity with MPI-ESM-LR (CMIP5) model data. The lowest rainfall occurred in Sydpur with 8.30 mm in winter for BMD data and only 2.17 mm rainfall occurred in Rajshahi and Ishurdi for MPI-ESM-LR (CMIP5) model data. Overall, the rainfall is decreasing in north-western region. MPI-ESM-LR (CMIP5) temperature projections are larger than BMD climate projection but MPI-ESM-LR (CMIP5) precipitation projections are smaller than BMD climate projections. Though there is a difference between the values of BMD and MPI-ESM-LR (CMIP5), the decisions are almost the same for a small region like north-western region of Bangladesh. That’s why, considering new MPI-ESM-LR (CMIP5) projections will be more helpful for decision makers as they have comparatively better representation of earth’s physical processes.
The authors express their appreciation to Mawlana Bhashani Science and Technology (MBSTU) research Cell for providing financial support that foster research. Special thanks go to Bangladesh Meteorological Department (BMD), World Climate Research Programme (WCRP) and Fifth Phase of Coupled Model Intercomparison Project (CMIP5) for providing data used in the study.
Bhuyan, M.D.I., Islam, M.M. and Bhuiyan, M.E.K. (2018) A Trend Analysis of Temperature and Rainfall to Predict Climate Change for Northwestern Region of Bangladesh. American Journal of Climate Change, 7, 115-134. https://doi.org/10.4236/ajcc.2018.72009