Trending Regional Precipitation Distribution and Intensity: Use of Climatic Indices ()
1. Introduction
Climate variability and its unfriendly concomitants are of growing concern worldwide. The climate of Bangladesh is undergoing changes and extremes are becoming more unprecedented every year. Floods, cyclones or droughts are often in its colossal mood here, which are aggravated by climate change and its variability being experienced more frequently in Bangladesh than ever before. Change of precipitation trends is in its way because of climate change. According to the intergovernmental panel on climate change (IPCC) report [1] , Bangladesh will experience 5% to 6% increase of rainfall by 2030. Bangladesh has been termed by IPCC as one of the most vulnerable countries in the world, which is prone to climate change. Global warming induced changes in temperature and rainfall are already evident in many parts of the world, as well as in this country [2] . High intensity of rainfall and uneven temporal and spatial distribution is creating flooding and longer dry spells are evoking droughty conditions [3] . As a whole, these have profound impact on agriculture [4] and food security of the country will be at risk [5] .
In the past three decades, higher Precipitation Concentration Index (PCI > 20) has been recorded in semi-arid and tropical humid environments in the cardamom hill slopes, south western ghats, India, which implies very long dry periods for up to 3 - 5 months [6] . During the last 50 years, Simple Daily Intensity Index (SDII) values for the countries of the western Indian Ocean have been showing significant negative trend for most of the observatory stations and positive trend in only a few of the stations [7] . Specifically for Bangladesh, although a number of studies have been carried out on average precipitation and temperature pattern [8] -[11] , research on intensity and distribution of rainfall has hardly been done.
2. Data and Methodology
Daily and monthly precipitation data of 35 BMD (Bangladesh Meteorological Department) stations (Figure 1) were used for this study. For the convenience of analysis, whole study area was divided into five regions as North Western Region (NWR), North Eastern Region (NER), Central Region (CR), South Western Region (SWR) and South Eastern Region (SER) (Figure 1).
The Simple Daily Intensity Index (SDII) [7] [12] is defined as follows:
Let RRwj be the daily precipitation total for the wet days W (RR ≥ 1 mm) in period j. If W is the number of the wet days within the period j, then:
(1)
The Precipitation Concentration Index [6] [13] , being proposed as an indicator of rainfall concentration and rainfall erosivity [14] , was calculated on an annual scale for each station according to Equation (2):
(2)
where Pi is the monthly precipitation in month i and Pt is the annual precipitation total. Table 1 lists different value ranges of SDII and PCI with their relative significances.
Mann-Kendall test [15] [16] was used to analyze the trends in annual rainfall total, annual number of rainy days, SDII and PCI; in addition, the Sen’s slope method [17] was applied to determine the magnitude of change. Confidence level of 95% was taken as threshold to classify the significance of positive and negative trends.
3. Results and Discussion
SDII and PCI both are extreme precipitation indices which measure intensity and distribution uniformity of rainfall events respectively. SDII depends on annual rainfall amounts and annual number of rainy days. Therefore, trends of both annual rainfall amounts and annual number of rainy days have effects on trends of SDII. Mann-Kendall trend test values for annual rainfall amounts, annual number of rainy days, SDII and PCI are shown in Table 2 and Figure 2. Results show insignificant negative trends in 15 stations, significant negative trends in 4 stations, insignificant positive trends in 14 stations and significant positive trends in 2 stations for annual rainfall amounts. For annual rainy days, the numbers are 11, 0, 15, and 9; for SDII the numbers are 17, 6, 12, and 0 and for PCI the numbers are 15, 6, 14, and 0 respectively.
(a) (b)
Figure 1. (a) Study area with regions: Bangladesh (b) BMD stations.
Figure 2. Summary of the results obtained for the trends of the four study parameters using Mann-Kendall test.
Table 2. Trends of annual rainfall total, annual number of rainy days, SDII and PCI for the study period at different BMD stations.
The asterisk (*) denotes statistically significant figures at 95% confidence level.
Intensive rainfall can be uniform if it occurs all over the time period, but if it occurs within a short instance of the time period then its distribution can be called irregular. Accordingly, when both SDII and PCI shows similar trend, rainfall will become uniform and less intensive (decreasing trend) or irregular and intensive (increasing trend). On the contrary, when the two trends are opposite, outcomes may be uniform and intensive or irregular and less intensive. Thus, in NWR and NER rainfall tends to be uniform and less intensive (Table 2). In other regions mixed effects are evident.
Six cases of SDII trend change related to the trend of annual rainfall amounts and annual number of rainy days are detected in this study (shown in Table 3). In case-I, trends of SDII increase although trends of both annual rainfall amounts and annual number of rainy days decrease. It may be due to the milder trend declination of annual rainfall amounts compared to the annual number of rainy days. In case-II, increasing trend of annual rainfall amounts is steeper than that of annual number of rainy days. Cases III and VI (Figures 3(a)-(c)) are
easy to understand as Equation (1) explains. Cases IV and V are the opposite of cases I and II.
The values of PCI shows irregular to strongly irregular precipitation distribution in North Western Region (NWR) and mostly strong irregularity in South Eastern Region (SER); also irregular distribution is evident in Central Region (CR), North Eastern Region (NER) and South Western region (SWR) (Table 4). While the values of SDII shows moderate to high precipitation intensity in NWR, moderate to strong intensity in CR and
Table 3 . Cases of SDII change (1971-2010).
Table 4. Average value of SDII and PCI for the study period at different BMD stations and their significance according to values (Table 1).
NER, mostly strong intensity in SER and low to very strong intensity in SWR (Table 4).
Spatial distribution of trends for annual rainfall amounts (mm), annual number of rainy days, SDII and PCI over the study period of 1971-2010 is shown in Figure 4. Annual rainfall amount (mm) decreases over the central region (maximum significant negative trend of −24.41 mm/year in Chandpur), and increases in south eastern region (maximum significant positive trend of 19.13 mm/year in Teknaf) (Figure 4(a)) and mixed changes prevail over other regions. Annual number of rainy days increases over eastern and northern parts (maximum significant positive trend of 0.889 day/year in Srimangal) and decreases in central part (Figure 4(b)). SDII shows mixed spatial trends and decrease over northern and southern part; also in central part decreasing pattern is prevalent (Figure 4(c)). Trends of PCI (Figure 4(d), Figure 3(d)) show decrease over northern and south eastern parts and increase over south western region.
As a whole, this paper shows a guideline to apply climate indices in detecting intensity and temporal distribution of precipitation pattern on regional scale. Our analysis yields some grimacing results for South Eastern Region (SER) and a little part of South Western Region (SWR) (Sitakunda, Sandwip, Hatya) since precipitation is of high intensity and strong irregularity there. Such irregular and highly intensive precipitation can have high erosion-potential. This may reduce the elevation of these regions over mean sea level making it prone to being engulfed by the sea as sea water surface is continually rising due to climate change [1] .
Due to erosion, silt charges in river water increase. Consequently conveyance capacity of river gets low as silts settle down to the river bed. This can cause longer lasting floods. Moreover, due to sudden highly intensive rainfall, flash floods and water clogging may occur in cities due to the lack of drainage facilities. Bangladesh is a country which largely depends on its agriculture and hydroelectricity. So the findings of this study can also be useful for developing various decision-support tools in different hydrologic, ecological and agricultural applications.
Acknowledgements
Contribution of Bangladesh Meteorological Department (BMD) for providing the necessary meteorological data is acknowledged here.
NOTES
*Corresponding author.