Precipitation Concentration in Bangladesh over Different Temporal Periods

Precipitation concentration is an important component of climate, and an unbalanced distribution of precipitation can yield excess or scarcity of water resources, which in turn can influence plant growth, flood risk, and water resource use. 0e precipitation concentration index (PCI) is a well-known indicator for the measurement of temporal precipitation in a short or long area. 0e purpose of this study was to analyze precipitation concentration rates in different regions of Bangladesh using the precipitation concentration index (PCI) and the inverse distance weighting method. In this study, the rainfall data from 30 meteorological observatory stations across Bangladesh were collected for the period 1980 to 2011. We defined periods of varying lengths (i.e., annual, supraseasonal, seasonal, and threeand two-month rainfall concentrations) and compared their PCI values. 0e results showed that precipitation concentrations were mostly irregular when rainfall was concentrated within two to four months of the year. Higher PCI values were mainly identified in the eastern region and have strong seasonal influences, whereas lower PCI values were mostly observed in the northern region. 0e analyses of periodic variation and precipitation in Bangladesh generally follow through the SW–NE direction due to the summer monsoon, while during the winter monsoon, they follow the N–S direction where JAS and JFM showed higher and lower PCI values. We observed variations in PCI among different regions using the Kruskal–Wallis test of the mean PCI on a decadal scale (1980–1989, 1990–1999, and 2000–2011). 0e result showed that significant changes in the precipitation occurred during the period of 1980–2011. At a two-month scale, significant changes were identified during transition periods where PCI values were lower from 2000 to 2011 than those in the earlier decades.


Introduction
Climate change due to global worming is a major concerning issue in the world.Precipitation is changing on global and regional scales by the influence of warming [1][2][3][4].For regions prone to hydrological disasters (e.g., Bangladesh), the implications of these changes are particularly significant [5].It has been suggested that water cycle modification is one of the most notable consequences of global atmospheric warming [6].Precipitation intensity, amounts, and patterns are expected to change, and extreme weather events such as droughts and floods are likely to occur more frequently [7].e amount and intensity of precipitation can influence soil erosion, slope instability, plant growth, and agricultural practices [8,9] as well as modify fluvial systems, groundwater recharge, water disposal, and hydroelectric projects [10,11].Precipitation conveyance and measurement can provide a better understanding of the exact mechanisms driving these phenomena for disaster prediction and water management [12].
e IPCC-2007 report found a precipitation increase from 1900 to 2005 north of latitude of 30 ° [13], and Bangladesh, which is one of the most flood-prone areas of the world, was highlighted as one of the most vulnerable countries [14].e variation of precipitation distribution in Bangladesh is a unique characteristic for climatic change.In the Bay of Bengal, the hydrological cycle is dominated by very high spatial and temporal variability; and predictions for the end of the 21st century show a high degree of uncertainty [15].Due to the monsoon season, precipitation concentration increased dramatically while that in the winter season decreased.For the case of dynamic combination, generally Bangladesh faced extreme floods and longer periods of droughts [5,16,17].
e northwestern part of the country, drought is a common phenomenon and became a growing concern in the recent years.
e report confirmed a precipitation decrease in tropical areas and an increase in areas affected by drought in tropical and subtropical areas since the 1970s [13].Several studies in the world indicate that global warming has strong influence on precipitation variation as well as extreme weather events like floods, droughts, and thunderstorms [18,19,20].Bangladesh is renowned for agricultural-based country where about 80% of people are directly or indirectly involved in agricultural activities as their main profession [21].Precipitation and its intensity are important factors on the Bangladesh climate for agricultural production [22].Due to global warming, hydrological changes are the most significant impacts of climate change in Bangladesh.e study of precipitation concentration intensity on seasonal variation is essential for agriculture, disaster mitigation, and water resources planning and management in Bangladesh in the context of global climatic change [23].
Although many studies in Bangladesh have considered general rainfall and temperature characteristics, few have studied the intensity and distribution of rainfall in this region [22].Furthermore, most research has been performed on an annual or six months basis and smaller temporal-scale analyses are lacking.Although there is no single research method that can comprehensively examine changes in precipitation over time, it is worth considering how different methods can be applied to certain situations [24].For example, some studies have applied a precipitation variability focus in regions influenced by extreme precipitation events [25].Several straightforward indicators can be used to assess variability and analyze hydrological processes [26].Among them, the precipitation concentration index (PCI) [27] is recommended since it provides information on long-term total variability of the rainfall amount [26,28,29].Iskander et al. [22] used the PCI to show an irregular precipitation distribution across Bangladesh, especially in the southeast region, and other trend analysis studies have examined long-term changes in precipitation over Bangladesh [30,31].
is study analyzed the spatial and temporal precipitation variability using the PCI calculated from dense precipitation datasets collected by the Bangladesh Meteorological Department (BMD).A geographic information system (GIS) was used for the development of a regional variation map to which we applied the inverse distance weighted method.
e PCI was calculated for annual, supraseasonal, seasonal, and three-and two-month data over a period of 30 years (1980-2011).

Study Area.
Bangladesh is a low-lying river-dominated country consisting primarily of flat plains.It occupies an area of 147,570 km 2 and geographically extends from 20 °34′ N to 26 °38′ N and from 88 °01′ E to 92 °41′ E. Most of the population lives in rural areas and directly or indirectly depends on agricultural activities.A sub-tropical monsoon climate is strongly active, and the country is sometimes affected by tropical cyclones [32].e country experiences a vast amount of rainfall annually.On average, 78% of rainfall in Bangladesh occurs during the monsoon and is caused by weak tropical depressions that are carried from the Bay of Bengal by wet monsoon winds [33,34].Precipitation concentration varies within the country, and the mean annual precipitation varies from a maximum of 5,690 mm in the northeast to a minimum of 1,110 mm in the west [21].Bangladesh has 30 meteorological observatories (Figure 1), all of which were included in this study.To better understand rainfall variation, the country was divided into four regions: north-northwest, central, east-southeast, and southsouthwest.

Study Process.
In this study, precipitation concentration changes at different temporal intervals were compared for the four regions of Bangladesh.Data were collected from the BMD from 1980 to 2011 and used as the input data for PCI analysis.Dataset quality is an important indicator for obtaining good results and was carefully controlled before the data were released.Generally, all stations were found to be homogeneous; therefore, the datasets from each station were included in this analysis.First, we performed a comparative correlation between the PCI values for the two-and three-month temporal intervals to understand specific variations.
en, we defined uniform, moderate, irregular, and highly irregular precipitation distributions and seasonal variations for specific regions.For the spatial analysis of precipitation, we interpolated data from all weather stations onto a regular grid.Finally, we analyzed the annual, supraseasonal, seasonal, and threeand two-month PCI values using the Kruskal-Wallis test to identify statistical significance among the decades.e major objective of this research was to analyze rainfall trend and irregular distribution over the study area at various temporal intervals.e geographic information system (GIS) tools have been used for the PCI mapping.Under the geostatistical analysis, the inverse distance weighted (IDW) method was considered in the study to interpolate the PCI data of the country.In interpolation, the power was considered as 2, the searching neighborhood was standard, the neighborhoods were at least 10 and neighbors to include was 15, major semiaxis was 1.52, minor semiaxis was 1.52, and the 2 Advances in Meteorology angle was 0. e inverse distance weighting (IDW) is an established deterministic method for the precipitation concentration index mapping and one of the most frequently used deterministic models in spatial interpolation [35][36][37].

PCI Methods.
PCI is an indicator of rainfall concentration and rainfall erosivity [28].e PCI values were calculated on an annual scale for each grid point using data from all stations based on the equations of Oliver [27]: where P i is the total precipitation in month i. e PCI was also calculated on a supraseasonal scale (Equation ( 2)) from January-June (JJ) and July-December (JD) and on a seasonal scale (Equation ( 3)) for the winter (November-February, NDJF), summer or premonsoon (March-June, MAMJ), and monsoon (July-October, JASO): We also calculated the three-and two-month PCI intervals (Equations ( 4) and (5), respectively) to break down the annual, seasonal, and supraseasonal precipitation distribution and concentration: As defined by Oliver [27], the PCI values of less than 10 represent a uniform precipitation distribution (i.e., low precipitation concentration), the PCI values from 11 to 15 denote a moderate precipitation concentration, the values from 16 to 20 denote an irregular distribution, and the values above 20 represent strong irregularity (i.e., high precipitation concentration).According to the analysis of the PCI formulae, on annual, seasonal, and supraseasonal scales, a PCI value of less than 10 indicates perfect uniformity of precipitation distribution (i.e., the same or highly similar amounts of rainfall in each month).At all scales, a PCI value of 16-20 indicates that the total precipitation is concentrated in half of the period, and a PCI value greater than 20 indicates that the total precipitation is concentrated in onethird of the period.For example, this last value indicates that most annual precipitation occurs within four months, most supraseasonal precipitation occurs within two months, and most seasonal precipitation occurs within one month.In other words, intensive rainfall is considered to be uniform if it occurs throughout the time period, but if it occurs within a narrow subset of the time period, then its distribution is considered to be irregular.
To investigate the temporal changes in PCI, the PCI values were calculated for ten decadal subperiods: 1980-1989, 1990-1999, and 2000-2011.e statistical significance of the difference between decadal periods at each grid point was assessed using a trend test with different levels of probability categorized as follows: exceptionally likely (p < 0.01), extremely likely (p > 0.05), very likely (p < 0.10), and very low probability (p > 0.10).

Annual PCI.
e annual PCI values were moderate with some irregularities (Figure 2).During the six-month analysis, the JJ (January-June) period was very irregular, whereas the JD (July-December) period showed a moderate precipitation distribution.A moderately irregular distribution was observed when the four regions shared similar climatic conditions.e distribution of the annual PCI varied from <14 in the north-east (along Sylhet, Srimangal) to >18 in the north and south-east (Figure 3(a)) near the Bay of Bengal, where a moderate-to-irregular precipitation distribution was observed.e western and central areas showed significant variations and irregular distributions.

Supraseasonal PCI.
e PCI on a supraseasonal scale demonstrated a complex precipitation distribution across the study area.During the JJ period, the PCI continued to show moderate-to-high irregularities.e north-west and south-southeast areas had higher PCI values (>20).e central (excluding Dhaka) and west areas also showed comparatively higher values, while low but irregularly distributed values were observed in the central to north-east regions (Figure 3(b)).During the summer (MAMJ), the central and western regions experienced strong storms, locally known as "Kalbaishaki" which are characterized by high rainfall and the potential for devastating damage.For the JD period, lower PCI values (<14) were observed over most of the study area, except for a small region in the south-east near Cox's Bazar (Figure 3(c)) and were characterized by a moderately irregular precipitation distribution.

Seasonal PCI.
ree distinct seasons are recognized in Bangladesh: the premonsoon or summer season (MAMJ), the monsoon/rainy season (JASO), and the postmonsoon or winter season (NDJF) [38].Normally, the premonsoon and postmonsoon seasons receive little rainfall.On average, 80% or more of the country's annual rainfall occurs during the monsoon season.For all regions, the postmonsoon/monsoon seasons showed strongly irregular/uniform precipitation distributions (Figures 3(d) and 3(f)).Rainfall in Bangladesh during the monsoon season is driven by tropical depressions in the Bay of Bengal [33].ese monsoon depressions have a south-to-north trajectory, with turns towards the northwest and west deflected by the Meghalaya Plateau.Owing to the height of this plateau, uplifted moisture condenses, leading to particularly heavy rainfall over northwestern Bangladesh [33].A moderate precipitation distribution was most common in the premonsoon season, although the north-west and south-east areas showed comparatively higher values (>12) than the west, central, and northeast areas (Figure 3(e)).Temperature has a significant influence on rainfall; during summer, temperatures can reach 45 °C or greater, and in winter, they can fall to 5 °C in some areas [21].

ree-Month PCI.
e three-month PCI values were highest in the north and south-east and lowest in the central and north-east regions, depending on the annual weather conditions.e JFM period showed irregular PCI values for all regions, with higher values in the south-east than those in the north-west (Figure 3(g)).
e AMJ period showed moderately uniform premonsoon PCI values (Figure 3(h)), which is a period that normally experiences increasing rainfall from April to reach moderate levels during May and June [39].Early rainfall during this period can sometimes cause flooding.e lowest PCI values during the JAS period were less than 10 (Figure 3(i)) and were characterized by uniform precipitation concentrations across the study area.Moreover, the periods of JFM and OND showed strong irregularities in the PCI values corresponding to the dry season [39].Overall, the north-west area showed the highest PCI values (>20), while the central, south, and east areas had moderate values and irregular distributions.Comparatively lower values were observed in the south-east, near the Chittagong Hill Tracts (Figure 3(j)), where the distribution was moderate.

4
Advances in Meteorology

Two-Month PCI.
e results from two-month data analysis showed relatively moderate-to-uniform precipitation distributions and no strong irregularities for any region (Figure 2).Moderate rainfall was distributed relatively evenly in JF, FM, MA, and AM, and MJ, JJ, JA, and AS were especially uniform.Meanwhile, SO, ON, ND, and DJ had moderate-to-irregular distributions.
e analysis showed that rainfall concentrations were most heterogeneous in nonrainfall seasons (JF, FM, MA, AM, ON, and ND) and in the transition period before and after the monsoon.For all regions, precipitation was most regular during the monsoon season (Figures 4(e)-4(h)).For the JF period, the south-southeast region showed a more moderate precipitation distribution than the north-west region, while the opposite was observed for the ON period (Figures 4(a) and 4(j)).During the SO period, some anomalies were found in the northwestern, central, and southeastern areas (Figure 4(i)).ese are consistent areas of high altitude where soil conditions and growth capabilities differ from the rest of the country.

Decadal Analysis of PCI.
e decadal analysis showed that the annual PCI values varied over time and exhibited an irregular distribution.Significant changes in the PCI values between 1980 and 1989 and between 1990 and 1999 were found for the extreme northwestern and southeastern regions.
e west-central and northeastern regions had similar values, but from 1990 to 1999, the central region changed significantly.From 2000 to 2011, the PCI changed significantly with the northwestern and southeastern regions displaying irregularities, while the central

3.7.
Signi cant Di erences in Regional PCI.We observed di erences in the mean annual precipitation among the four regions.e decadal analysis revealed a signi cant di erentiation during the decades of 1980-1989 and 1990-1999, while 2000-2011 showed no signi cant di erences (Table 1).For the supraseasonal analysis, signi cant di erentiation was observed during 1990-1999 for the JJ period and during 2000-2011 for JD period (Figures 7 and 8).
Di erentiation was observed for MAMJ in all decades except 2000-2011.Di erentiation was observed for NDJF 8 Advances in Meteorology have adverse impacts on lives and livelihoods of millions of people [4].A positive trend in annual rainfall and negative trend in winter rainfall (Figures 9 and 10) means that the rainfall of Bangladesh is concentrating in the monsoon and premonsoon months [5].e annual trend of PCI has increased by maximum in majority of the stations for most of the periods whereas the SupraSeasonal PCI of July-December has decreased in two third of the stations (Table 2).Moreover,  Advances in Meteorology

Implications of Variation of PCI in Bangladesh.
Rainfall variability in space and time is one of the most relevant characteristics of the climate of Bangladesh where hydrological disaster is a common phenomenon [5].e volume, timeliness, distribution, and duration of rainfall in each season are major concerns to farmers.Most of the people in our country area depend on crop and livestock production for livelihood and household incomes [42].Rainfall-related hazards and disasters (like drought and ood) require farmers to implement measures in altering  Advances in Meteorology 11  12 Advances in Meteorology the cropping system, planting or sowing times, crop varieties, and change crop types in the country [43].e PCI can be used as an indicator of hydrological hazard risks such as oods and droughts [44].PCI can help quantifying the relative distribution of the rainfall patterns [28].

Conclusion
Bangladesh experiences a monsoon climate and has a oodplain topography, making it prone to hydrological disasters.In this study, we considered the rainfall distribution over di erent spatial and temporal scales from 1989 to 2011.Advances in Meteorology e results showed that precipitation is the most uniform during the monsoon season.For all regions, the annual PCI values were moderate and irregular, and the six-month (JJ and JD) PCI values were irregular to moderate.For the seasonal and three-month analyses, the values were generally moderately irregular.Finally for shorter twomonth intervals (MJ, JJ, JA, and AS), most PCI values showed a uniform precipitation distribution.In general, the northwestern and southeastern areas of Bangladesh showed the highest PCI values, re ecting an irregular precipitation distribution.
e results showed that the rainfall in Bangladesh is moderately seasonal, with an intensity increasing from the premonsoon to the monsoon season, driven by weak tropical depressions from the Bay of Bengal.Most  Advances in Meteorology premonsoon rainfall occurs in the central and western regions, where agricultural production is extensive.e decadal analysis (1980-1989, 1990-1999, and 2000-2011) showed that the annual PCI values have varied over time and exhibited irregular distributions.
e north-west and south-east regions had strong irregularities during 2000-2011, while the shorter JJ and JD six-month periods showed signi cant and moderate irregularities, respectively.e seasonal analysis showed that the JASO period exhibited a uniform subtropical monsoon climate; therefore, the three-month JAS period also showed a uniform distribution.
e two-month PCI analyses indicated that precipitation variation was mostly found in seasonal transition periods.For short time periods, precipitation was strongly linked to season.Short-term variations also increased during the decade 2000-2011, during which the OND period saw the most irregular precipitation distribution.16 Advances in Meteorology

Figure 1 :
Figure 1: Map of the study area.
same conditions were found during Monsoon periods.A negative trend of PCI means concentration of rainfall during monsoon in some parts of Bangladesh is increasing.erefore, this indicates that the irregularities of Rainfall during Monsoon Season (June-September) are less.During Winter Season (November-February) a positive trend of PCI values are evident.erefore this indicates that the irregularities of Rainfall during Winter Season are high (Figure11).
a Letters denote the months of the year (e.g., JJ is January-June and JD is July-December).* In nity of number.