Seismic Status in Bangladesh

Vietnam Journal of Earth Sciences, 40(2), 178-192, Doi:10.15625/0866-7187/40/2/12266 178 (VAST) Vietnam Academy of Science and Technology Vietnam Journal of Earth Sciences Seismic Status in Bangladesh Syed Mustafizur Rahman1*, Md. Habibur Rahman2, Md. Omar Faruk3, and Md. Sultan-Ul-Islam4 1Department of Applied Physics and Electronic Engineering, University of Rajshahi, Rajshahi 6205, Bangladesh 2CEGIS, Dhaka 1212, Bangladesh 3Department of ICE, Pabna University of Science and Technology, Pabna 6600, Bangladesh 4Institute of Environmental Science, University of Rajshahi, Rajshahi 6205, Bangladesh Received 12 February 2018; Received in revised form 01 April 2018; Accepted 5 April 2018 ABSTRACT Seismic status in Bangladesh has been investigated using earthquake data recorded by the global network of USGS during 1980 to 2016. Seismicity parameters such as magnitude of completeness 𝑀𝑐, 𝑏-value and a-value are being estimated. It has observed that the overall 𝑏-value in and around Bangladesh is of 0.84, which is seemed to be seismically active zone. As, reliable 𝑏-value assessment can lead to better seismic hazard analysis, reliable magnitude of completeness 𝑀𝑐 can lead to 𝑏-value assessment of an area, this work has dealt and estimated magnitude of com- pleteness 𝑀𝑐 using various techniques for the whole region for a reliable estimation. Estimated 𝑀𝑐 is obtained to be around 3.9-4.7, which lead to 𝑏-value of 0.93. Spatial variations of 𝑀𝑐 and 𝑏-value have been investigated for 1 o×1o horizontal and vertical rectangular regions for the study area between 18-29°N and 84-95°E. Estimated 𝑀𝑐 and 𝑏- value along with 𝑎-value are then averaged for the common regions in the pair of horizontal and vertical regions. Re- sults are then being presented in the form of maps. The findings resemble as, the 𝑀𝑐 is low at the border line of N-W Bangladesh, and a line from Cox’s Bazaar to Sylhet through Hill tracts. Remain parts belong to the 𝑀𝑐 value of 4.1- 4.2, thus the 𝑏-value obtained is varying from 0.68 to 1.2, where, the value is higher at region in Chittagong and Barisal division that extends toward north through part of Dhaka to Sylhet and lower at Rajshahi, Rangpur and part of Khulna division, while 𝑎-value is varying from 5.0 to 7.2 mostly from west to east. Keywords: earthquake; seismicity; magnitude; completeness. ©2018 Vietnam Academy of Science and Technology 1. Introduction * Earthquake is one of the most natural dev- astating events that can hurl people around and destroy lives and properties. The study of earthquake distribution in space and time in a region is known as seismicity. Seismic activi- ties are being referred to frequency and mag- nitude of earthquakes experienced over a pe- riod of time. Realistic assessment of seismic activities in Bangladesh may assist to reduce *Corresponding author, Email: smrahman@ru.ac.bd the risk from this catastrophic disaster. Earth- quake catalogues in this regard are the only sources as the most important products for studying seismological activities those can support to understand earthquake physics and let to learn seismotectonics, seismicity or seismic hazard of an area. Even in modern time it is still difficult to obtain most reliable catalogues. Earthquake catalogue is basically the result of recorded signals of seismometers and processed by a variety of techniques and assumptions (Zuniga and Wiemer, 1999), hence adequate care should have been taken Syed Mustafizur Rahman, et al./Vietnam Journal of Earth Sciences 40 (2018) 179 to assess the quality, consistency or homoge- neity before using it to scientific analyses (Hafiez, 2015). In order to avoid such com- plexities, the present analysis intends to work with one catalog for better uniformity. The frequency-magnitude distribution (FMD) of earthquakes introduced by Guten- berg and Richter (1944) known as G-R law is the basis as well as the basic relation for any seismicity studies. In order to understand meaningful interpretation of frequency- magnitude distribution in an earthquake cata- log, the magnitude of completeness, 𝑀𝑐 is de- fined as the minimum magnitude above which all earthquakes within a certain region are re- liably recorded (Naylor, et al., 2010). The G- R law is written as below. 𝑙𝑜𝑔10𝑁(𝑀) = 𝑎 − 𝑏𝑀 (1) where, 𝑀 is the magnitude, 𝑁(𝑀) is the num- ber of earthquakes occurred in a specific time with magnitudes 𝑀 ≥ 𝑀𝑐 , 𝑎 is the earth- quake productivity, and 𝑏 describes the rela- tive distribution of small and large earth- quakes. The 𝑏-value in the Gutenberg-Richter power law is an indicator which describes seismic status of an area. However, there are difficulties to determine reliable 𝑏-value (Felzer, 2006), particularly setting magnitude of completeness 𝑀𝑐 which can lead to im- proper 𝑏-value estimation unless 𝑀𝑐 is deter- mined properly. This research work intends to estimate 𝑏-value and magnitude of complete- ness 𝑀𝑐 in Bangladesh using the earthquake catalogs retrieved from USGS (USGS, 2012). Few initiatives were being taken in the past to define seismic hazard map, earthquake cata- log, national building code, peak ground ac- celeration and seismicity analysis in Bangla- desh (GSB, 2018; Siddique, 2015; Al- Hussaini, 2006). However, the works are yet to seem as much more meaningful inputs. In order to estimate meaningful seismicity in Bangladesh a location map and epicenters of occurred earthquakes over the years in the study area are shown in Figure 1. In addition, there are several plausible ex- planations in the observation of variations in 𝑏-values according to tectonic or geologic set- ting of an area. Therefore, a description of the geological overview of the study area has in- corporated in the following section. 2. Geological Setting of Bangladesh Bangladesh belongs to South Asia and lies between 20°34’-26°38´N and 88°01’- 92°41’E. The area of the country is approxi- mately 147,570 km2 with more than 710 km long coastlines. It covers about 80% of the Bengal Basin. The land area is following a downward slope of 1-2° from north-west to south-east direction. Tectonic framework of the region is shown in Figure 2 that entails the existence of plate boundaries, shelf, fault, trough, threshold, long hinge zone and the complicated river basin system. Physiographically, the study area is divid- ed into: Territory Hilly regions (east and north-eastern frontier), Pleistocene Terraces (N-W and central part), Tippera surface, Tista Fan (north eastern part), Floodplains and Del- taic plain of the Ganges-Brahmaputra-Meghna delta complex, Sylhet Depression and Inland marshes (scattered all over Bangladesh) etc. (Rashid, 1991; Reimann, 1993). Holocene un- consolidated sediments (sands, silts, clays, gravels and peats) from a few hundred to thousands of meters cover the Floodplains and the Delta. The whole basin area is criss- crossed by several basement controlled fault configuring the present structural and geo- morphic setup of the country (Hunting, 1981). The Bengal Basins are bounded in the north by the Dauki fault and Bangladesh-Burma subduction zone in the east. Beside these sev- eral faults like hinge zone, Bogra fault, Gan- ges and Jamuna lineaments, Korotoya fault are prominent structures can trigger the earth- quakes in the region. 3. Data and Methods This work has used the source parameters of earthquake data of the study area for the duration from 1990-2016 recorded by USGS using global seismic network. Under Earth- quake Hazards Program, USGS has been rec- Vietnam Journal of Earth Sciences, 40(2), 178-192 180 orded the millions of earthquakes in the world. It is believed that the ANSS Compre- hensive Earthquake Catalog (ComCat) is a re- liable source in the world. Earthquake data are downloaded from USGS for the present re- search as shown in Figure 1. Figure 1. Study area and the map of earthquake epicenters during 1990-2016, retrieved from USGS Syed Mustafizur Rahman, et al./Vietnam Journal of Earth Sciences 40 (2018) 181 Figure 2. Tectonic framework of Bangladesh (after Banglapedia, 2012) 3.2. Magnitude of Completeness Magnitude of completeness 𝑀𝑐 is the min- imum magnitude at where most of the earth- quakes preferably 100% in a space-time vol- ume are detected. Assessment of a correct magnitude of completeness 𝑀𝑐 is crucial since too high value of 𝑀𝑐 can lead to under- sampling by discarding usable data, while too low value can lead to erroneous or biased seismicity parameters by using incomplete da- ta (Mignan and Woessner, 2012). A number of contributions have provided various techniques to compute 𝑀𝑐 upon valid- ity of the G-R law (Wyss et al., 1999; Wiemer and Wyss, 2000; Cao and Gao, 2002; Woess- ner and Wiemer, 2005; Amorese, 2007). Computation of 𝑀𝑐 is straightforward and based on readily accessible parametric catalog data. The most basic way is to estimate 𝑀𝑐 by fitting a G-R model to the observed frequen- cy-magnitude distribution. The magnitude at where the FMD departs from the G-R law is taken as an estimate of 𝑀𝑐 (Zuniga and Wyss, 1995). In few cases a visual evaluation could lead to a correct estimate of completeness magnitude. On the contrary, it has been seemed difficulties in visual estimation of completeness (Naylor et al., 2010). Spatio- temporal heterogeneities can cause to change in 𝑀𝑐, which is being observed in frequency magnitude distributions (Wiemer and Wyss, 2000 and Mignan et al., 2011). There are both opinions that FMD has been observed as to be scaled as approximately magnitude 0 event or the events which can be only detected within 10 m form the source (Abercrombie and Brune, 1994), on the other hand, few contribu- tors have suggested changes in scaling at higher or smaller magnitude events (e.g., Lomnitz-Adler and Lomnitz, 1979; Utsu, 1999 and Aki, 1987). However, the changes in slope of G-R model are not seemed to be rele- vant for the estimate of 𝑀𝑐. It is believed that dominant factor changing the slope of G-R model is incompleteness in reporting for smaller magnitudes (Wiemer & Wyss, 2000). The work to be done here is slightly different as small and/or very small (<3.0 M) events are not available from the catalogues to be used but magnitude completeness 𝑀𝑐 and 𝑏-value are to be learned for the study area. In this context the popular techniques to estimate 𝑀𝑐 are being employed to observe the 𝑀𝑐 in the present analysis. The techniques based on va- lidity of G-R law are being explained below. 3.2.1. Maximum Curvature Technique (MAXC) The Maximum Curvature (MAXC) tech- nique (Mignan and Woessner, 2012, Wyss et al., 1999 and Wiemer and Wyss, 2000) is non parametric technique but fast and straightfor- ward way to estimate 𝑀𝑐 and consists in de- fining the point of the maximum curvature by computing the maximum value of the first de- rivative of the frequency-magnitude curve (FMD). 𝑀𝑐 = 𝑚 while, 𝜕(𝑁(𝑚) 𝜕𝑚 = 𝑚𝑎𝑥 (2) In practice, this matches the magnitude bin with the highest frequency of events in the non-cumulative FMD. Despite the easy ap- plicability of this approach 𝑀𝑐 can be under- estimated in the case of gradually curved FMDs. Vietnam Journal of Earth Sciences, 40(2), 178-192 182 3.2.2. Goodness-of-Fit Test (GFT) The Goodness-of-fit test (GFT) proposed by Wiemer and Wyss (2000), calculates 𝑀𝑐 by comparing the observed FMD with syn- thetic ones. The goodness-of-fit is evaluated by the parameter 𝑅, absolute difference of the number of events in each magnitude bin be- tween the observed and synthetic G-R distri- butions. Synthetic distributions are calculated using estimated 𝑎-value and 𝑏-value of the observed dataset for 𝑀 ≥ 𝑀𝑐𝑜 as a function of ascending cutoff magnitude 𝑀𝑐𝑜. 𝑅(𝑎, 𝑏, 𝑀𝑐𝑜) = 100 − ( ∑ |𝐵𝑖−𝑆𝑖| 𝑀𝑚𝑎𝑥 𝑀𝑐𝑜 ∑ 𝐵𝑖𝑖 100) (3) where, 𝐵𝑖 and 𝑆𝑖 are the observed and predict- ed cumulative number of events in each mag- nitude bin. 𝑀𝑐 is found at the first magnitude cutoff at which the observed data for 𝑀 ≥ 𝑀𝑐𝑜 is modeled by a straight line for a fixed confidence level, e.g. 𝑅= 90% or 95%. 3.2.3. 𝑀𝑐 by 𝑏-value stability (MBS) Cao and Gao (2002) estimated 𝑀𝑐 using the stability of the 𝑏-value as a function of cutoff magnitude 𝑀𝑐𝑜, referred to as MBS by Woessner and Wiemer (2005). This model is based on the assumption that 𝑏-value esti- mates ascend for 𝑀𝑐𝑜 < 𝑀𝑐 and remain con- stant for 𝑀 ≥ 𝑀𝑐𝑜. If 𝑀𝑐𝑜 < 𝑀𝑐, the resulting 𝑏-value is incorrect. As 𝑀𝑐𝑜 approaches 𝑀𝑐, the 𝑏-value approaches its true value and re- mains constant for 𝑀𝑐𝑜 > 𝑀𝑐. 𝑀𝑐 is defined as the magnitude for which the change in 𝑏-value ∆𝑏 between two succes- sive magnitude bins is smaller than 0.03. Woessner and Wiemer (2005) have shown that this principle is unstable since the fre- quency of events in single magnitude bins can vary strongly. In order to satisfy such objec- tive measure and to stabilize numerically, Woessner and Wiemer (2005) have used the 𝑏-value uncertainty 𝜕𝑏 according to Shi and Bolt (1982) as: 𝜕𝑏 = 2.3𝑏2√ ∑ (𝑀𝑖−〈𝑀〉)2 𝑁 𝑖=1 𝑁(𝑁−1) (4) with 〈𝑀〉 being the mean magnitude and 𝑁 the number of events. 𝑀𝑐 is then defined as the first magnitude increment at which ∆𝑏 = |𝑏𝑎𝑣𝑒 − 𝑏| ≤ 𝜕𝑏 (5) The arithmetic mean 𝑏𝑎𝑣𝑒 is calculated from b-values of successive cutoff magnitudes 𝑀𝑐 in half a magnitude range 𝑑𝑀 = 0.5 such as 𝑏𝑎𝑣𝑒 = ∑ 𝑏(𝑀𝑐𝑜)∆𝑚/𝑑𝑀 𝑀𝑐𝑜+𝑑𝑀 𝑀𝑐𝑜 (6) for a bin size ∆𝑚 = 0.1. Large magnitude ranges are preferable, and would be justified for FMDs that perfectly obey a power-law. 3.3.4. 𝑀𝑐 from the Entire Magnitude Range (EMR) Entire magnitude range (EMR) method in- cludes the events below 𝑀𝑐. This method con- sisting of two parts: the G-R law for the com- plete part and the cumulative normal distribu- tion for the incomplete part of the non- cumulative FMD. The model attempts to re- produce the entire frequency-magnitude dis- tribution, thus fits the incompletely observed part. The EMR approach is explained as the non-cumulative FMD can be described by the intensity λ (normalized number of events) at magnitude 𝑚 as 𝜆(𝑚) = 𝜆𝑜(𝑚)𝑞(𝑚) (7) with 𝜆𝑜(𝑚|𝛽) = 𝑒 −𝛽𝑚 where, 𝛽 = 𝑏𝑙𝑜𝑔10 and 𝑞(𝑚) is a detection function with 0 ≤ 𝑞 ≤ 1 . 𝑞 is commonly de- fined as the cumulative normal distribution of mean 𝜇 and standard deviation 𝜎 (Ogata and Katsura, 1993, 2006 and Iwata, 2008), where 𝑞(𝑚|𝜇, 𝜎) = ∫ 1 √2𝜋𝜎 𝑒 −(𝑥−𝜇)2 2𝜎2 𝑑𝑥𝑚 −∝ (8) Equation 6, (using Eqs. 7-8) provides a model to fit the FMD over the entire magni- tude range where the magnitude completeness is only implicit with 𝑀𝑐(𝑛) = 𝜇 + 𝑛𝜎 (9) where 𝑛 indicates the confidence level. 𝑛 = 0, means that 50% of the events are detected Syed Mustafizur Rahman, et al./Vietnam Journal of Earth Sciences 40 (2018) 183 above 𝑀𝑐, similarly 𝑛 = (1,2,3) means that 68%, 95% and 99% of the events are detected respectively. The parameters 𝜃 = (𝛽, 𝜇, 𝜎) are simultaneously obtained by maximizing the log-likelihood function 𝑙𝑜𝑔𝐿(𝜃) = ∑ 𝑙𝑜𝑔𝑓(𝑚𝑖|𝜃)𝑖 with the normalized density function 𝑓(𝑚|𝜃) = 𝑐𝜆(𝑚|𝜃), 𝑐 being a normalization factor. The model becomes as following (Ogata and Katsura, 2006): 𝑓(𝑚|𝛽, 𝜇, 𝜎) = 𝛽𝑒 (−𝛽(𝑚−𝜇)−𝛽2 𝜎2 2 ) 𝑞(𝑚|𝜇, 𝜎) (10) 4. Seismic Status Estimation Spatial variation of seismicity parameters 𝑀𝑐 and 𝑏-value of the study area has been es- timated using the Eqs.1-10. In order to ob- serve variations of the parameters, the study area was divided into twelve uniform horizon- tal and twelve uniform vertical rectangular re- gions as shown in Figure 3 to assess seismici- ty parameters for each rectangular regions. It is believed that the average value would re- flect the seismic status of the common region as shown (C. cell) in Figure 3 for the pair of horizontal and vertical rectangular regions. Figure 3. Schematic diagram of 12 horizontal (H1-12) and 12 vertical (V1-12) rectangular regions and common re- gion as common cell for vertical and horizontal rectangular pair for the assessment of seismicity parameters 4.1. Estimation of Seismicity in Bangladesh Figure 4 shows the frequency magnitude distribution (FMD), cumulative frequency dis- tribution (CFD) and linear fitting of G-R law of earthquake events retrieved from USGS as shown in Figure 1 for the whole study area. The 𝑏-value and 𝑎-value are being obtained as 0.84 and 6.54 respectively from the analysis. This is the primary and overall estimation of the study area. As mentioned earlier that 𝑀𝑐 means a great deal for proper estimation of 𝑏- value. In order to study a reliable estimation of 𝑀𝑐 four techniques as mentioned earlier in Eqs. 2-10 are applied to present catalog and the results of 𝑀𝑐 estimation, are shown in Ta- ble 1 and in Figures 5(a-d). Estimated magnitude of completeness 𝑀𝑐 as shown in Figure 5 is varying from 3.8-4.4 (Ta- ble 1). Catalog used does not contain low or very low magnitude events. Rather it contains the events of the study area greater than magni- tude 3.1. If the highest 𝑀𝑐 is being considered for further analysis the number of total events significantly decreased. On other hand 𝑀𝑐 es- timations using all the techniques are seemed to be around 4.0. Since spatial variation of seismic status of the study is one of the impetuous be- hind the work, this work has been intended to keep the 𝑀𝑐 as low as possible. As a result the maximum number of events can be involved in the estimation of seismicity. In this line MAXC technique is appeared to be the right choice in this analysis. Hence, using 𝑀𝑐=3.9 obtained through MAXC the FMD, CMD and linear G-R Vietnam Journal of Earth Sciences, 40(2), 178-192 184 fitting over CMD once again have been esti- mated for the whole study area and shown in Figure 6. Estimated 𝑏- and 𝑎-values are of 0.93 and 7.1 respectively, where 𝑏-value is found to be close to 1.00 which reiterates the area as seismically active zone. Figure 4. Earthquake magnitude distribution of the study area a) FMD and b) CFD and linear fitting of G-R law Table 1. Estimated magnitude of completeness using different techniques Techniques MAXC MBS GFT EMR Estimated Mc 3.9 3.8 4.4 4.2 Figure 5. Estimated magnitude of completeness 𝑀𝑐 using a) MAXC, b) GFT, c) MBS and d) EMR techniques Syed Mustafizur Rahman, et al./Vietnam Journal of Earth Sciences 40 (2018) 185 Figure 6. Estimation of 𝑏-value for Bangladesh using 𝑀𝑐=3.9. a) normalized frequency magnitude and cumulative frequency distributions, b) linear fitting of G-R law 4.2. Spatial Variation of Seismicity in Bang- ladesh In order to observe spatially distributed 𝑀𝑐 and 𝑏-value the study area has divided into eleven horizontal and five vertical rectangular regions as explained in Figure 3. Objective behind the consideration of horizontal and vertical rectangles is to cover most seismicity effect from all directions. Seismicity estima- tions apparently may mislead as to be estimat- ed for horizontal and vertical cells, however, seismicity parameters are to be derived for common regions of the pair of horizontal and vertical rectangles over the study area. In addition, contour or surface map to be derived using seismicity parameters for common re- gions would influence the nearby regions. The scheme would have also allowed a little com- putational advantage. Separating data according to rectangular regions from the main earthquake catalog magnitude of completeness 𝑀𝑐s are computed and shown in Table 2. Using computed 𝑀𝑐s for the horizontal and vertical rectangular re- gions, 𝑏-value and 𝑎-value are also estimated as shown in Table 2. Later the average for the common regions of the pair of horizontal and vertical rectangles, 𝑀𝑐, a-value and 𝑏-value are being estimated and shown in Table 3. Table 2. Estimated seismicity parameters 𝑀𝑐 , 𝑎-value and 𝑏-value for the horizontal (a) and rectangular (b) regions (a) (b) Horizontal rectangular regions Vertical rectangular regions Lat oN Long oE N. of Events Mc b-value a-value Lat oN Long oE N. of Events Mc b-value a-value 18 85-95 40 4.2 0.90 5.2 18-29 84 85 4.0 0.58 3.8 19 85-95 95 4.5 0.98 6.0 18-29 85 219 3.9 1.40 8.0 20 85-95 71 4.2 0.77 4.8 18-29 86 163 3.9 0.70 4.8 21 85-95 131 3.9 0.91 5.8 18-29 87 130 3.9 1.10 6.6 21 85-95 131 3.9 0.91 5.8 18-29 87 130 3.9 1.10 6.6 22 85-95 261 4.1 1.10 7.1 18-29 88 88 3.9 0.86 5.3 23 85-95 322 3.9 0.97 6.3 18-29 89 45 4.6 0.67 4.0 24 85-95 338 3.9 0.86 6.0 18-29 90 131 4.1 1.20 7.0 25 85-95 187 3.9 0.93 5.9 18-29 91 132 4.8 1.20 7.3 26 85-95 205 4.1 1.20 7.1 18-29 92 255 4.0 1.10 6.8 27 85-95 502 3.9 0.92 6.3 18-29 93 268 4.2 1.30 7.7 28 85-95 215 3.9 0.77 5.2 18-29 94 667 3.9 1.10 7.1 29 85-95 240 4.0 1.10 6.8 18-29 95 423 4.2 0.83 5.9 Syed Mustafizur Rahman, et al./Vietnam Journal of Earth Sciences 40 (2018) 186 Table 3. Spatial distribution of seismicity parameters, varying with latitude (19-30)°N and longitude (85-96)°E Lat oN Long oE Mc b-value a-value 18.50 84.50 4.10 0.74 4.50 18.50 85.50 4.05 1.15 6.60 18.50 86.50 4.05 0.80 5.00 18.50 87.50 4.05 1.00 5.90 18.50 88.50 4.05 0.88 5.25 18.50 89.50 4.40 0.79 4.60 18.50 90.50 4.15 1.05 6.10 18.50 91.50 4.50 1.05 6.25 18.50 92.50 4.10 1.00 6.00 18.50 93.50 4.20 1.10 6.45 18.50 94.50 4.05 1.00 6.15 18.50 95.50 4.20 0.87 5.55 19.50 84.50 4.25 0.78 4.90 19.50 85.50 4.20 1.19 7.00 19.50 86.50 4.20 0.84 5.40 19.50 87.50 4.20 1.04 6.30 19.50 88.50 4.20 0.92 5.65 19.50 89.50 4.55 0.83 5.00 19.50 90.50 4.30 1.09 6.50 19.50 91.50 4.65 1.09 6.65 19.50 92.50 4.25 1.04 6.40 19.50 93.50 4.35 1.14 6.85 19.50 94.50 4.20 1.04 6.55 19.50 95.50 4.35 0.91 5.95 20.50 84.50 4.10 0.68 4.30 20.50 85.50 4.05 1.09 6.40 20.50 86.50 4.05 0.74 4.80 20.50 87.50 4.05 0.94 5.70 20.50 88.50 4.05 0.82 5.05 20.50 89.50 4.40 0.72 4.40 20.50 90.50 4.15 0.99 5.90 Lat oN Long oE Mc b-value a-value 20.50 91.50 4.50 0.99 6.05 20.50 92.50 4.10 0.94 5.80 20.50 93.50 4.20 1.04 6.25 20.50 94.50 4.05 0.94 5.95 20.50 95.50 4.20 0.80 5.35 21.50 84.50 3.95 0.75 4.80 21.50 85.50 3.90 1.16 6.90 21.50 86.50 3.90 0.81 5.30 21.50 87.50 3.90 1.01 6.20 21.50 88.50 3.90 0.89 5.55 21.50 89.50 4.25 0.79 4.90 21.50 90.50 4.00 1.06 6.40 21.50 91.50 4.35 1.06 6.55 21.50 92.50 3.95 1.01 6.30 21.50 93.50 4.05 1.11 6.75 21.50 94.50 3.90 1.01 6.45 21.50 95.50 4.05 0.87 5.85 22.50 84.50 4.05 0.84 5.45 22.50 85.50 4.00 1.25 7.55 22.50 86.50 4.00 0.90 5.95 22.50 87.50 4.00 1.10 6.85 22.50 88.50 4.00 0.98 6.20 22.50 89.50 4.35 0.89 5.55 22.50 90.50 4.10 1.15 7.05 22.50 91.50 4.45 1.15 7.20 22.50 92.50 4.05 1.10 6.95 22.50 93.50 4.15 1.20 7.40 22.50 94.50 4.00 1.10 7.10 22.50 95.50 4.15 0.97 6.50 23.50 84.50 3.95 0.78 5.05 23.50 85.50 3.90 1.19 7.15 Lat oN Long oE Mc b-value a-value 23.50 86.50 3.90 0.84 5.55 23.50 87.50 3.90 1.04 6.45 23.50 88.50 3.90 0.92 5.80 23.50 89.50 4.25 0.82 5.15 23.50 90.50 4.00 1.09 6.65 23.50 91.50 4.35 1.09 6.80 23.50 92.50 3.95 1.04 6.55 23.50 93.50 4.05 1.14 7.00 23.50 94.50 3.90 1.04 6.70 23.50 95.50 4.05 0.90 6.10 24.50 84.50 3.95 0.72 4.90 24.50 85.50 3.90 1.13 7.00 24.50 86.50 3.90 0.78 5.40 24.50 87.50 3.90 0.98 6.30 24.50 88.50 3.90 0.86 5.65 24.50 89.50 4.25 0.77 5.00 24.50 90.50 4.00 1.03 6.50 24.50 91.50 4.35 1.03 6.65 24.50 92.50 3.95 0.98 6.40 24.50 93.50 4.05 1.08 6.85 24.50 94.50 3.90 0.98 6.55 24.50 95.50 4.05 0.85 5.95 25.50 84.50 3.95 0.76 4.85 25.50 85.50 3.90 1.17 6.95 25.50 86.50 3.90 0.82 5.35 25.50 87.50 3.90 1.02 6.25 25.50 88.50 3.90 0.90 5.60 25.50 89.50 4.25 0.80 4.95 25.50 90.50 4.00 1.07 6.45 25.50 91.50 4.35 1.07 6.60 25.50 92.50 3.95 1.02 6.35 25.50 93.50 4.05 1.12 6.80 25.50 94.50 3.90 1.02 6.50 25.50 95.50 4.05 0.88 5.90 26.50 84.50 4.05 0.89 5.45 26.50 85.50 4.00 1.30 7.55 26.50 86.50 4.00 0.95 5.95 26.50 87.50 4.00 1.15 6.85 26.50 88.50 4.00 1.03 6.20 26.50 89.50 4.35 0.94 5.55 26.50 90.50 4.10 1.20 7.05 Syed Mustafizur Rahman, et al./Vietnam Journal of Earth Sciences 40 (2018) 187 26.50 91.50 4.45 1.20 7.20 Lat oN Long oE Mc b-value a-value 26.50 92.50 4.05 1.15 6.95 26.50 93.50 4.15 1.25 7.40 26.50 94.50 4.00 1.15 7.10 26.50 95.50 4.15 1.02 6.50 27.50 84.50 3.95 0.75 5.05 27.50 85.50 3.90 1.16 7.15 27.50 86.50 3.90 0.81 5.55 27.50 87.50 3.90 1.01 6.45 27.50 88.50 3.90 0.89 5.80 27.50 89.50 4.25 0.80 5.15 27.50 90.50 4.00 1.06 6.65 27.50 91.50 4.35 1.06 6.80 27.50 92.50 3.95 1.01 6.55 27.50 93.50 4.05 1.11 7.00 27.50 94.50 3.90 1.01 6.70 27.50 95.50 4.05 0.88 6.10 28.50 84.50 3.95 0.68 4.50 28.50 85.50 3.90 1.09 6.60 28.50 86.50 3.90 0.74 5.00 28.50 87.50 3.90 0.94 5.90 28.50 88.50 3.90 0.82 5.25 28.50 89.50 4.25 0.72 4.60 28.50 90.50 4.00 0.99 6.10 28.50 91.50 4.35 0.99 6.25 28.50 92.50 3.95 0.94 6.00 28.50 93.50 4.05 1.04 6.45 28.50 94.50 3.90 0.94 6.15 28.50 95.50 4.05 0.80 5.55 29.50 84.50 4.00 0.84 5.30 29.50 85.50 3.95 1.25 7.40 29.50 86.50 3.95 0.90 5.80 29.50 87.50 3.95 1.10 6.70 29.50 88.50 3.95 0.98 6.05 29.50 89.50 4.30 0.89 5.40 29.50 90.50 4.05 1.15 6.90 29.50 91.50 4.40 1.15 7.05 29.50 92.50 4.00 1.10 6.80 29.50 93.50 4.10 1.20 7.25 29.50 94.50 3.95 1.10 6.95 29.50 95.50 4.10 0.97 6.35 4.3. Seismic Status Map of Bangladesh Table 2 and 3 show the seismicity parame- ters at different locations in Bangladesh, par- ticularly at 144 regions, the common area of vertical and horizontal pair rectangular regions. Using these parameters, 𝑀𝑐 , 𝑎- and 𝑏-values as shown in Table 3 contour maps along with the surface maps for Bangladesh polygon are being derived and shown in Figures 7-9. Figure 7. Spatially distributed magnitude of completeness in Bangladesh Figure 8. Spatially distributed 𝑎-value in Bangladesh Syed Mustafizur Rahman, et al./Vietnam Journal of Earth Sciences 40 (2018) 188 Figure 9. Spatially distributed 𝑏-value in Bangladesh Estimated magnitude completeness 𝑀𝑐 us- ing maximum curvature technique is of 3.9 and using MBS is 3.8 as shown in Figure 5(a- b). However, using two other methods as mentioned through the Eqs. 3-6, 𝑀𝑐 estima- tions were observed little high or greater than Syed Mustafizur Rahman, et al./Vietnam Journal of Earth Sciences 40 (2018) 189 4.0. It has observed from the earthquake cata- log that if higher values are being set to 𝑀𝑐, it is evident that the significant number of earthquake events were to be found beyond threshold level. Magnitudes of earthquake events less than 3.0 were not available in the catalog. Hence, in order to increase the partic- ipation of maximum number of events in the analysis, MAXC technique is appeared to be better one and used to estimate the 𝑀𝑐. Esti- mated 𝑀𝑐s are varying from 3.8-4.4, which are seemed to be well estimated. Recent con- tributions have shown the almost similar esti- mation of magnitude completeness of the ar- ea. Das et al. (2012) have estimated the value of 𝑀𝑐 for N-E India in the zone VII and VIII as 3.9 and 3.7 respectively. The zones VII and VIII are basically the northern and southern parts of Bangladesh. Kolathayar et al. (2012) have shown that 𝑀𝑐 is varying from 4.25-4.5 for Bangladesh re- gion along with India and adjoining area. While Khan et al. (2011) have approximated 𝑀𝑐 value around 3.0 for N-E India. From all above analyses including the present analyses it can be said that the magnitude completeness of the study area is close to 4.0. This work has suggested the same results of magnitude com- pleteness as obtained and shown in Figure 5(a-d) as varying from 3.8-4.2 using various techniques. It is also notable that Das et al. (2012) have shown 𝑀𝑐=3.9 in zone VII using four catalogs that include historical seismicity catalog as well and it was difficult to convert to a uniform magnitude scale, while the pre- sent analysis for the same area has shown al- most similar magnitude completeness, 𝑀𝑐=3.9 using MAXC technique for a single catalog. Kolathayar et al. (2012) have done the same job taking historical and instrumental data but estimated slightly higher magnitude of com- pleteness varying 4.25-4.5 for Bangladesh re- gion. It is essential to learn a reliable estima- tion of 𝑀𝑐 of an area though it can vary with time and space. There are limitations to inte- grate all historical and instrumental data from different sources for seismicity analysis. Ad- dition of time, spatial variation in seismicity and heterogeneity of earth could make the work too complicated. However, it has been said in most contributions that the assessment becomes more convincing while number of events can be increased in the catalog. Present research has emphasized on increasing num- ber of events only from one source in order to keep it bias free. Above contributions as men- tioned earlier have estimated seismicity for the whole area but not shown spatial variation of 𝑀𝑐 for Bangladesh preciously. This work has estimated over all magnitude of complete- ness 𝑀𝑐=3.9 of the study area and defined the spatial variation of magnitude of complete- ness varying from 3.90-4.45 for the whole Bangladesh. Figure 7 shows a reliable magni- tude completeness map, which can be em- ployed further whenever required, with spa- tially distributed 𝑀𝑐 along with 0.05 interval contour lines. It has observed that 𝑀𝑐 is low at the border line of N-W Bangladesh, and a line from Cox’s Bazaar to Sylhet through Hill tracts. Khan et al. (2011) have shown 𝑏-values in N-E India for zone I, where 𝑏-value is varying from 0.5-0.7. This zone (24.5-25.2oN and 90- 92oE) is partially common to the study area considered in this work. While the present analyses have presented the 𝑏-value varying from 0.77-1.15 in Table 3 and Figure 9, which is slightly higher. However, the variability in the seismic activity rate across the whole of India and adjoining areas has quantified in another contribution made by S. Kolathayar et al. (2012). This quantification has covered the total study area of the present analysis, where, 𝑏-value distributions in Bangladesh have shown as varying from 0.7-0.8. Present anal- yses have estimated almost the same 𝑏-value as varying 0.77-1.15 as shown Figure 9 except the region the central part of Barisal and Khulna divisions, where the 𝑏-value is ap- Vietnam Journal of Earth Sciences, 40(2), 178-192 190 peared to be greater than 1.1. There are few more contributions to assess the 𝑏-value of S- E Asia including Bangladesh and in most of the contributions, the 𝑏-value has shown vary- ing from 0.6-1.3 (Siddique, 2015 and Al- Hussaini, 2006). Hence, the 𝑏-value is seemed to be well estimated and quantified with an interval equal to 0.005 through spatially dis- tributed 𝑏-value as shown in Figure 9. It was so far not yet visualized before the spatially varied seismicity parameters in Bangladesh. Present work has developed spatially varying magnitude of completeness 𝑀𝑐, 𝑏-value and 𝑎-value in Bangladesh. The 𝑎-value varying form 4.95-7.20 has also presented and shown in Figure 8. 5. Conclusions The work may appear to be disadvanta- geous as used only the USGS earthquake cata- log for the analysis. But it is advantageous be- cause the events are being recorded, transmit- ted or processed uniformly by one organiza- tion and catalog is biased by processing or transmitting mechanisms only from one or- ganization. If other catalogs were to be inte- grated there would be different mechanisms to be used. Even different type of recording in- struments can cause further problems along with instrumental drifts. In such cases, the work may appear to be complicated to convert into a unique scale. The number of events in the catalog may be another issue but a total of 2606 events and magnitudes from 3.0-8.0 can be accounted as reasonable for seismicity analysis. Indeed integration of other catalogs and conversions into unique scale could pro- duce the better analysis. Earthquake events and caused damages are not seemed to be uniform to all directions from the source. Apparently, earthquake dis- tributions vary from place to place. It depends mostly upon the geologic condition of an area as sediments and geologic structure varies from one area to another. Findings of this work were basically primitive measures of seismicity for uniform long horizontal and vertical areas. Later the results are being inte- grated for common area, and to present the estimations in the form of maps. Spatially dis- tributed seismicity parameters as 𝑎-value, 𝑏- value and 𝑀𝑐 distributions of the country have been estimated and presented in maps. These maps might be valuable aid for engineering constructions and seismic hazard estimation. Estimated 𝑏-value obtained 𝑏>1.0 is being in- dicated a significant proportion of small earthquakes to the large one for the whole N- E Bangladesh, where central part of Barisal and Chittagong divisions including the port city Chittagong is being visualized as the highest state of 𝑏-value (𝑏=1.15) in the coun- try. It would be far better for 𝑎-value, 𝑏-value and 𝑀𝑐 distributions in the country if the cata- log could contain small or very small magni- tude earthquake events. Neither local magni- tude distributions recorded at seismic stations in Bangladesh were available, nor were the seismic networks found to be dense enough. On the contrary, small events are not the threats for damages. Within the limitations, this research work was intended to produce reliable 𝑎-value, 𝑏-value, and 𝑀𝑐 distributions. In this context, the work has contemplated and em- ployed different techniques to obtain the mag- nitude of completeness, 𝑀𝑐. An overall 𝑎=7.1, 𝑏=0.93 and 𝑀𝑐=3.9 values of seismicity pa- rameters estimation in S-E Asia and Bangla- desh indicate that the area is of a highly active seismic zone. Spatially distributed 𝑀𝑐 and 𝑏- value in Bangladesh presented in this work might be a useful aid for further development of seismological activities in the area. Acknowledgements Authors acknowledge Earthquake Hazards Program of USGS for making downloadable earthquake data available in their site. Syed Mustafizur Rahman, et al./Vietnam Journal of Earth Sciences 40 (2018) 191 References Abercrombie R.E., and Brune J.N., 1994. Evidence for a constant b-value above magnitude 0 in the southern San Andreas, San Jacinto, and San Miguel fault zones and at the Long Valley caldera, California. Geophys. Res. Lett., 21(15), 1647-1650. Aki K., 1965. Maximum likelihood estimate of b in the formula log N=a-b M and its confidence limits. Bull. Earthquake Res Inst., Tokyo Univ., 43, 237-239. Aki S., 1987. On nonparametric tests for symmetry. Ann. Inst. Statist. Math., 39, 457-472. Al-Hussaini T.M., 2006. Seismicity and Seismic Hazard Assessment in Bangladesh: Reference to Code Pro- visions. Meeting on Seismic Hazard in Asia ICTP, Trieste, Dec. 4-8. Amorese D., 2007. Applying a change-point detection method on frequency-magnitude distributions. Bull. Seismol. Soc. Am., 97(5), 1742-1749. Doi:10.1785/0120060181. Banglapedia, 2012. The National Encyclopedia of Bang- ladesh. Framework, retrieved on 31 Aug 2017. Cao A.M., and Gao S.S., 2002. Temporal variations of seismic b-values beneath northeastern Japan island arc. Geophys. Res. Lett., 29(9), 481-483. Doi:10.1029/2001GL013775. Das R., Wason H.R., and Sharma M.L., 2012. Temporal and spatial variations in the magnitude of complete- ness for homogenized moment magnitude catalogue for northeast India. J. Earth Syst. Sci., 121(1), 19-28. Felzer K.R., 2008. Simulated aftershock sequences for a M 7.8 earthquake on the southern San Andreas fault. Seismol. Res. Lett., 80, 21-25. GSB, 2018. Seismic Zone Map of Bangladesh. ortal.gov.bd/common_document/a6e75ad2_5acd_4f e3_911d_c9d25a7e349e/BD_Sciesmic- zonemap(NBC).pdf, retrieved on 31 March 2018. Gutenberg B., and Richter C.F., 1944. Frequency of earthquakes in California, Bull. Seismol. Soc. Am., 34, 184-188. Gutenberg B., and Richter C.F., 1956. Earthquake mag- nitude, intensity, energy and acceleration (second paper). Bull. Seismol Soc. Am., 46(2), 105-145. Hafiez H.E.A., 2015. Estimating the magnitude of com- pleteness for assessing the quality of earthquake cat- alogue of the ENSN. Egypt. Arab J. Geosci., 8(1), 9315-9323. Doi:10.1007/s12517-015-1929-x. Hunting Geology and Geophysics Ltd., (1981), Interpre- tation and Operations report on an aeromagnetic survey in Bangladesh, Borehamwood, Hertfordshire, England. Iwata T., 2008. Low detection capability of global earthquakes after the occurrence of large earth- quakes: investigation of the Harvard cmt catalogue. Geophys. J. Int., 174(3), 849-856. Doi:10.1111/j.1365-246X.2008.03864.x. Kagan Y.Y., 2002. Seismic moment distribution revisit- ed: I. statistical results. Geophys. J. Int., 148(3), 520- 541. Doi: 10.1046/j.1365-246x.2002.01594.x. Khan P.K., Ghosh M., Chakraborty P.P., and Mukherjee D., 2011. Seismic b-Value and the Assessment of Ambient Stress in Northeast India. Pure Appl. Ge- ophys., 168(10), 1693-1706. Doi:10.1007/s00024- 010-0194-x. Kolathayar S., Sitharam T.G., and Vipin K.S., 2012. Spatial variation of seismicity parameters across In- dia and adjoining areas. Nat Hazards, 60(3), 1365- 1379. Doi:10.1007/s11069-011-9898-1. Lomnitz-Adler J., and Lomnitz C., 1979. A modified form of the Gutenberg-Richter magnitude-frequency relation. Bull. Seism. Soc. Am., 69(4), 1209-1214. Marsan D., 2003. Triggering of seismicity at short time- scales following Californian earthquakes. J. Geophys. Res., 108, B5, 2266. Doi:10.1029/2002JB001946. Mignan A., 2011. Retrospective on the Accelerating Seismic Release (ASR) hypothesis: Controversy and new horizons. Tectonophysics, 505(1), 1-16. Doi:10.1016/j.tecto.2011.03.010. Mignan A., and Woessner J., 2012. Estimating the mag- nitude of completeness for earthquake catalogs, Community Online Resource for Statistical Seismic- ity Analysis. Swiss Seismological Service, ETH Zurich, 145p. Doi:10.5078/corssa-00180805. Avail- able at Naylor M., Orfanogiannaki, K., and Harte D., 2010. Ex- ploratory data analysis: magnitude, space, and time. Community Online Resource for Statistical Seismic- ity Analysis, 42p. Doi:10.5078/corssa-92330203. Available at Ogata Y., and Katsura K., 1993. Analysis of temporal and spatial heterogeneity of magnitude frequency Vietnam Journal of Earth Sciences, 40(2), 178-192 192 distribution inferred from earthquake catalogues. Geophys. J. Int., 113(3), 727-738. Doi:10.1111/j.1365-246X.1993.tb04663.x. Ogata Y., and Katsura K., 2006. Immediate and updated forecasting of aftershock hazard. Geophys. Res. Lett., 33, 10, L10305. Doi:10.1029/2006GL025888. Rashid H., 1991. Geography of Bangladesh, University Press Ltd, Bangladesh; 2nd edition, 545p. Reimann K.U., 1993. Geology of Bangladesh. Ger- bruder Bornt Ramerg, Berlin, Germany. 160p. Siddique S., 2015. Gutenberg-Richter recurrence law to seismicity analysis of Bangladesh. IABSE-JSCE Joint Conference on Advances in Bridge Engineer- ing-III, August 21-22, Dhaka, Bangladesh. Shi Y., and Bolt B.A., 1982. The standard error of the magnitude-frequency b-value. Bull. Seismol. Soc. Am., 72(5), 1667-1687. USGS, 2012. Earthquake Hazards Program. https://earthquake.usgs.gov/earthquakes/search/, USA, retrieved on 20 April 2017. Utsu T., 1999. Representation and analysis of the earth- quake size distribution: a historical review and some new approaches. Pure Appl. Geophys., 155(2), 509-535. Wiemer S., and Wyss M., 2000. Minimum magnitude of complete reporting in earthquake catalogs: examples from Alaska, the western United States, and Japan. Bull. Seismol. Soc. Am., 90, 859-869. Doi:10.1785/0119990114. Woessner J., and Wiemer S., 2005. Assessing the quality of earthquake catalogues: Estimating the magnitude of completeness and its uncertainty. Bull. Seismol. Soc. Am., 95(2), 684-698. Doi:10.1785/012040007. Wyss M., Hasegawa A., Wiemer S., and Umino N., 1999. Quantitative mapping of precursory seismic quiescence before the 1989, M7.1 off-Sanriku earth- quake, Japan. Annali Di Geoflsica, 42(5), 851-869. Zuniga F.R., and Wyss M., 1995. Inadvertent changes in magnitude reported in earthquake catalogs: Their evaluation through b-value estimates. Bull. Seismol. Soc. Am., 85, 1858-1866. .Zuniga F.R., and Wiemer S., 1999. Seismicity patterns: Are they always related to natural causes? Pure Appl. Geophys., 155(2), 713-726.