See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/245189640

A wind map of Bangladesh

Article  in  Renewable Energy · April 2004

DOI: 10.1016/j.renene.2003.10.002

CITATIONS READS

63 10,615

3 authors, including:

Jahangir Khan Tariq Iqbal

BC Hydro Memorial University of Newfoundland
49 PUBLICATIONS   2,034 CITATIONS    340 PUBLICATIONS   3,835 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

NESTNET View project

Solar Energy Storage System View project

All content following this page was uploaded by Jahangir Khan on 19 October 2020.

The user has requested enhancement of the downloaded file.

 Renewable Energy 29 (2004) 643–660
www.elsevier.com/locate/renene

A wind map of Bangladesh

M.J. Khan
, M.T. Iqbal, S. Mahboob
Faculty of Engineering, MUN, St. John’s, NL, Canada A1B3X5
Received 8 September 2003; accepted 3 October 2003

Abstract

Utilization of wind energy in Bangladesh has been slow mainly due to lack of quality
wind data. Recent measurements in some places have shown significant wind energy poten-
tials in Bangladesh. In this paper, a wind map is presented which incorporates several micro-
scale features, such as terrain roughness, elevation etc. with a mesoscale model. Several
meso-maps were obtained from global databases and a suitable model was chosen and modi-
fied for a 30-m elevation. Ground data from various sources were collected and modified for
height and land condition adjustments based on local knowledge and GIS information. It
was found that, the generated wind map and the modified ground data resemble. Annual
average wind speed at 30 m height along the coastal belt is above 5 m/s. Wind speed in
northeastern parts is above 4.5 m/s while inland wind speed is around 3.5 m/s for most part
of Bangladesh. Small-scale wind turbines could be installed and tested in locations such as
St. Martins Island, Cox’s Bazar, Patenga, Bhola, Barguna, Dinajpur, Thakurgaon and
Panchagar.
# 2003 Elsevier Ltd. All rights reserved.

Keywords: Wind map; Wind energy in Bangladesh; Global database; Ground measurement; Meso and
microscale wind modeling

1. Introduction

Bangladesh is a south Asian country located between 20 300 N–26 380 N and
v v

88 040 E–92 440 E, bordered by the Bay of Bengal at the south, Myanmar at the
v v

south-east and India at the west, north and north-east. Its population is about 133
million, with a growth rate of 1.59% and population density of 800 persons per sq.
km (July 2002). About 80% of the country’s population depends on traditional

Corresponding author. Tel.: +1-709-737-8809.
E-mail address: mjakhan@engr.mun.ca (M.J. Khan).

0960-1481/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.renene.2003.10.002
644 M.J. Khan et al. / Renewable Energy 29 (2004) 643–660

methods of agriculture for livelihood. Only 44.3% of the people are literate and
35.6% (1995–1996) live below the poverty level [1,2].
About 73% of the total energy consumption (1999) of Bangladesh comes from
traditional biomass fuels (wood 65%, crop residue 22.1%, animal dung 7.8%) [3].
Bangladesh Power Development Board (BPDB) and several Independent Power
Producers (IPP) are responsible for generating electricity for the national grid
(73.08% and 26.92%, respectively). Total installed capacity of BPDP in 1999 was
3603 MW (fossil fuel: 92.45%, hydro: 7.55%) [2,4]. The transmission loss in some
cases was as high as 25%.
Only about 18% of the population (25% in urban areas and 10% in villages) is
connected to the grid and the demand for electricity is increasing by 300 MW per
year [2,3]. The need for a huge investment implies that grid-connected electricity
supply for the rural areas is not going to happen in the near future and 80% of the
population would remain off-grid for the days to come. Considering these facts,
several organizations have been actively working on renewable energy alternatives
for rural areas of Bangladesh [4–8]. Successful implementation of solar-
photovoltaic units has been achieved and stand-alone wind driven battery-charging
station and wind–solar hybrid energy sources are also being considered [2,7].
Assessment of wind energy resources and installation of wind energy conversion
systems (WECS) in Bangladesh has long been hindered due to lack of reliable wind
speed data. There is no reported wind map of Bangladesh, that could be relied
upon and used for wind energy assessment [5,6]. Weather data collected by the
meteorological offices around the country are considered as low quality data (char-
acterized by low tower height, obstacles around the tower and poor data recording
mechanism). Some organizations have recently started wind data measurement
activities in various parts of the country [5,6]. A long-term surface measurement
(3–5 year average) is vital for making a quality wind resource assessment. Before
this is done, a wind map proposed in this paper, which is developed with alterna-
tive means, could prove to be useful.
It is understood that, wind speed in Bangladesh is not as high as many other
countries in the world (Denmark, England, USA or Canada). However, develop-
ment of a reliable wind map is of paramount importance for three reasons. Firstly,
it would help identify suitable sites with higher wind energy concentration where
the demonstration scale WECS could be installed. Secondly, a wind map would
serve the purpose of indicating candidate sites where more measurements could be
carried out. Last but not the least, research for low-speed wind turbines is well
underway and such systems might find applications in low-wind areas inside the
country [9]. A wind map presented in this paper, in spite of its dependence on some
crucial assumptions, would be a useful starting point for any interested person or
group.
In this paper, a wind map of Bangladesh is developed by introducing several
microscale features to a mesoscale wind model obtained from a number of global
databases. A general description of the geography of Bangladesh is followed by
two sets of wind maps collected from two different global databases. Surface data
was collected from various sources and modified for height and terrain conditions.
 M.J. Khan et al. / Renewable Energy 29 (2004) 643–660 645

The most suitable wind map (generated by global database) was also modified for
similar adjustments. Finally, a comparison of modified surface data and the pro-
posed wind map was carried out to demonstrate its range of validity.

2. Wind resources and geography of Bangladesh

Total area of Bangladesh is around 144,000 sq. km (land: 133,910 sq. km, water
bodies: 10,090 sq. km) with a long shoreline along the Bay of Bengal (580 km).
Most of the country lies within the delta formed by the Ganges and Brahmaputra
rivers, originating in the Himalayas mountain range. The land is characterized as
flat and fertile with alluvial soil deposited by the floodwaters, recurring almost
every year.
The southeastern part of the country consists of hilly terrains (less than one-
tenth of the nation’s territory) attaining elevations of as high as 100 m. Small and
sparsely distributed eroded hills (Madhupur Tract, north central regions and The
Barind, straddling in the northwestern part) attain elevations of about 60 m
(Fig. 1). The Chittagong Hill Tracts District (southeastern regions) is covered with

Fig. 1. Elevation map of Bangladesh.

646 M.J. Khan et al. / Renewable Energy 29 (2004) 643–660

broad leaf evergreen trees. The Sundarbans, located at the southwestern part of the
country is a mangrove forest with rich natural diversity. The plain areas within the
countryside contain several species of deciduous trees and about 70% of the land is
used for cultivation of various crops (rice, jute, cotton, sugarcane etc.) (Fig. 2). The
land is crisscrossed by hundreds of rivers and flooding is a regular phenomenon in
Bangladesh.
The climate of Bangladesh could be described with two major seasonal varia-
tions: mild winter and hot-humid summer. The dry-winter season lasts for about 5
months (November–March) and gentle north-easterly land breeze blows over the
terrain during this period. The summer spans from April to October with frequent
small-scale storms occurring almost every day. During the months of June–
October, monsoon rains cause a large quantity of precipitation. Strong south-
westerly sea breeze blows over the land during the summer season.
The population distribution within the country is relatively even. The density is
exceptionally high in the four major cities in the country (Dhaka (capital city),
Chittagong, Rajshahi, and Khulna). The Chittagong Hill Tracts District and the

Fig. 2. Terrain profile of Bangladesh.

 M.J. Khan et al. / Renewable Energy 29 (2004) 643–660 647

Sundarbans are almost entirely uninhabited. The countryside contains evenly loca-
ted houses surrounded by trees and crop fields.
The Tropic of Cancer passes through the middle of Bangladesh and the terrain is
subjected to the northeast trade winds (primary origin). During the summer, wind
blows from the southwest direction bringing moisture from the Bay of Bengal and
causing significant monsoon effect throughout the country (secondary origin).
Tropical cyclones are also common at the beginning of the summer season. All the
studies show that, wind speed is generally lower in the winter and higher in the
summer. Wind also exhibits a diurnal cycle, generally having a peak at noon and a
dip at evening (tertiary origin).
Wind has three major classes of origin: primary, secondary and tertiary [11]. The
primary or global origin of wind resource is accounted for pressure differences
across the earth’s surface, gravitational forces, inertia of the air, the earth’s
rotation and friction with the surface. In a simpler form, four atmospheric forces:
pressure force, Coriolis force (due to earth rotation), inertial force (due to global
scale circular motion) and frictional forces (with the earth’s surface) determine glo-
bal perspective of wind motion. Secondary sources of wind include hurricanes,
monsoon circulation, and cyclones. Thirdly, diurnal variations, thunderstorms, tor-
nadoes etc. determine short-term, small-scale wind variations.
A worldwide wind atlas generated with NOAA/CIRES–Climate Diagnostic
Center’s (CDC) search engine (NCEP reanalysis product), reveals that the surface
v v
mean wind speed over Bangladesh (20 N, 90 E) is within the range of 4–6 m/s
(Fig. 3) [14]. Surface measurements have also established that, wind speed in the

Fig. 3. Overview of worldwide wind speed (m/s) variations at surface level. (CDC Derived NCEP Rea-
nalysis Products (Surface Level), averaged over Jan 1948 to Jun 2003) [14].
648 M.J. Khan et al. / Renewable Energy 29 (2004) 643–660

coastal areas is generally higher than in the inland, except the northeastern part.
The western part of the country has relatively higher winds than the east [5,6].

3. Wind map generation using global databases

One of the very first steps towards harnessing energy from the wind is to make
extensive assessment of wind energy potential and cost analysis for a site of inter-
est. A suitable location could then be identified from several candidate sites. All of
these analyses need a reliable and accurate wind map. Long record and dense
measurement network based information incorporated with satellite data is gener-
ally used for such purposes. However, when data is inadequate or inaccurate sev-
eral other alternatives could also be explored [12].
A microscale wind map considers the effects of near-by obstacle, terrain rough-
ness, orography and thermal flow with a scale of 1–10 km, which is the ultimate
goal of a site-specific wind resource analysis [11–13]. On the other hand, a mesos-
cale map has a grid size of the order of 10–100 km. Large-scale climatological data
is regionalized in order to develop a meso-map, where, local effects defined above
are not considered. Various software tools, such as WAsP, KAMM could be used
to model high-resolution wind resource maps [12,13].
Development of a wind map requires information from various sources: local
knowledge, regional surface measurements, global climatological databases, com-
putational fluid dynamics (CFD) analysis, and correlation analysis [12]. An over-
estimation may result in an economic disaster and an underestimation might leave
a good potential unexplored.
Combination of long, dense network of surface measurements with alternative
methods is the de facto standard for wind atlas modeling. For the region under
study, there is no such reliable record. Therefore alternative methods need to be
utilized until such data could be collected. Global databases containing long-term
(5–25 years) information could be used with relative ease. These databases are gen-
erally accessible to everyone, climatologically stable and could be used for almost
any place of the world. On the other hand, ground data collected for a short per-
iod could still prove to be very useful when properly collected and modified by
means of measure-correlate-predict (MCP) method [12]. In this paper these two
aspects (global database and surface measurement correlation) are considered for
generating a wind map that essentially incorporates microscale features embedded
in a mesoscale map.
3.1. Wind map using NASA–SSE database

Surface meteorology and Solar Energy (SSE) data set of the National Aeronaut-
ics and Space Administration’s (NASA) Earth Science Enterprise Program main-
tains a renewable energy resource web site in collaboration with the CAMNET
Energy Technology Center [15,16]. The data output of this site could be used in
conjunction with the RETScreen1 International Renewable Energy Project Analy-
sis Software [16]. Unlike surface measurements, this data set is a consistent 10-year
 M.J. Khan et al. / Renewable Energy 29 (2004) 643–660 649
v v
global average on a 1 by 1 (about 100  100 km) grid. The SSE data,which is
essentially an average over the entire area of the cell may not represent a particular
site within the grid. However, this database is an excellent and easy-to use source
which could be used for any preliminary study for renewable energy resource esti-
mation.
v
Two sets of wind speed data for 10 and 50 m heights were gathered for 20.5 N–
v v v
26.5 N and 87.5 E–93.5 E. This data after rearranging in a 7  7 matrix (49 data
v v
points, each for one 1  1 grid), were used in Matlab [17] to create contour plots
shown in Figs. 4 and 5. As expected, the coastal and northern regions appear as
high-wind areas when compared with the main land.
3.2. Wind map using NOAA/CIRES–CDC database

The Climate Diagnostic Center (CDC) database maintained jointly by the

National Oceanic and Atmospheric Administration (NOAA) and the Cooperative
Institution for Research and Environmental Sciences (CIRES) is an excellent source

Fig. 4. Wind map of Bangladesh using NASA–SSE-data set at 10 m height (m/s).

650 M.J. Khan et al. / Renewable Energy 29 (2004) 643–660

Fig. 5. Wind map of Bangladesh using NASA–SSE-data set at 50 m height (m/s).

of weather information. The search mechanism in CDC allows one to collect a wide
range of climate data from many different sources.
Wind speed contour plots for various elevation levels (surface, 1000, 850 mbar
etc.) were gathered from various sources (NCEP/DOE AMIP-II, GFDL R30S14,
CDC Derived NCEP Reanalysis) with CDC’s search engine. Since Bangladesh is
mostly flat with 5–10 m average elevation from the sea level for most part (Fig. 2),
the surface level plots deemed suitable. Two plots with NCEP reanalysis and
GFDL database were chosen for further studies (Figs. 6 and 7).
The CDC derived NCEP reanalysis (Fig. 6), shows close match with the NASA–
SSE maps (Figs. 4 and 5), in terms of general pattern of high-wind and low-wind
regions. The plot of GFDL data deviates greatly from the others. The NASA–SSE
plots (even at 50 m height), shows lower wind speeds when compared with the
ground data given in Table 1. This leaves, the CDC derived NCEP reanalysis
based wind map (Fig. 6) as the most suitable one for further investigation (modifi-
cation for elevation, terrain conditions and validation).
 M.J. Khan et al. / Renewable Energy 29 (2004) 643–660 651

Fig. 6. Wind map of Bangladesh at surface level (m/s) with CDC Derived NCEP Reanalysis Products
(Surface Level), averaged over Jan 1948 to Jun 2003 (min: 1.87, max: 5.52).

4. Collection, analysis and modification of surface data

Systematically collected, quality wind data containing information on wind

speed, direction, turbulence measure etc. is almost entirely absent for most parts of
the country. Bangladesh Meteorological Department has been collecting wind
speed data for a long period (1961 onward) from many weather stations spread
throughout the country. But these are usually meant for weather forecasting and
are unsuitable for wind energy assessment purposes owing to three major reasons.
Firstly, measurement tower height is too low (5–10 m), whereas wind turbines gen-
erally have higher hub heights (10–30 m). Secondly, obstacles such as, near-by
buildings, houses and plantation, surround many weather stations. Thirdly, data
recording mechanisms in many stations were not automated and are subject to
erroneous processing. However, since these are averaged over a very long period
(more than 30 years), use of these data with proper adjustments for elevation and
652 M.J. Khan et al. / Renewable Energy 29 (2004) 643–660

Fig. 7. Wind map of Bangladesh at surface level (m/s) with CDC Derived GFDL Data for Surface
Wind Speed, averaged from Jan to Dec (min: 3.29, max: 7.61).

terrain conditions is still worth consideration. Whatever data were available from
different sources is given in Table 1 [5,6,10].
Recently, many different local and international organizations, such as, Bangla-
desh University of Engineering and Technology (BUET), Bangladesh Center for
Advanced studies (BCAS), Local Government Engineering Department (LGED),
Bangladesh Atomic Energy Commission (BAEC), Rural Electrification Board
(REB), Institute of Fuel Research Development (IFRD), Bangladesh Council of
Scientific and Industrial Research (BCSIR), Energy Technical Support Unit
(ETSU, Harwell, UK), GTZ (Germany) have started wind resource assessment
projects such as Wind Energy Study (WEST), Solar and Wind Energy Resource
Assessment (SWERA), Technical Expertise for Renewable Energy Application
(TERNA), Wind Energy Resource Mapping (WERM) program etc. Most of these
projects are ongoing and final results are not published. Available wind speed data
and measurement tower heights for different projects are given in Table 1 [5,6].
Table 1
Surface data of annual average wind speed (m/s) of different places in Bangladesh
No. Place Latitude–Longitude Meteorological office WERM TERNAb IFIDb
(1961–1993) (1996–1997) (1995–1997) (2001–2002)
v v
1 Barisal 22 400 N–90 280 E 2.66a – – –
v v
2 Bhola Island 22 400 N–90 400 E 2.44 (at 7 m) – – –
v 0 v
3 Char Fassion 22 12 N–90 420 E – 3.28 (at 10 m) 4.07 – –
(at 25 m)
v v
4 Comilla 23 280 N–91 100 E 2.21 (at 6 m) 2.78a – – –
v v
5 Cox’s Bazar 21 280 N–91 590 E 2.42 (at 10 m) 3.81a 3.34 (at 25 m) – –
v v
6 Dinajpur 25 330 N–88 450 E 2.83a – – –
v 0 v
7 Hatia Island 22 23 N–91 230 E 2.08 (at 6 m) 3.74a – – –
v v
8 Khepupara 21 500 N–89 500 E 2.36 (at 10 m) 4.24a – – –
v 0 v
9 Khulna 22 48 N–89 300 E 2.89a – – –
v v
10 Kuakata 21 550 N–90 080 E – 4.54 (at 25 m) – –
v 0 v
11 Kutubdia 21 48 N–91 520 E 2.16 (at 5 m) 2.32a 4.18 (at 25 m) – –
v v
12 Mongla 22 220 N–89 470 E 2.20a – – –
v 0 v
13 Noakhali 22 52 N–91 080 E – 2.96 (at 25 m) – –
v v
14 Patenga 22 160 N–91 450 E 2.45 (at 5 m) 7.48a 3.85 (at 25 m) 5.76 –
v 0 v
15 Rangamati 22 36 N–92 110 E 2.15a – – –
v v
16 Sandip 22 280 N–91 300 E 2.76a – – –
v 0 v
17 St. Martin Island 20 42 N–92 170 E – – – 4.79
v v
M.J. Khan et al. / Renewable Energy 29 (2004) 643–660

18 Sylhet 24 500 N–91 520 E 2.38a – – –

v 0 v
19 Teknaf 20 50 N–92 150 E 2.16 (at 5 m) 3.17a 2.96 (at 25 m) – –
a
Collected from Mahfuz et al. [10]. Apparently these data are derived from meteorology department data and placed in that category. Tower height is
assumed to be 7 m.
b
Since recent data, tower height is assumed to be 25 m.
653
654 M.J. Khan et al. / Renewable Energy 29 (2004) 643–660

As seen in this Table 1, recent measurements done with proper elevation show
higher wind speeds when compared with the meteorological office data for the
same site (Cox’s Bazar, Kutubdia, Patenga, Teknaf). Indications are, wind speed in
many parts of the country is higher than previously thought [5,6,10,18]. Therefore,
it could be concluded that, the meteorological office data when scaled up for elev-
ation and terrain adjustments, can still be used for wind resource assessment.
Wind data measured in a place surrounded by near-by obstacles such as hills,
tall tress, houses etc. are characterized by lower magnitude, higher turbulence and
lower power concentration. Similar affects are observed when wind measurement is
done near-surface locations (Fig. 8a).
The variation of wind speed with elevation could be described with two math-
ematical models: logarithmic law and power law [11]. The logarithmic law is a
combination of theoretical and empirical research originated from boundary layer
flow in fluid mechanics. Although this law is meant for vertical wind speed vari-
ation modeling in a particular terrain condition, here it is used for incorporating
both vertical and lateral variations providing means for scaling a data from one
elevation and terrain class to another elevation and land condition. The log law is
given by the following equation:
  
Z Zr
Uz ¼ ln ln Ur ¼ cfl  Ur ð1Þ
Zo Zor

where, Z and Zo are modified elevation and roughness length of new terrain,
respectively. Zr and Zor are similar parameters for reference (measurement) site.
The log law correlation/correction factor (cfl) is therefore a term that converts
wind speed (Ur) from reference area to a new value (Uz) that would suit a near-by
location. To identify the roughness length of a particular site, a terrain profile
index versus roughness length correlation (Fig. 8b) is done using the rule of thumb
information found in [11,19].

Fig. 8. (a) Downstream wind speed variation due to obstacles. (b) Terrain classification in terms of a ter-
rain profile index.
 M.J. Khan et al. / Renewable Energy 29 (2004) 643–660 655

The power law is a similar model that represents vertical wind speed profile. It is
given by [11]:
 a
Z
Uz ¼ Ur ¼ cfp  Ur ð2Þ
Zr

Here, cfp is the power law correlation/correction factor. The power law exponent
a, varies with parameters such as elevation, time of day, season, terrain, tempera-
ture, wind etc. A model correlating this exponent to surface roughness is given
by [11]:
a ¼ 0:096log10 Zo þ 0:016ðlog10 Zo Þ2 þ 0:24 ð3Þ

The logarithmic law (1) is used for converting reported wind data (specially the
meteorological office data) with specified measurement tower height to a new elev-
ation of 30 m. The measurement (reference) area terrain index is taken from 5.25
to 7, which is attributed to locations with near-by obstacles. Since most of the land
is covered with vegetation (crops, few trees, small village homes and water bodies)
a terrain profile index from 4 to 6 is chosen for the near-by areas. Since this
method corrects the available data (to some extent) and also correlates other mea-
surements (Table 2), the term correction and correlation are used interchangeably.
To test the validity of this approach wind data from the meteorological office is
scaled up for elevation correction and compared with WERM project data for four
separate locations. These are the only sites, where multiple reliable data could be
gathered, corrected and compared (Table 2).
To determine the terrain profile index, local knowledge and Geographic Infor-
mation System (GIS) database of Environmental Systems Research Institute
(ESRI) [20] was used, as shown in Fig. 9. Calculation of the roughness length from
the terrain index is done by a curve fitting equation that represents Fig. 8b.

Table 2
Terrain profile analysis for four test locations
Place Meteorological WERM Estimated Error (%) Estimation method
office data wind speed wind speed
data (m/s) Uz (m/s) at
Wind Tower
at 25 m Z ¼ 25 m
speed, Ur height Zr
(m/s) (m)
Cox’s Bazar 2.42 10 3.34 3.22 3.61 Terrain profile index: 6.5
Zo ¼ Zor ¼ 0:6245 m
Kutubdia 2.16 5 4.18 3.69 11.77 Terrain profile index: 6.25
Zo ¼ Zor ¼ 0:5139 m
Patenga 2.45 5 3.85 4.04 5.12 Terrain profile index: 6
Zo ¼ Zor ¼ 0:416 m
Teknaf 2.16 5 2.96 3.33 12.62 Terrain profile index:
5.5Zo ¼ Zor ¼ 0:258 m
656 M.J. Khan et al. / Renewable Energy 29 (2004) 643–660

Fig. 9. Analysis of sites of interest for determining terrain profile index with GIS database of ESRI [20].

As seen in Table 2, the range of error of this estimation method for these loca-
tions is 4 to +20%. The wind data from different sources were compared, ana-
lyzed and a new data set is prepared for making corrections with the logarithmic
law. Where multiple sources of information were available, recent reports (WEST,
TERNA, IFID project) were preferred over meteorological office measurements.
However, the source of data reported in [10] could not be verified and were not
taken into consideration where alternate information was available.

5. The proposed wind map and its validity

The CDC derived NCEP reanalysis based wind model of Bangladesh (Fig. 6) is
considered as a more suitable map when compared with other global data based
models. However, it also needs some height and terrain modifications to take into
account for the land profile and elevations attributed to Bangladesh (Fig. 2). The
plot of CDC derived NCEP reanalysis was gridded into a 13  13 matrix (with 169
v v
data points representing 0:5  0:5 resolution). Then the power law (Eq. (2)) was
used to modify the model to elevate it to 30 m. Necessary height adjustments were
introduced to the gridded data for the hilly districts in the southeast and eroded
highlands in the north. The value of the power law exponent a, for most parts of
the country was found to be 0.15 when Eq. (3) is used with land profile index of
4.5 (Zo ¼ 0:07 m, equivalent to cropland with occasional trees).
The modified CDC derived NCEP reanalysis data was plotted in Matlab. Fig. 10
shows the developed wind map at 30 m elevation, which effectively incorporates the
surface roughness microscale conditions typical to Bangladesh. This model highly
 M.J. Khan et al. / Renewable Energy 29 (2004) 643–660 657

Fig. 10. Wind map of Bangladesh at 30-m elevation.

correlates with the corrected surface data of Table 3, as seen in Fig. 11. Deviations
in the northeastern part could not be analyzed further as there is only one
measurement available in that region. However, most of the values in the coastal
region show a close match with the wind speed outlined in Fig. 10. The widely
accepted notion that, wind speed in Bangladesh is higher than previously thought,
is also reflected in this model. As predicted, the coastal and northeastern regions
are found to be higher-wind areas compared to the mainland.

6. Conclusion

In this paper a wind map of Bangladesh incorporating several microscale fea-

tures, such as terrain roughness, elevation etc. is presented. A mesoscale map,
obtained from several global climatological databases is adjusted for a 30-m elev-
ation, and terrain condition, which would suit the land profile of Bangladesh. The
available data from surface measurements were investigated, corrected, correlated
and compared with this model. In the absence of a long, dense network of ground
measurement based wind data, several logical assumptions were made during the
 658

Table 3
Corrected wind speed (m/s) data set for 30 m elevation and typical countryside terrain of Bangladesh using the logarithmic law
Location Ref. wind Tower height Reference Reference area Near-by area Near-by area Correction Estimated wind
speed, Zr (m) area terrain roughness length terrain profile roughness factor, cfl speed at
Ur (m/s) profile index Zor (m) index length Zo (m) Z ¼ 30 m
Barisal 2.66 7a 6.25 0.5139 5.5 0.2586 1.8203 4.8419
Bhola Island 2.44 7 6.25 0.5139 5.5 0.2586 1.8203 4.4414
Char Fassion 4.07 25 6 0.4166 5.25 0.1963 1.2284 4.9994
Comilla 2.21 6 6.25 0.5139 6 0.2586 1.9349 4.2751
Cox’s Bazar 3.34 25 6.5 0.6245 4.65 0.0862 1.5862 5.2977
Dinajpur 2.83 7a 6 0.4166 5.85 0.3642 1.5634 4.4243
Hatia Island 2.08 6 5.5 0.2586 4.15 0.0307 2.1897 4.5546
Khepupara 2.36 10 5.75 0.3318 4.15 0.0307 2.0215 4.7708
Khulna 2.89 7a 5.5 0.2586 5.15 0.1742 1.5610 4.5112
Kuakata 4.54 25 5.5 0.2586 5 0.1440 1.1679 5.3024
Kutubdia 4.18 25 6.25 0.5139 5.25 0.1963 1.2947 5.4120
Mongla 2.20 7a 6 0.4166 4.25 0.0359 2.3507 5.1716
Noakhali 2.96 25 6.25 0.5139 4.75 0.1010 1.4658 4.3387
Patenga 3.85 25 6 0.4166 5 0.1440 1.3040 5.0203
Rangamati 2.15 7a 7 0.8888 6.15 0.4734 2.0103 4.3222
Sandip 2.76 7a 6 0.4166 5.15 0.1742 1.8248 5.0365
M.J. Khan et al. / Renewable Energy 29 (2004) 643–660

St. Martin Island 4.79 25a 5.25 0.1963 4.5 0.0664 1.2611 6.0408
Sylhet 2.38 7a 5.25 0.1963 5 0.1440 1.4938 3.5552
Teknaf 2.96 25 5.5 0.2586 4 0.0194 1.6064 4.7550
a
Seven meters for older measurements and 25 m for more recent measurements.
 M.J. Khan et al. / Renewable Energy 29 (2004) 643–660 659

Fig. 11. Comparison with corrected surface data (Table 3).

development of this model. Further ground measurements should be carried out

before reaching decisive conclusions. The proposed wind map resembles closely
with the correlated surface measurement data. As indicated by some recent mea-
surements, the wind speeds in many locations were found to be higher than older
meteorological department assessments. The coastal areas and the northeastern
regions contain areas with high wind, and small-scale wind energy conversion sys-
tem could be built and tested in those locations. It indicates that the coastal belt of
Bangladesh has an annual average wind speed above 5 m/s whereas the wind speed
in the northeastern districts is around 4.5 m/s. More measurements and demon-
stration type wind turbine installations could be made in areas such as St. Martin
Islands, Cox’s Bazar, Kutubdia, Patenga, Bhola, Sandip, Barguna, Kuakata,
Dinajpur, Thakurgaon and Panchagar.

References
[1] Official web page of the Government of the People’s Republic of Bangladesh. Available from URL:
http://www.bangladeshgov.org/.
 660 M.J. Khan et al. / Renewable Energy 29 (2004) 643–660

[2] The Bangladesh travel guide. Available from URL: http://www.discoverybangladesh.com/.

[3] Ibrahim M, Anisuzzaman M, Kumar S, Bhattacharya SC. Demonstration of PV micro-utility sys-
tem for rural electrification. Solar Energy 2002;72(6):521–30.
[4] Bangladesh Power Development Board (BPDP), Bangladesh. Available from URL: http://
www.bd-pdb.org/.
[5] Sustainable Rural Energy (SRE) program, Local Government Engineering Department (LGED),
Government of Bangladesh. Available from URL: http://www.lged.org/sre/.
[6] Shakti, Energy & Environment Related Information Source of Bangladesh. Available from URL:
http://shakti.hypermart.net/index.html/.
[7] Grameen Shakti. Available from URL: http://www.grameen-info.org/grameen/gshakti/.
[8] Rural Electrification Board (REB), Bangladesh. Available from URL: http://www.bangladeshgo-
v.org/reb/solar_program.htm/.
[9] 2002 research review. National Renewable Energy Laboratory (NREL), 1617 Cole Blvd, Golden,
CO, July 2002. Available from URL: http://www.nrel.gov/.
[10] Khan SI, Mahfuz MU, Aziz T, Zobair NM. Prospect of hybrid wind system in Bangladesh. Second
International Conference on Electrical and Computer Engineering ICECE 2002, 26–28 December,
Dhaka, Bangladesh. 2002.
[11] Manwell JF, McGowan JG, Rogers AL. Wind energy explained. West Sussex, England: John Wiley
& Sons Ltd; 2002 ISBN 0-471-49972-2.
[12] Landberg L, Myllerup L, Rathmann O, Petersen EL, Jorgensen BH, Badger J, et al. Wind resource
estimation—an overview. Wind Energy 2003;6(3):261–71.
[13] Petersen EL, Mortensen NG, Landberg L, Hojstrup J, Frank HP. Wind power meteorology. Wind
Energy Department, Risø National Laboratory, Frederiksborgvej 399, P.O. 49, DK-4000 Roskilde,
December 1997. Available from URL: http://www.risoe.dk/.
[14] NOAA-CIRES Climate Diagnostics Center: 325 Broadway, R/CDC1, Boulder, CO 80305-3328.
Available from URL: http://www.cdc.noaa.gov/.
[15] NASA, SSE database. Available from URL: http://eosweb.larc.nasa.gov/sse/RETScreen/.
[16] CAMNET Energy Technology Center, RETScreen1 International Renewable Energy Project
Analysis Software. Available from URL: http://www.retscreen.net/.
[17] The MathWorks, Inc., 3 Apple Hill Drive, Natick, MA 01760-2098, US. Available from URL:
http://www.mathworks.com/.
[18] Rahman MF. Prospects of wind energy in Bangladesh. Renew Energy 1996;9(1–4):806–9 September.
[19] Wind Energy Resource Atlas of Southeast Asia September 2001. The World Bank Asia Alternative
Energy Program (ASTAE), 1818 H Street, NW, Washington, DC 20433, USA. Avaialble from
URL: http://www.worldbank.org/ASTAE/.
[20] Environmental Systems Research Institute (ESRI), 380 New York Street, Redlands, CA 92373-
8100, USA. Available from URL: http://www.esri.com/.

View publication stats