ARTICLE IN PRESS

Renewable Energy 31 (2006) 2329–2341

www.elsevier.com/locate/renene

Data Bank

A pre-feasibility study of wind resources in

Kutubdia Island, Bangladesh
Shafiuzzaman Khan Khadem, Muhtasham Hussain
Renewable Energy Research Centre, Energy Park, Faculty of Science, University of Dhaka, Dhaka, Bangladesh
Received 10 May 2005; accepted 17 February 2006
Available online 17 April 2006

Abstract

Kutubdia is one of the coastal islands in Bangladesh. The wind speed data at the wind monitoring
station of Bangladesh Meteorological Department (BMD), Kutubdia, located in a built-up area,
appears to be low, but at another location, near the seashore, Bangladesh Centre for Advanced
Studies (BCAS) finds that the wind energy availability should be reasonably high. Considering the
surface roughness, obstacle condition and terrain information of the island, a micro-scale prediction
has been done using (100
100 m2) grid cells in the Wind Atlas Analysis and Application Program
(WAsP) to develop monthly and annual wind atlas and also a wind resource map which shows that at
50 m height, the annual wind speed over Kutubdia coast varies from 5.1 to 5.8 m/s. Southern and
Eastern sides of Kutubdia appear to be promising for wind electricity generation using large turbines
where wind power density at a height of 50 m or higher is found to be above 200 W/m2 annually over
the year September 1996 to August 1997. It is found that at 30 m height, the coast side of Kutubdia
should be sustainable for small turbines.
r 2006 Published by Elsevier Ltd.

Keywords: Wind resource assessment; WAsP analysis; Kutubdia Island; Bangladesh

1. Introduction

To avoid burning of fossil fuels which pollutes the atmosphere and to go for non-
exhaustible energy supplies, the demand for utilizing wind energy, solar energy and other

Corresponding author. Tel.: +1 880 2 9677125.

E-mail address: skkhadem@yahoo.com (S.K. Khadem).

0960-1481/$ - see front matter r 2006 Published by Elsevier Ltd.

doi:10.1016/j.renene.2006.02.011
 ARTICLE IN PRESS
2330 S.K. Khadem, M. Hussain / Renewable Energy 31 (2006) 2329–2341

renewable resources has been growing steadily. Countries having locations where the wind
speed is higher than 5 m/s or so have been going for wind energy generation. Smaller
generators are being designed to work for speeds above 2.5 m/s. Windmills for
water pumping or other mechanical activities continue to be used where wind speed is
2 m/s or so.
In Bangladesh, adequate information on wind speed and particularly at hub heights
of wind machines is not available. A previous study showed that the wind monitoring
stations of Bangladesh Meteorological Department (BMD), situated in built-up
areas, measure low wind speed near the ground level at a height of around 10 m.
Hussain et al. analysed using data from 14 locations of BMD and surveyed wind
speed in coastal locations using handheld anemometers and suggested that
Chittagong–Noakhali seacoast and coastal islands may have a good many locations with
a fairly high wind speed [1,2]. After that, some measurements were done by Bangladesh
Centre for Advance Studies (BCAS), GTZ—a German Organization in collaboration
with Rural Electrification Board (REB) and Local Government Engineering
Department (LGED), Bangladesh Council for Scientific and Industrial Research (BCSIR)
and now some studies are being carried out by LGED and Bangladesh Power
Development Board (BPDB) which show that the coastal belt and islands are suitable
for wind generators [3–5].
Kutubdia, an island close to the eastern coast of Bangladesh, with an area of
215.8 km2, has the status of a sub-district having a population of about 0.12 million.
Most of the island is surrounded by embankments (some acting as roads) and there are
cyclone shelters all over the island. The main occupations of the population of Kutubdia
are agriculture, livestock, forestry, fishery and salt-making. Several rice husking mills,
shrimp farms, salt-processing plants, primary and higher secondary schools and a health
complex are situated in this island. BPDB installed some isolated generating sets totaling
1.4 MW. But only 160 KW is running and this serves a small number of consumers
including only 2.6% of households and for a limited number of hours in the
evening [6].
In Kutubdia, the BMD station (91.851E, 21.821N) finds the annual average wind
speed over the last 10 years (1991–2000) to be 1.8 m/s at a height of 13 m (the
building of the met station is 11 m high and the anemometer height is 2 m above
the roof). But BCAS finds that their measurement over a year at a different site,
close to the seashore, (91.841E, 21.861N) leads to an annual value of 4.4 m/s at a height of
25 m [4]. One expects that at 50 m height higher speeds should occur. On the
other hand, NASA data averaged over 11 latitude
11 longitude area shows a lower
value of 3.6 m/s at 50 m height around Kutubdia [7]. The present work attempts to assess
the wind resource over the island and to predict good locations for wind generators. For
this purpose, Wind Atlas Analysis and Application Program (WAsP) [8], developed at
RISO National Laboratory, Denmark, has been employed for vertical and horizontal
extrapolation which uses time series of wind data along with information on surface
characteristics of the location, obstacles due to buildings and trees close to the met station
and the orography. It may be mentioned that WAsP program has been employed as a
micro-scale modeling tool to develop the wind atlas map in more than 100 countries and
territories around the world [9,10]. Logarithmic law [11,12] was also employed for vertical
extrapolation but it shows remarkable difference between measured and predicted values
for BCAS location.
 ARTICLE IN PRESS
S.K. Khadem, M. Hussain / Renewable Energy 31 (2006) 2329–2341 2331

WAsP has been used to develop the frequency response of the wind speed, wind rose,
monthly and annual Weibull parameters (shape factor, K, and scale factor, A) and the
wind atlas at 50 m height from the BMD and BCAS data of Kutubdia. An attempt has
been made to predict the wind speeds for the BCAS location at 25 m height from the wind
atlas developed using BMD data and to compare them with experimental data (Table 3). A
close agreement found between them validates the technique. The monthly and annual
wind atlas over Kutubdia for four roughness conditions and for five heights have been
developed and a pre-feasibility analysis has been done to obtain the wind speed and power
density for the selected locations at 50 m height from BCAS station measured at 25 m
height.

2. Wind speed data for Kutubdia

Three-hourly time series data of wind speed and direction for 10 years (1991–2000) of
BMD, Kutubdia, has been collected. BMD uses a cup anemometer made by Casella,
takes readings manually and anemometer checking is also done manually by
comparing with a handheld new anemometer. Raw data at 10-min intervals for the
BCAS location has been collected from September 1996 to February 1998. BCAS also used
a cup anemometer, testing and calibration were done at Harwell, England, during May
1996 in accordance with the user’s guide provided by Campbell Scientific Ltd. As the
measurements were taken only for 12–18 months, further calibration was not done. Fig. 1

Fig. 1. (a) Kutubdia Island showing the positions of BCAS and BMD stations at Kutubdia and (b) map of
Bangladesh showing the location of Kutubdia.
 ARTICLE IN PRESS
2332 S.K. Khadem, M. Hussain / Renewable Energy 31 (2006) 2329–2341

shows the position of the monitoring stations in the map of Kutubdia Island and the
location of Kutubdia Island in Bangladesh. The distance between the two monitoring
stations is around 3 km.

3. Data analysis

Both BMD and BCAS data have been analyzed to get the diurnal, monthly and seasonal
variations. It has been found that during the windy months (May to August), the diurnal
variation in wind speed is low but for the rest of the months it is high (Fig. 2). Fig. 3 shows
that the monthly variation in wind speed is also high. Measured wind speed for BMD
station, Kutubdia, is 1.9 m/s for the period of September 1996–August 1997 with a
standard deviation of 0.36 m/s over the 10 years. As energy depends on the cube of speed,
v3 , it is therefore expected that available energy should be higher in Kutubdia than for
locations having the same annual wind speed with a low speed variation. Fig. 4 shows
remarkable variations in wind direction between the months of June and October at BMD,
Kutubdia.

4. Prediction of wind speed

Ground-level obstacles such as vegetation, buildings and topographic features tend to

slow down the wind near the surface of the earth. Since the effect of these obstacles
decreases with height above ground, wind speeds tend to increase with height above
ground. This variation of wind speed with height which is called wind shear could be

Fig. 2. Hourly variation of wind speed.

 ARTICLE IN PRESS
S.K. Khadem, M. Hussain / Renewable Energy 31 (2006) 2329–2341 2333

Fig. 3. Monthly variation of wind speed.

Fig. 4. Wind rose, at BMD position using 10 years data, for the months of (a) June—average speed: 2.8 m/s and
(b) October—average speed: 0.9 m/s.

described by the logarithmic law

u Z
V¼ ln . (1)
k Z0
pffiffiffi
Here u is the friction velocity which is equal to t=r where t is the surface shearing stress
representing the force exerted on the surface by air and r is the density of air. Z is the
 ARTICLE IN PRESS
2334 S.K. Khadem, M. Hussain / Renewable Energy 31 (2006) 2329–2341

Fig. 5. Elevated position of the surface changes due to vegetation.

Table 1
Predicted wind speed at 25m from BMD data and with measured values of BCAS

Month (September 1996–August 1997) Measured Predicted

BMD (13 m) BCAS (25 m) (using log law & BMD data) (25 m)

Jan 1.4 3.8 2.3

Feb 1.5 3.5 2.4
Mar 1.7 3.9 2.7
Apr 1.5 3.6 2.4
May 2.6 5.0 4.0
Jun 2.6 6.1 4.1
Jul 2.6 6.3 4.1
Aug 2.3 5.5 3.6
Sep 1.4 3.6 2.2
Oct 1.5 4.0 2.3
Nov 1.1 3.5 1.8
Dec 1.5 3.7 2.3

Annual 1.9 4.4 2.9

height and Z0 is the aerodynamic surface roughness height. The value of Z0 for different
types of surfaces varies widely between fractions of a cm to a meter [11,12]. Again, the air
flow between the lower atmosphere and tall vegetation changes the elevated position of the
active surface as shown in Fig. 5 and therefore the logarithmic wind profile equation
(eq. (1)) should be modified as [11]

u Zd
V ¼ ln (2)
k Z0
 ARTICLE IN PRESS
S.K. Khadem, M. Hussain / Renewable Energy 31 (2006) 2329–2341 2335

where d is the zero plane displacement ðd ¼ 2h=3Þ and h is the height of the vegetation
stand. The ratio of eq. (2) for two heights was then used to predict the wind speed at a
height of 25 m using the data of BMD (Table 1).
It shows that there is a remarkable difference between prediction from log law and the
measured values. A field visit was done for both of the locations and it has been found that
the BMD station is surrounded by obstacles like trees, a cyclone shelter building and other
community buildings. On the other hand, BCAS station is in the seashore area with very
few obstacles of low height. In general, log law is used to get rough estimation or
prediction of wind speed for a certain height. But there is no way to consider the obstacles
and terrain effect in the log law. Beside these effects, hourly and monthly wind speed
variation is also high for Bangladesh. So, we find that log law is not accurate enough for
time series hourly energy assessment. Hence, WAsP program has been used for the
assessment.

5. Wind energy assessment using WAsP

WAsP is based on the physical principles of flow in the atmospheric boundary

layer and takes into account the effects of different surface roughness conditions,
sheltering effects due to buildings and other obstacles, and the modification of the wind
imposed by the specific terrain height variations around the met station. Latitude,
longitude and anemometer height are the pre-requirements to input the data for any
location. Considering the effects for the obstacles, roughness and terrain, the WAsP
develops a wind atlas for a region around 100 km2 in area [8]. Fig. 6 shows the WAsP
methodology to generate regional wind climatology or wind atlas and to predict wind
climate.

5.1. Models for obstacles, roughness and terrain effect

In the WAsP, base map of any location is used to draw the roughness and contour lines
according to the land use information and terrain description of the location. Based on the
surface characteristics and corresponding roughness values [8,11,12], WAsP develops
roughness models both for BMD and BCAS locations. These roughness models show that
the roughness value for coastal areas lies between 0.001 and 0.03 m, and as the met station
is situated within a built-up area, the roughness value lies between 0.1 and 0.4 m. It may be
mentioned that WAsP develops the roughness model considering the land use information
around 5–10 km area for all directions from the point of interest.
For BMD, Kutubdia, as the mast is on the roof of the met station building (mentioned
above), hill effect has been introduced in the WAsP model by modifying the ground into
1:5 slope and it is done by drawing two separate closed contour lines one for the building
height and another one for the ground height above the sea level [13]. In this case, the
anemometer height to be used should be its height from the roof. It may be mentioned that
the ground elevation over Kutubdia varies from 0 to 4 m only above the sea level, and for
BMD and BCAS locations, it is 3 and 2 m, respectively.
Again, the BMD station is situated in an inhabited locality and the sheltering effect, due
to obstructions for wind flow by buildings and trees close to the met station, is to be
considered for that location. The height and depth of an obstacle, its angular position and
distance from the monitoring mast and the porosity (0–1) have to be taken into
 ARTICLE IN PRESS
2336 S.K. Khadem, M. Hussain / Renewable Energy 31 (2006) 2329–2341

Fig 6. WAsP methodology.

consideration to develop the obstacle model. In this model, the obstacles that are found
within the slope (1:5) range with heights equal to or less than the height of the met station
need not be considered [14]. For Kutubdia BMD, only one obstacle, a cyclone shelter
building having an area of (8
10) m2 and height of 13 m, located very close to the met
station—around 10 m distance in the north direction—is higher than the met station and
has been considered. For BCAS location, obstacles are very less than that of BMD and
there are only some rows of trees with low density. Fig. 7 shows the roughness and
obstacles models developed in the WAsP program.
 ARTICLE IN PRESS
S.K. Khadem, M. Hussain / Renewable Energy 31 (2006) 2329–2341 2337

Fig 7. (a) Roughness model and (b) obstacles model for the BCAS location in the WAsP program.

Using the logarithmic wind profile, surface-layer similarity laws, geostrophic drag law
and the geostrophic wind, the stability model, the roughness change model, shelter model,
the orographic model from the terrain information and analyzing the time series raw data
of wind speed and direction, we obtained from WAsP analysis the average wind speed,
power density, frequency response, wind rose and then a wind atlas/regional wind climate
in the form of Weibull parameters.

5.2. Developed wind atlas

After analyzing the artificial hill effect due to the met building, effect of the single obstacle
close to the anemometer mast, roughness of the locality and terrain information the monthly
and annual wind atlas over Kutubdia Island has been developed using 1 year (September
1996–August 1997) as well as 10 years (1991–2000) three-hourly time series data of BMD and
1 year (September 1996–August 1997) 10-min interval time series data of BCAS. The annual
wind atlas at different heights with roughness change using both BMD and BCAS data for 1
year (September 1996–August 1997) have been shown in Table 2. Results show that the wind
speed values are almost same for both of the atlas but the power density values for BMD atlas
are quite higher than that of BCAS atlas. Fig. 8 shows the wind direction for the annual
observed wind climate of BCAS, Kutubdia, at a height of 25 m and the developed wind atlas
for Kutubdia at a height of 50 m for the roughness of 0.03 m.

5.3. Comparison between measured and predicted values

A comparative study was done between the measured values of the wind speed at BCAS
location and the predicted values at that position for 25 m height from the wind atlas
obtained for September 1996–August 1997 period using BMD data. Prediction for BMD
 ARTICLE IN PRESS
2338 S.K. Khadem, M. Hussain / Renewable Energy 31 (2006) 2329–2341

Table 2
Annual wind atlas for Kutubdia using 1 year data of BCAS and BMD showing the predicted wind speed (U) and
power density (P) for four roughness classes and different heights

Height Parameter BCAS (Annual Wind Atlas) BMD (Annual Wind Atlas)

Roughness Roughness

0.00 m 0.03 m 0.10 m 0.40 m 0.00 m 0.03 m 0.10 m 0.40 m

10.0 U (m/s) 4.9 3.4 3.0 2.3 4.6 3.2 2.8 2.2
P (W/m2) 164 68 44 21 227 103 68 32
25.0 U (m/s) 5.4 4.1 3.6 3.1 5.0 3.8 3.4 2.8
P (W/m2) 208 104 76 45 283 153 111 65
50.0 U (m/s) 5.8 4.7 4.3 3.7 5.4 4.4 4.0 3.4
P (W/m2) 251 141 108 71 334 192 149 99
100.0 U (m/s) 6.3 5.6 5.1 4.4 5.8 5.2 4.7 4.1
P (W/m2) 330 220 166 109 441 281 212 139
200.0 U (m/s) 6.9 6.9 6.3 5.4 6.4 6.4 5.8 5.0
P (W/m2) 473 445 326 206 635 560 419 264

Fig. 8. (a) Annual wind climate for BCAS, Kutubdia, at a height of 25 m and (b) annual wind atlas over Kutubdia
at a height of 50 m for roughness 0.03 m.

position also has been done using the wind atlas developed by BCAS data for the same
period. Table 3 shows that the measured and predicted values of wind speed and power
density for BMD are much closer than that for BCAS location. It has also been found that
as the obstacles, terrain and roughness effects are high for BMD station and data also
manually recorded for three-hourly intervals, therefore wind atlas obtained by BCAS data
will be more accurate than that of BMD. Also BCAS wind atlas gives lower prediction of
power density. So it’s better to use the conservative way. And the prediction should
become more accurate if data were available for a longer period.

5.4. Wind resource map for Kutubdia—selection of good locations

The wind atlas developed by BCAS data for 1 year, shown in Table 2, was used to
generate the wind resource map at 50 m height with the help of (100
100) m2 grid cells.
WAsP calculates and shows the Weibull parameters, wind speed and power density for
each of the grid cells taking into account the roughness and terrain effect but not the
sheltering effect for the desired location. The wind resource map for Kutubdia at 50 m
height is shown in Fig. 9 along with the minimum, maximum and average values of Weibull
 ARTICLE IN PRESS
S.K. Khadem, M. Hussain / Renewable Energy 31 (2006) 2329–2341 2339

Table 3
Measured and predicted values of wind speed (U) and power density (P) for BMD and BCAS locations for the
period of September 1996–August 1997.

Month BMD (13 m) BCAS (25 m)

Measured (U/P) Predicted (using Measured (U/P) Predicted (using

BCAS atlas) (U/P) BMD atlas) (U/P)

Jan 1.4/03 1.6/04 3.8/059 3.5/055

Feb 1.5/04 1.5/04 3.5/044 3.5/053
Mar 1.7/12 1.8/09 3.9/086 4.0/102
Apr 1.5/06 1.6/08 3.6/074 3.1/057
May 2.6/68 2.3/23 5.0/183 4.8/332
Jun 2.6/51 2.7/22 6.1/207 5.2/393
Jul 2.6/33 2.9/23 6.3/224 5.2/251
Aug 2.3/21 2.5/20 5.5/182 4.4/148
Sep 1.4/03 1.6/06 3.6/053 3.0/037
Oct 1.5/09 1.8/11 4.0/097 2.8/059
Nov 1.1/02 1.5/03 3.5/036 2.7/035
Dec 1.4/03 1.5/04 3.7/049 3.6/071

Annual 1.9/24 2.0/14 4.4/125 4.3/190

Fig. 9. Wind resource at 50 m height above the ground level for Kutubdia Island.
 ARTICLE IN PRESS
2340 S.K. Khadem, M. Hussain / Renewable Energy 31 (2006) 2329–2341

Table 4
Predicted wind speed (U) and power density (P) for BMD, BCAS and other three other locations in Kutubdia at
50 m height using wind atlas of BCAS

Month BMD BCAS Site 1 Site 2 Site 3

Jan 4.5/097 4.5/093 5.1/141 4.9/120 4.9/120

Feb 4.2/072 4.2/069 4.6/095 4.6/090 4.4/086
Mar 4.4/122 4.5/121 4.5/123 4.7/145 4.4/122
Apr 4.0/108 4.0/102 4.1/121 4.2/122 4.0/104
May 5.5/262 5.6/259 5.5/273 5.9/322 5.5/276
Jun 6.5/274 6.7/298 6.6/286 7.5/405 7.2/373
Jul 7.0/307 7.2/324 7.2/334 8.0/447 7.7/407
Aug 6.1/264 6.2/268 6.2/295 6.9/393 6.7/368
Sep 4.2/081 4.2/080 4.4/093 4.6/108 4.4/099
Oct 4.8/153 4.8/150 5.4/190 5.4/212 5.3/201
Nov 4.1/061 4.1/060 4.7/099 4.6/086 4.6/091
Dec 4.2/082 4.2/080 4.9/130 4.7/109 4.7/115
Annual 5.1/183 5.2/183 5.5/208 5.7/244 5.5/222

Table 5
Predicted wind speed (U) in m/s and power density (P) in W/m2 at different heights

Height (m) BMD, U/P BCAS, U/P Site 1, U/P Site 2, U/P Site 3, U/P

30 4.6/138 4.7/132 5.0/169 5.4/211 5.2/189

50 5.2/183 5.2/183 5.5/208 5.7/244 5.5/222
65 5.5/219 5.6/224 5.7/246 5.9/274 5.8/255
75 5.7/245 5.8/249 5.9/268 6.0/291 5.9/274
90 5.9/277 6.0/285 6.1/298 6.2/314 6.1/302

parameters, wind speed and power density for all over Kutubdia. Monthly wind speed and
power density obtained for three good locations, considering only the terrain and roughness
effects, have been shown in Table 4. As the selected sites are close to the seashore, it may be
expected that trees and buildings may not be high enough for obstacle effect.
Wind resource classification system, developed by US National Renewable Energy
Laboratory, is perhaps the most widely used and designed for a nominal height is 50 m,
[15] whereas the standard height for large wind turbines is now 65 m or more [16].
Therefore, wind speed and power density for different heights have been estimated for the
selected sites of Kutubdia which are shown in Table 5. Based on these classifications, it is
found that the selected sites are in the class of fair and good at 30 m height and are
sustainable for small turbines [16] whereas at 50 and 60 m heights, the power densities are
marginal and fair [15,16] for large turbines.

6. Conclusion

The computed wind speed of BMD from BCAS data and BCAS from BMD data have
been found to be close to the measured values (Table 3). This validates the WAsP
technique employed. As power density is lower for BCAS than for BMD, therefore in
 ARTICLE IN PRESS
S.K. Khadem, M. Hussain / Renewable Energy 31 (2006) 2329–2341 2341

order to make conservative prediction, wind atlas obtained from BCAS data has been
considered for detail assessment and analysis. Hence, the estimations shown in Tables 4
and 5 should be dependable. However, the wind speed varies from year to year, therefore
measurements at 50 m height and above at selected locations would be welcome. If
sheltering effect is present, the speed would be lowered and sites without serious obstacles
may be selected for wind generators. This study and analysis provides an assessment of
wind energy over Kutubdia Island. From Fig. 8 and Tables 4 and 5, it appears that the
Southern and Eastern coast parts of Kutubdia are better than other parts for wind
generators. At the coast side for monsoon (May–August) period, the wind speed varies
from 5.5 to 8.0 m/s and for rest of the month it varies from 4.0 to 5.4 m/s. And at the South
and Eastern side the power density is found to be over 200 W/m2 for the year September
1996–August 1997. Therefore, it can be concluded that

 measurement at 50 m height or above for a long period is needed for feasibility study of
large turbine installations,
 small turbines at 30 m height or above will be more sustainable and
 South and Eastern part of the island should have more potentiality for wind electricity
generation.

Acknowledgment

The authors are very much thankful to UNEP/GEF for providing funds for Solar and
Wind Energy Resource Assessment (SWERA), Bangladesh, project under which the
analysis has been done. Bangladesh Meteorological Department is thanked for making
their data available.

References

[1] Hussain M, Alam S, Reza KA, Sarkar M. A study of the wind speed and wind energy availability in
Bangladesh. Energy Convers Manage 1986;26(3/4):321–7.
[2] Sarkar M, Hussain M. The potential of wind electricity generation in Bangladesh. Renew Energy
1991;1(5/6):855–7.
[3] Wind Data Evaluation Final Report by GTZ ‘‘TERNA Wind Energy Programme’’, Bangladesh, 1997.
[4] Wind Energy Study (WEST) Project, Bangladesh Centre for Advanced Studies, Final Report, Bangladesh, 1998.
[5] Wind Energy Resource Mapping, LGED, http://www.lged-rein.org/wind/werm.htm
[6] Hossain, A. IIFC report ‘‘Remote Area Power Supply Systems (RAPSS)’’, 2001.
[7] NASA, SSE database, available from URL: http://eosweb.larc.nasa.gov/sse
[8] Troen I, Petersen EL. European Wind Atlas. Denmark: Risoe National Laboratory; 1989.
[9] The World of Wind Atlases–Wind Atlases of the World, http://www.windatlas.dk
[10] Solar and Wind Energy Resource Assessment, project of UNEP/GEF, http://swera.unep.net
[11] Manwell JF, McGowan JG, Rogers AL. Wind energy explained. New York: Wiley; 2002.
[12] Oke TR. Boundary layer climates. London: Routledge; 1996.
[13] Lars L. The mast on the House. Wind Energy 2000; 3.
[14] Khadem S.K. Badger J. Ullah S.M. Aditya S.K. Ghosh H.R. Hussain M. The effect of obstacles close to the
anemometer mast located on a building on wind flow in the WAsP model. RETRUD-03 conference,
Kathmandu, Nepal, 2003 p. 323–26.
[15] Wind Energy Resource Atlas of Sri Lanka and the Maldives, National Renewable Energy Laboratory, USA,
2003, http://www.nrel.gov
[16] Wind Energy Resource Atlas of Southeast Asia, The World Bank Asia Alternative Energy Program, 2001,
http://www.worldbank.org/ASTAE