Content uploaded by Sayantan Das
Author content
All content in this area was uploaded by Sayantan Das on Dec 19, 2018
Content may be subject to copyright.
Geographical Review ofIndia 75 (3) September-2013, 281 -296
Assessment of Impervious Surface Growth in the Mula-mutha
Watershed in Maharashtra
Sayantan Das', Amit Dhorde2, Anargha Dhorde3
ABSTRACT: The Mula-Mutha watershed is located in the western Maharashtra which nests Pune urban
agglomeration at its central part. For the assessment of impervious surface growth of this watershed, satellite
images of 1989 and 2008 have been used. An integrated application of remote sensing techniques using
Normalized Difference VegetationIndex (NDVI) and Geographic Information System (GIS) has been carried
out to identify the amount of impervious surface and its variations. The Impervious surface analysis reveals
that there has been almost three-fold increase of the built up area under very high impervious class representing
the Pune city and its surroundings between 1989 and 2008. The entire scenario is indicating the requirement
of immediate remediation through pollutant mitigation and resource restoration in those areas.
Keywords :impervious surface, watershed, normalized difference vegetation index, built up area.
Introduction
For the last couple of decades or so, large scale human activities have brought about an
enormous change on the Earth's surface. The effects of urbanization and industrialization are clearly
noticeable, especially in the urban areas. But their indirect influences can be observed in the
adjacent areas of the city too. Land-use changes occur at many places and it is observed that
vegetation and barren land have changed into non-evaporating and non-transpiring impervious
concrete surfaces. Natural land surface cover has been removed continuously and this bas led to a
simultaneous increase in impervious surface cover. It not only hampers the natural eco-system, but
also has long term effects on the climate, natural resources, economy and culture of the surrounding
areas.
The impervious surfaces have been increasing rapidly and those are mainly associated with
the cities and heavily populated areas. United Nations (2002) has projected doubling of the global
urban population by 2030. The study of spatio-temporal change in impervious surfaces would
facilitate the understanding of awatershed containing abig urban environment and its relationship
with urbanization and industrialization.
IResearch Scholar, Department of Geography, University of Calcutta
2Assistant Professor, Department of Geography, University of Pune
3Assistant Professor, Department of Geography, Nowrosjee Wadia College, Pune
282 GEOGRAPIDCALREVJEW OF IND TA VOL. 75
Mapping of Impervious Surface: Impervioussurface mapping is an attempt to track the pattern
of land use changes. Impervious surfaces are defined as any surface which does not allow water to
infiltrate and are primarily associated with transportation (streets, highways, parking lots and
sidewalks) and building rooftops. Imperviousness directly affects the amount of runoff to streams
and lakes and is related to non-point source pollution and water quality of surrounding water
bodies. The amount of impervious surface in alandscape is an indicator of environmental quality
(Arnold and Gibbons, 1996).ln anatural landscape,the maximum amount of runoff occurs after the
beginning of astorm. As impervious surface area increases, the storm water coming off of them
increases velocity,quantity, temperature, and pollution load. Any one of these attributes contributes
to the degradation of natural hydrology and water quality. Noticeable degradation of water bodies
begins when the watershed reaches 10-20 percent imperviousness. Quantifying and analyzing
impervious surfaces is an important step in determining the current condition of awatershed. It can
serve as a key ingredient to carry out further research for determining land use planning implications
and directing future decision-making processes for ecologically sensitive zones. Ageneral
relationshipshowing the health status of awatershed decreasing with increasingimpervious coverage
was given by Schuler in l994 (fig. l). He stated that the health of awatershed can be categorized as
protected, impacted and degraded, depending upon the percentage of impervious surface available.
Protected '"""'*"
Fig. IRelationship of imperviousness to the health status of the area
(Modijiedfrom Schuler, 1994)
,.,.
30
Punt' Dislricl in
M11l1111
.. ? ... 111
20
N
A
N
A
0100 200 300 400
--===--c===ikm
100
MULA-MUTHA CATCHMENT
Maharashtra
in India
.......
N
A
50 10 15 20
--==--===ikm
Fig. 2Location Map of the Study Area
N0.3 ASSESS?1EXTOF ThlPER\ lOUS SURFACE GROWTH ........ 283
He identified certain threshold values in order to define the health status of the watershed. If the
region has less than 10 percent o!- unperv ious surface then it is treated as protected, if it is from 10
percent to 25 percent then it is treated as impacted and above 25 percent of impervious surface
indicates adegraded state. This categorization is followed by almost the entire scientific community
working on the studies related 10 imperviousness. The results of the impervious surface analysis
can be used to guide planning emphasis within each local area. For areas in the lower impervious
zone. emphasis should be placed in preventive measures that retain existing natural systems, using
techniques hke open space planning and stream buffers.
The main aim of the study is to bring about spatio-temporal variations in impervious surfaces
of the Mula-Mutha watershed during the period of 1989-2008. To fulfill this, following objectives
have been outlined:
•To extract impervious surfaces from satellite data.
•To evaluate the Normalized Difference Vegetation Index (NDVl) and impervious surface
within the watershed for the study period.
Data Acquisition
Landsat TM 1989
IRS P6 LISS ill -2008
!
Preprocessing of the RS data
Geometric Correction
Rau..imetnc and Atmospheric correction
Reprojection
fr"•?,:emen1
I
..
..
l'sD\. l:::e;er Water Masking
Combine function
Fig. 3. Flaa-?
Extraction of Impervious
surface area (ISA)
'1:1£ the methodology adopted for the study. Source :Prepared by the authors
284 GEOGRAPIDCALREVIEW OF INDIA VOL.75
Study Area: The area under consideration is Mula-Mutha watershed (l 8° 18 'N -18°44 'N,
73°20'E -74°20'E) which includes Pune city, which is rapidly growing in terms of the urban space.
The location of the watershed with reference to India and Maharashtra is shown in the figure-2.
Mula and Mutha rivers originate from Sahyadri ranges lying in the west of the watershed and they
are the tributaries of the Bbima River. The region lies in the Deccan trap volcanic province. Mula
and Mutha rivers and all their tributaries are of seasonal type. They receive water from the monsoon
rainfall. The climate of the watershed is dry but invigorating. The area lies in the rain shadow zone
of the Western Ghats. Semi-arid climate with monsoon rainfall is the common characteristic of the
region.
Methodology
Data Acquisition and Preprocessing: For the present work, Landsat 4 TM and IRS P6 LISS ID
images were used (table l). The geometric correction of Landsat 4TM (1989) image was done by
usingERDAS Imagine 9.2 software. The IRS P6 LISS III (2008) imageacquiredfrom National Remote
Sensing Centre (NRSC), Hyderabad had already been corrected geometrically. For radiometric and
atmospheric correction of the images of 1989 and 2008, the models built in ERDAS imagine 9.1 by
Dhorde and Wakhare (2008) was employed.
TABLE- I:Satellite Data used for the Study
Satellite Sensor Date
IRSP6 LlSSlll 14111February, 2008
LANDSAT 4TM ll III February, 1989
Source :Prepared by the authors
Reprojection and Resampling: The images were reprojectedto Geographic Latitude/Longitude
with the WGS 84 spheroid and datum. IRS P6 LISS III image of2008 had aspatial resolution of23.5
meter whereas the resolution of Landsat 4 TM 1989 image was 30 meter. In order to make both these
images comparable the IRS image was resampled and brought down to 30m resolution. This process
can be attempted in both ArcGIS as well as ERDAS imagine softwares.
Image Enhancement: Simple enhancement techniques like contrast-stretch, histogram
equalization etc. were performed for visual interpretation.
Obtaining Subset for the Area of Interest: The area of interest (AOI) was obtained with the
help of avector layer prepared in the same projection as that of the images. This AOI is the
boundary of the Mula-Mutha watershed. Data subsetting was done for the images of Landsat 4
TM and IRS P6 LISS ill. These subsets were further employed for obtaining desired parameters.
Calculation of Normalized Difference Vegetation Index (NDVI): The NDVI calculation was
done in ERDAS imagine by running the indices option. Normalized difference vegetation index uses
N0.3 ASSESS\1£'1. TOF Th1PER\10US SURFACE GROWTH .285
the combraarion of'oand 3n63...(\ omm) i.e red band and band 4(0.76-0.90mm) i.e. near infrared
band for ?: ?Thi irr-? 'D\l i, representative of plant assimilation condition and its
pbotesynthetic appara ..
"US -".'-. -.l:.ld le: mass concentration (Groten, 1993).NOVI is representative
.of plan, pbo.o?-=ilietic e-?I· snme varying due to the changes in meteorological and
environmental p.a.--amete.: He:?..:h_vegetation will have ahigh NDVT value. Bare soil and rock
reflect sumla: ?-e- ??red aad so "ill have ?'DVT values near zero. Clouds, water,
and sno- are dJ-e in tz.at they reflect more visible energy than infrared
.UVI \':2: ,?'D\ 1for Landsat T11. ETM+ and IRS LISS lll,
1'1R -R
?lR-R
ctT":1n11?11 ?method for comparingvegetation
:. :). Ia tais research. cells with low NDVT
..... -?-;-es-pondco the cells of potential
zes •;;:;:z::? ?tt? ..... egetanon in the Great Plains and remains as
e?:. p)a=,1covers in the satellite data.
? for Landsat 4nt and IRS LISS m Images: Out of anumber
of methods •..:..-d ""eralllre and employed in the extraction of Impervious Surface Area
(ISA .arn:"'? r. _=:!OD ilie basic factor of vegetation cover was selected for the extraction of
impervi as s:.:.rf -?'Tl" for me concerned watershed, An integrated application of satellite remote
sensing :ed:ci.,_.... ::. ??; .Tormahzed Difference Vegetation Index (NOVI) and Geographic
Info? _ystem ulS is carried out to estimate the amount of impervious surface and their
variation f ?the time period spanning twenty years (1989-2008). Here impervious surface is derived
using ihe :bssifiec Iaad use image and the ?'DVI image. The NOVI image obtained is further used
to caiegonze me levels of imperviousness and then employed to generate the percentage of ISA
per ca:e-; -: of mv1. TI-.e table 2below is showing Translation of NOVI range into imperviousness
range
=?· r. is. observed that the values for water and developed areas (i.e. impervious
surface zeaes b:!ve coincided with each other. So water might have got classified as an impervious
layer ·- to the bailt up structures. Thus, abias might have been introduced in the analysis
pan. S??task involved was to mask out water from the NOVI rasters.
\\attt Mas •?: For doing water masking ArcGIS spatial analysis capabilities are used. lo
thi,- process. other ?d use classes except the water bodies have been eliminated. The table
gh en ve -e3) illustrates bow the water class was separated from the NDVI integer raster
286
High
Low
GEOORAPHICALREVIEW OFINDIA
TABLE-2 :Translation of ND VI range into Imperviousness range
lmpemoaaea Values NDVI Interpretation
-1
-0.9
-0.8
-0.7 Barren areas of
High lmpcrviousness -0.6 rock/sand/water
(Before Water masking) -0.5
-0.4
-0.3
-0.2
-0.1
0
0.1
Medium 0.2
Imperviousness 0.3 Built up area
0.4
0.5
0.5 to I0.6 0.5 to 1
Low imperviousness 0.7 HighVegetation
0.8
0.9
1
VOL.75
High
Low
Source :Prepared by the authors
TABLE-3 :Separating water class from NDVI raster
Ruter! Ralter2 Reclas of Combhle Ruter Reslllt Value
NDVl Integer Supervised classified I-Densely built-up, sparsely built-up, agriculture,
file (1989 &2008) barren, fallow, hills, hill slopes
0-Water bodies
Source: Prepared by the authors
N0.3 ?ESS?tD-"TOFL\lPER\10CSSURFACEGROWTH .287
Combining ?l)\l integer and Water class: Combining is done by theArcGIS software which
works O:l.? \\-di the ?TI,? ?.e;er values. So. it was necessary to convert the decimal NOVI values
into the ?:eger values. Thi, con, ersion was performed in the ERDAS Model builder wherein a
simple model was built to convert >.TI\ 1 decimal values to NOVI integer values. For the combine
funcnon. u.-a;er .:::."?.?ec. and ?TI\ 1 images were used. Interestingly, the nature of the NOVI raster of
having higher \alues for vegetation and lower values for impervious area reversed after it was
combined ... in die .,,.a.er raster. The tables below (table 4a and 4b) are showing that.
TABLE-4a: Differences between NDVI Integer raster and Combine raster for 1989
l\1)\1 NDVI An:GIS Combine
lnterpffbriN Integer Value Combine vahle hterpretatien
Impervious -100 ILmpervious
To To
Vegetation 100 365 Vegetation
Source :Prepared by the authors
Table-4b :Differences between NDVI Integer raster and Combine raster for 2008
'.\1)\'J I.iDVI ArcGIS Comblae
lnterpretatioa 1ateca-\'.a. c.. 1ueva1ae laterpretatloo
Impervious -65 IImpervious
To To
Vegetation 35 II (.!: Vegetation
Source: Prepared by the authors
After combimag. )::-eal:-u;-,alues sere obtained for 1989 and 2008 combined images on the
basis of narura, breaz ::m.::.oc 'b: using :\rcGIS and these values are the indicators of imperviousness
of different par's .;:· ?e "atershed,
Impervieusness Indices: Deriving impervious surface area in square kilometer and percentage
was the .5nzl go-..:.
of the present study. For this the break-up values obtained were used to find the
area under each d3.SS and finally to compute the per cent imperviousness. The method given by
Sleavia (1999 Qi Prisloe et al. (2000) was adopted for the computation. Accordingly the impervious
surfa ... earea itlnn each class is calculated as:
?1)\'kbss .._...
:OVlclass""'
288 GEOGRAPHICAL REVIEW OF INDIA VOL.75
NDVIclass1Satca
is the total area of impervioussurface for aspecific NDVI combine class.
ND Vic lass=• is the total area for the same VDVI combine class.
With this method, NDVI data are converted to aGIS polygon format and overlaid on impervious
surface data extracted from the supervised classified image. This process was repeated for each
land use type for each NDVI combine class, resulting in aset of regionally calibrated impervious
surface coefficients.
Results and Discussion
Impervious Surface Area for 1989: NDV1 values for the 1989 image ranges from -1 to +1.
From these values, NDVI integer values were extracted which ranged from -100 to +100 (Fig. 4).
Higher NDVI values indicated highly vegetative areas whereas lower values corresponded with the
built up surfaces.
By combining the NDVI integer and water raster (Fig. 5) the combine image for 1989 was
obtained. The combined image took the values ranging from 1to 365. The combined image is
shown in the fig. 6. From this image, four specific break up values were assigned on the basis of
their potential imperviousnessviz. Ito 30 (Moderately Impervious Surface), 30-45 (High Impervious
Surface), 45-60 (VeryHigh Impervious Surface)and above 60 (Low Impervious Surface).
From the combined image it was found that moderately impervious surfaces are in the western
side of the watershed near the big reservoirs indicating the vegetative parts. High impervious
surface dominated most parts of the imageand mainly included the barren land. Very high impervious
layers can be observed in ascattered manner. Low impervious surfaces can be found at or near the
water bodies.
Built up class was extracted from the supervised classified imageof 1989 (fig. 7). The percentage
of built up for every combined break up class is calculated and the results are shown in the table 5.
As observed from the given table, very high impervious class contains almost 15 percent built
up area. The highest built up area percentage is found in high impervious class (15.5 percent).
Table-S :Impervious Surface Area (ISA) Calculation, 1989
Break updua Built up area Total area IS coefficient ISA percent
(in sq.km) (in sq.km)
Very High Impervious 83.30 562.60 0.1480 14.80
High Impervious 76.40 493.57 0.1547 15.47
Moderately Impervious 192.54 1385.79 0.1389 13.89
Low impervious 25.98 309.46 0.0839 8.39
Source :Prepared bythe authors from satellite data
.....
High: 100 ?0
0 5 10 20t<ilomoton 20Klomoten -Other Objects i
-- •• DWater Bodies
Low: ·100
..... ......
Fig. 4NDVT Integer Image, 1989 Fig. SWater Raster, 1989 '
..... ...... '"" I
?
?
I
"-- ?
?
?O..V- .,.ao
030·45 legend
0 5 10 20t<lomeln -41-to 0 5 10 20 l<JlomttOII
•• oeo-las •• •Built up
-?- ...... ··-
Fig. 6 Combined lmage, 1989 Fig. 7 Built up Area, 1989 ?
-..J
V,
N0.3 •
---- ?ITh-:-Or !!\!PER\ lOUS SURFACE GROWIB .289
Whereas. :1.e ...od...? imperioas class comprises more area than any other class and possessing
almost !- pc:-:-· -.,::ii:? area. As expected, the low impervious class contains the lowest percentage
built up area ( percent),
H? adiscrepancy can be noticed. Normally, very high impervious class (Class I) is expected
to have the h!ghest span of built up area. However, maximum proportion of built up area is found in
the high impervious surface class (Class II). It is observed that areas under Class Iappear to be
distributed in ascanered pattern as opposed to an idealistic clustering one. There is no compactness
identified in this class. So. built up areas occur in ascattered manner in this class. Low impervious
class unexpectedly has asignificant proportion of built up area, indicating occurrence of man-made
structures around the water bodies.
It is evident from these results that there is less concentration of impervious layers even in
the higher impervious classes. This indicates that Pune city didn't experience an enormous amount
of constructional activities before 1989.
ImperviousSurface Area for 2008: ).1)\'l values for the 2008 imageranges from -0.65 to +o.35.
From these values. XDVI integer values were extracted which ranged from -65 to +35 (Fig. 8).After
extracting the ).'DVI integer. again the water masking was carried out.
The combined image values ranged from 1to 142 (Fig. l0). From this image, four specificbreak
up classes were assigned again on the basis of their potential imperviousness viz. 1 to 15 (Moderately
Impervious Surface .15-r High Impervious Surface). 27-50 (Very High Impervious Surface) and
above 50 (Low Impervious Surface.
As compared to the 1989 combined image. the distribution of moderately impervious surfaces
were the same in the western side of the watershed. but in the downstream it has increased
considerably. High impervious surface again dominated amajor part of the watershed. Very high
impervious layers have been found in acompact manner. especially in the middle part where Pune
city is situated. Occurrence of low impervious surfaces remained more or less unchanged.
The percentage of built up for every combined break up class is calculated and the results are
shown in the table-6. As indicated in the table, very high impervious class contains 43.4 percent
built up area. 16.4 percent impervious surface area percentage is found in high impervious class.
The moderate impervious class comprises of only 9percent built up Area. Low impervious class
contains the most negligible percentage built up area 0.5 percent.
Here. expectedly the very high impervious class contains the highest percentage built up area.
As compared with the situation in 1989, the built up percentage under very high impervious class
has increased strikingly with the increase in built up area. The built up percentage in high impervious
layer is also greater than before. The moderate impervious class contains lower impervious surface
percentage than 19 9. Asignificant decrease in built up percentage is marked in the low impervious
class v.1:!: \e:: less area under built up.
,,..,...
,.. ...... ,.,,.,,.
..... ???-'-???????????-l·????????-,
7.
0
...
.....
Legend
-Built up
Il>QIIIIII
-0111111Ohj1>1
I•
W•lcir llotll••
Fig. 9Water Raster, 2008
......
05 10 20Kllomo10!$
··-
--
0510
•l• .... 1.-.mtH1
._..
('.Qf111>K...iC1aNV-
.,.,?
C 31021
•U·liO
t'J&0-1'2
"""
Fig. 8 NDVI Integer Imagi:, .1008
0 5 10
--
..
Fig. IO Combined Image, 2008 Fig. 11 Built up Area, 2008
N
'.0
VI
290 GEOGRAPIDCAL REVIEW OF INDIA
Table-6 :Impervious Surface Area (ISA) Calculation, 1989
VOL.75
Break up class Built up area Total area IS coefficient ISA percent
(in sq.km) (ID sq.km)
Very High Impervious 155.66 358.35 0.4344 43.44
High Impervious 228.54 1393.90 0.1640 16.40
Moderately Impervious 102.38 1126.10 0.090 9.09
Low Impervious 0.67 145.00 0.0046 0.46
Source :Prepared by the authors from satellite data
These results show that the constructional activities have increased rapidly in the last two
decades in the watershed. The city and its surroundings have experienced alarge scale change that
might have affected the residents.
Impacts of Occurrence of Impervious Surfaces: The imperviousness in different areas of the
watershed can be identified and specific actions can be taken on the basis of their vulnerability.
The main reason behind the imperviousness of an area is the built up surface developed there. So
more percentage area under built up indicates more impermeability as far as the impact is concerned.
According to Schuler (1994), awatershed is:
•Protected -at 10 percent or less impervious surface area,
•Impacted -at 10 percent-25 percent impervious surface area and
•Degraded -at greater than 25 percent impervious surface area.
By observing fig. 12a, we can assess the impact of the impervious surfaces in 1989. It can be
recognized that the watershed is largely impacted with 10 percent-25 percent areas under impervious
surface where preventive planning measure should be accompanied with afocus on site design
considerations that reduces runoff and imperviousness. Very less amount of area is found under
protected zone with less than 10 percent or less impervious surface area where emphasis should be
placed in preventive measures that retain existing natural systems, using techniques Like open
space planning and stream buffers. No area is found under degraded zone in 1989.
Fig. 12b is showing the impact of increasing impervious surfaces in 2008. It is clearly evident
from the image that the degraded areas with more than 25 percent impervious surface occur in Pune
city located in the central part of the watershed and Pimpri-Chindhwad city in its north-west. For
this zone, the remediation through pollutant mitigation and resource restoration should be given
importance. The other areas having the built up units are under the impacted zone with almost no
occurrence of the protected zone. So, it can be clearly stated that for the last 20 years, built up areas
have been increasing quite rapidly in and around the city region. Therefore the changing condition
of the watershed under concern is quite evident.
296 GEOGRAPHICAL REVIEW OF INDIA VOL.75
74'00'E
O 5 1o 20 30 40 Kilometers
11111111111c:::J111111ii=::J
Leaend
lmpervlo111 Surface Area 'Y.
Protected -1-lO'Y•
Impacted 10-25%
Dqraded ->25'Yo
7?'0"E
Fig. 12a Impervious Surface Area, 1989
7.-o'O"E
Degraded ->25%
Le&end
1.mpervkJus Surface Area •;.
Protected -1-10%
,....,...N
10-250/o
Impacted
O 5 1o20 30 40 Kilometers
111111111c=::::11111-=:::J
W30'0'E
Fig.12b. lmpervious Surface Area, 2008
N0.3 291
SIIIIlDUI)and c·ooctmil:ia
The ?undettak=n has helped to understand changes in impervious surfaces in the Mula-
Mutha "arersbed over the :?, rwo decades. The study also highlights the expansion of Pune city
and surrouadings over the same period. Impervious surface mapping is an attempt to track the
pattern of land U5e changes The increasing amount of impervious surfaces indicates the expansion
of urban areas with anumber of industrial and residential projects.
Major findings of the Study:
•Very high impervious class contains almost 15 percent built up area in 1989; whereas, in
200 .-13.4 percent built up area is found in the very high impervious class. So there is
almost three times increase in the built up area under this class in the last two decades.
•The built up percentage under the high impervious class has increased slightly, only about
Ipercent. These areas are mostly found in the fringes of the city.
•In 1989, most of the areas in the watershed were impacted with 10 percent-25 percent areas
under impervious surface. No area was found in the degraded zone in 1989 image. The
occurrence of protected area is also very less.
•In 2008, asignificant proportion of the impacted area has transformed into the degraded
area with more than 25 percent of the impervious surface cover. In addition to this, certain
new areas. which have been developed in the past two decades, are in the degraded zone.
All the degraded areas are found in and around the Pune city and Pimpri-Chinchwad
indicating rapid constructional activities in these areas in the recent past. The areas under
the protected class are negligible in 2008.
Conclusion
It has been noticed that almost three times increase in the built up area has occurred in the
Pune and Pimpri-Chinchwad cities. Nowadays. impervious surface is considered as an indicator of
environmental fitness. In the Mula-Mutha Watershed, for the last two decades, alot of changes
have occurred as far as the impervious surface coverage is concerned. An appreciable proportion
of these built up units is found in the degraded areas with more than 25 percent built up surfaces
have been set up in those areas. Apart from these areas there are also some fringe areas which can
be pointed out as impacted areas. So, attention should be given to those sites also where new
constructions have been going on.
In the Mula-Mutha Watershed, due to the presence of Pune and Pimpri-Chinchwad cities, the
developmental activities cannot be stopped. But to nullify the effects of the hazardous incidents,
some immediate actions could be taken like designing constructions that reduce runoff and
imperviousness etc. Optimum use ofresources should be given importance. Sustainable development
should be encouraged which will give people achance to use the existing resources in the long
term basis l>\'lUlOllt obstructing the present use of the resources.
292 GEOGRAPIDCALREVIEW OF INDIA VOL.75
Acknowledgements
We thank Department of Geography, University of Pune for providing all the support required
during the study. We are also grateful to National Remote Sensing Centre (NRSC),Hyderabad for
providing the needed satellite imagery.
References
Arnold C.L. Jr, Gibbons C.J. (1996) :'Impervious Surface Coverage: the Emergence of aKey Environmental
Indicator'. Journal of the American Planning Association, Vol.-62, pp 243-258.
Dhorde, A.G. and WakhareA. S. (2008): 'Satellite-based Assessment of the Urban Heat Island of Pune City',
Major Research Project Report, Submitted to University of Pune, Pune
Groten, S. M. E. (1993) :NOVI -'Crop Monitoring and Early Yield Assessment of Burkina Faso'.
International Journal of Remote Sensing, Vol. 14, pp 225-242.
Prisloe, M., Giannotti, L., and W. Sleavin (2000) :'Determining Impervious Surfaces for Watershed Modeling
Applications'. Proceedings of the 8'k National Non-point Source Monitoring Workshop, Hartford.
pp 10-14.
Rouse, J. W. Jr., R.H. Haas, J.A. Schell and D. W. Deering. (1973) :'Monitoring Vegetation Systems in the
Great Plains with ERTS'. In: Proceedings of the 3"' ERTS Symposium, NASA SP-351 1:pp 309-
317.
Schueler T.R (1994) :'The Importance of Imperviousness'. Watershed Protection Techniques, I, pp 100-111.
Sleavin, W.J.(1999) :'Measuring Impervious Surface in Connecticut Using Planimetric GIS Data'. Thesis,
University of Connecticut, Storrs, Connecticut, pp 126.
United Nations (2002): 'World Urbanization Prospects: The 2001 Revision', New York, United Nations.
Wang, Q. Adiku, S. Tenhunen, J. Granier, A., (2005) :'On the Relationship ofNDVI with Leaf Area Index
in aDeciduous Forest Site'. Remote Sensing of Environment, 94 (2), pp.244-255.
