COMPARASION OF NDBI AND NDVI AS INDICATORS OF [620611]

COMPARASION OF NDBI AND NDVI AS INDICATORS OF
SURFACE URBAN HEAT ISLAND EFFECT IN LANDSAT 8
IMAGERY: A CASE STUDY OF IASI
Paul Macarof 1, Florian Statescu2

1PhD student: [anonimizat]., „Gheorghe Asachi” Technical University of Iasi, Faculty of
Hydrotechnical Eng ineering, Geodesy and Environmental Engineering, e -mail
[anonimizat],
2Professor Dr. Eng., Gheorghe Asachi”Technical University of Iasi, Faculty of
Hydrotechnical Engineering, Geodesy and Environmental Engineering, e -mail
[anonimizat]

Key words: NDBI, NDVI, Land surface temperature, Surface urban heat
island, Landsat -8

Abstract: This study compares NDVI (normaliz ed difference vegetation index)
and the normalized difference built -up index (NDBI) as indicators of surface urban heat
island (UHI) effects in Landsat -8 OLI imagery by investigating the relationships betwee n
the LST (land surface temperature) , NDBI and NDVI. The urban heat island (UHI)
represents the phenomenon of higher atmospheric and surface temperatures occurring in
urban area or metropolitan area than in the surrounding rural zone s due to urbanization.
With the development of remote sensing technology, it has become an important approach
to urban heat island research. Landsat data were used to estimate the LST, NDBI and N DVI
from four seasons for Iasi municipality area. This paper indicates than there is a strong
linear relationship between NDBI and LST, whereas the relationship between LST and
NDVI varies by season. This paper suggests, NDBI is an accurate indicator of su rface UHI
effects and can be used as a complementary metric to the traditionally applied NDVI.

1.Introduction

The urban heat island (UHI) represents the phenomenon of higher
atmospheric and surface temperatures occurring in urban area or metropolitan ar ea
than in the surrounding rural zones due to urbanization ( Oke & Voogt , 2005). UHI
is most noticeable during the summer and winter. The main cause of the UHI effect
is from the modification of land surfaces (Solecki et al., 2005 ). Waste heat
generated by energy usage is a secondary contributor (Li, Zhao, 2012). With the

rising of these urban or metropolitan environment issues, s cientists have paid more
attention on UHI research. Therefore, how to monitor, analyze, and evaluate the
urban heat island effect quantitatively has become one of the prime important
research among current urban climate and environment studies.
With the d evelopment of remote sensing technology, it has become an
important approach to urban heat island research. In earlier thermal remote sensing
studies, much weight has been placed on using NDVI as the major indicator of
urban climate. Results was indicated there is a linear relationship between Land
surface temperature and NDVI. According to Carlson, this negative correlation
between them is valuable for urban climate studies (Carlson et al., 1994). For all
that, NDVI measurements are topic to seasonal varia tions who may influence the
results of surface urban heat island analysis. Furthermore, the relationship between
NDVI and LST is well known to be nonlinear. Thus, NDVI alone may not be a
enought metric to study surface UHI quantitatively. According to stud ies the
relationship betw een percent impervious surface area (ISA) and LST, by using
Landsat TM and ETM+ d ata in an urbanized environment. Yuan et al. found that
percent ISA was an accurate indicator of surface UHI effects with strong linear
relationships between LST and percent ISA for all the four seasons (Yuan et al.,
2011). The normalized difference build -up index (NDBI) represents one of the
major land cover types, that is, build -up areas (Zha, Gao and Ni, 2005). Morever,
like NDVI, NDBI is very simple and easy to obtain, so it is feasible to use NDBI to
substitute for percent impervious surface area for study of surface urban heat
island. NDBI can be used as indicator of intensity of development and as indicator
of urban impervious surface.

2.Data an d methods

2.1. Study Area

Iași is the largest city in eastern Romania and the seat of Iași County.
Located in the historical region of Moldavia, Iași has traditionally been one of the
leading centres of Romanian social, cultural, academic and artistic li fe. The city is
positioned on the Bahlui River. This is an affluent of Jijia that flows into the Prut
River. This is one of the "legendary city of the seven hills", namely Bucium,
Cetățuia, Galata, Copou, Șorogari, Breazu and Repedea, just like so many cit ies
around world, one such example being Rome. The local climate is continental with
low rainfall and with large temperature differences between the seasons. Summer is
hot and it lasts from the end of the month of May up to the half of September.
Autumn is a short season, of transition. In the second half of November there is
usually frost and snow. Winter is a freezing season with temperatures dropping to
–20 șC (https://en.wikivoyage.org/wiki/Iași).

Study Area is geographically situated on latitude 47°1 2'N to 47°06'N and
longitude 27°32'E to 27°40'E.

Figure 1. Study Area (wikipedia, rotravel)

2.2. Landsat data

The main objective of this study was to compare the relationships of LST
to NDBI and NDVI using Landsat data obtained from four different seas ons using
Iasi metropolitan area as a case study.
Landsat 8 measures different ranges of frequencies along the
electromagnetic spectrum – a color, although not necessarily a color visible to the
human e ye. Each range is called a band. Landsat 8 has 11 ban ds. In the thermal
infrared ( TIR) are band 10 and 11, which see heat. According NASA, i nstead of
measuring the temperature of the air, like weather stations do, they report on the
ground itself, which is often much hotter (NASA Landsat Science).

Table 1 -Landsat data
Nr. c rt. Path Row Date
1 182 27 2016 -02-11
2 182 27 2016 -05-01
3 182 27 2016 -08-05
4 182 27 2016 -11-09

2.3. Data processing

The Normalized Difference Vegetation Index (NDVI) , according to Rouse
is a numerical indicator that uses the visi ble (Vis) and near -infrared (NIR) bands of
the electromagnetic spectrum, and is adopted to analyze remote sensing
measurements and assess whether the aim being observed contains live green
vegetation or not ( John Rouse, 1973). The NDVI algorithm subtracts the red (R)
reflectance values from the near -infrared (NIR) and divides it by the sum of them .

NDVI=
RED NIRRED NIR

NDBI in below equation is a normalized difference built -up index which is
used to extract built -up area. Beginning, this index, was used for TM da ta by Zha,
Gao and Ni, in 2003 .
NDBI=
NIR SWIRNIR SWIR

11

3. RESLUTS AND DISCUSSION

3.1. LST, NDBI and NDVI spatial patterns

LST (land surface temperature) is used to establish the temperature
distribution at the change global, regional and local scale. Also it’s used in climate
and acclimate change models in particular. LST, calculated from remote sensing
data is used in a lot of sphere of science, like: climate change, hydrology, forestry,
urban planning, oceanography , agriculture etc. Obtaining surface temperatures and
using them in different analysis is important to determine the problem associated
with the environment (Orhan et al. 2014).
The LST maps are presents in figure 1.
LST statistical data for the four dates are shown in Table 1.

Table1. Statistical data of LST for four seasons (oC)
Seasons Minimum Maximum Variations Mean Standard
Deviation
Winter 2.71 10.52 7.81 5.48 0.79
Spring 17.46 36.71 19.25 23.99 2.24
Summer 24.12 40.20 16.09 29.86 1.94
Autumn 5.52 14.51 8.99 9.89 0.81

The most important parameters are mean and variations. These two
parameters reflect change extent of the LST which spring and summer are much
higher than autumn and winter. This also indicates spring and summer SUHI
effects are more striking which au tumn and winter are not so obvious .

Figure 2. LST maps for four seasons (a -2016 -02-11; b -2016 -05-01;
c-2016 -08-08; d -2016 -11-09)

NDBI statistical data for the four dates are shown in Table 2. The
variations of NDBI are higher summer and spring than s maller winter and autumn.

Table 2. Statistical data of NDBI
Seasons Minimum Maximum Variations Mean Standard
Deviation
Winter -0.22 0.38 0.60 0.02 0.029
Spring -0.38 0.32 0.70 0.10 0.089
Summer -0.38 0.42 0.80 0.07 0.076
Autumn -0.23 0.41 0.64 0.01 0.026

The NDBI maps are presents in figure 2. The NDBI in bulit -up area of
summer and spring are obviously higher than winter and autumn.

Figure 3. NDBI maps for four dates (a -winter, b -spring, c -summer, d –
autumn)

As can be seen from Table 2 and Figure 3, the NDBI values for four
seasons are quite close, these indicate NDBI varies less with the season.
NDVI maps and statistical data for the four dates are shown in Fig. 4 and
Tab. 3.

Table 3. Statistical data of NDVI
Seasons Minimum Maximum Variations Mean Standard
Deviation
Winter -0.06 0.27 0.33 0.05 0.03
Spring -0.04 0.63 0.66 0.26 0.12
Summer -0.03 0.61 0.64 0.22 0.11
Autumn -0.02 0.38 0.40 0.09 0.05

Figure 4.NDVI maps for four dates (a -winter, b -spring, c -summer,
d-autumn)

The parameters: va riations, mean and standard deviations of NDVI for four
seasons showing that this index are higher spring and summer than autumn and
winter when all the vegetation was defoliated. The variations for spring and
summer are over 0.6 while for autumn and winte r are about 0.3 -0.4. The mean for
May-August is about 0.2 -0.3 while for November -February is 0.05 -0.1. These all
indicate NDVI has evident seasonal change which is exactly different to NDBI.

3.2. LST relationships to NDBI and NDVI

Figure 7 indicates cons istent linear patterns between LST and NDBI for all
the seasons, while in Figure 8 the relationship between LST -NDVI is non -linear
and strongly affected by season.
Research over the past two decades, according to Price, has shown that the
surface radiant t emperature response is determined by both surface vegetati on
cover and soil water content (Goetz, 1994; Price, 1990). The variation in the pixel
temperatures may be mostly explained to built -up site amounts and characteristics
since vegetated surfaces vary less in temperature than urban surfaces. Moreover,
NDBI is sensitive to the build -up area .

Irregular patterns with NDVI values are shown in the scatter plots for the
autumn and winter data. These results indicate that although NDVI may be used for
analysi s of surface UHI e ffects during spring and summer. So, NDBI is suitable for
LST studies for all the seasons (Gillies et al., 1997).

Fig. 5 Scatter plots of NDBI and LST for four dates

A linear relationship is shown between LST and NDBI for all the four
seasons (Fig.7), suggesting the fluctuation in LST can be accounted for very well
by NDBI for all seasons.
No e vident association between the surface UHI and the NDVI, which
further implies using NDVI to understand surface UHI is complicated since NDVI
itself suffers strong seasonal changes. The NDBI reflects seasonal surface UHI
fluctuations in similar linear patterns, which makes it a valuable l tool for analysis
and understanding of multi -temporal surface UHI effects. Figure 8 show
relationship of NDVI a nd mean LST for four dates.

Fig. 6. Scatter plots of NDVI and LST for four dates

Figure 7.Relationship of NDBI and LST for four dates

Figure 8.Relationship of NDVI and LST for four dates

4. Conclusions

This paper investigated the relationships between the LST, NDBI and
NDVI in Iasi. Results indicate there is a strong linear relationship between NDBI

and LST for all the four date, whereas the relationship between LST and NDVI
varies by season. The linear relationship between the LS T and NDBI suggests that
urban zone accounts for most of the fluctuation in land surface temperature
dynamics. Therefore, NDBI is an accurate indicator of surface UHI effects and can
be used as a complementary metric to the traditionally applied NDVI.

References

Carlson T., Gillies R., Perry M., "A method to make use of thermal infrared temperature
and NDVI measurements to infer surface soil water content and fractional
vegetation cover," 1994.
Chen X., Zhao H., Li P., Yin, Z. , Remote sensing image -based analysis of the relationship
between UHI and LU/LC changes, 2006.
Dihkan M., Karsli F., Guneroglu, A., Guneroglu N ., Evaluation of surface urban heat
island (SUHI) effect on coastal zone: The case of Istanbul Megacity.
Fabrizi R., Bonafoni, Biondi R ., Sat ellite and Ground -Based Sensors for the Urban Heat
Island (UHI) Analysis in the City of Rome, 2010.
Gillies R., Carlson T ., Cui J ., "A verification of the ‘triangle’ method for obtaining
surface soil water content and energy fluxes from remote measurements of the
normalized difference vegetation index, NDVI and surface temperature,"
International Journal of Remote Sensing, 1997.
Goetz S. J ., "Muti -sensor analysis of NDVI, surface temperature and biophysical variables
at a mixed grassland site", 1997.
Imhoff M. L., Zhang, P., Wolfe R. E., Bounoua L ., Remote sensing of the urban heat
island effect across biomes in the continental USA, 2010.
Kriegler F. J., Malila W. A., Nalepka R. F., Richardson W ., Preprocessing
transformations and their effects on multispect ral recognition. In Remote Sensing
of Environment, 1969.
Liu K., Su H., Zhang L., Yang H., Zhang R., Li X. , Analysis of the Urban Heat Island
Effect in Shijiazhuang, China Using Satellite and Airborne Data, 2015.
Ogashawara I., Bastos V. , Quantitative Appr oach for Analyzing the Relationship between
Urban Heat Islands and Land Cover, 2012.
Orhan O., Ekercin S., Dadaser -Celik F ., Use of Landsat Land Surface Temperature and
Vegetation Indices for Monitoring Drought in the Salt Lake Basin Area, 2014.
Price J. , "Using spatial context in satellite data to infer regional scale evapotranspiration,"
1990.
Purevdorj T. , Tateishi R., Ishiyama T ., "Relationships between percent vegetation cover
and vegetation indices," 1998.
Rouse J., Haas R., and Schell J. , Monitoring the vernal advancement and retrogradation
(greenwave effect) of natural vegetation, Texas A and M University, College
Station, 1974.
Streu tker D. , A remote sensing study of the UHI of Houston, Texas, 2002.
Voogt J. A. and Oke T. R ., "Thermal remote sensing of urban climates," 2003.
Weng Q., Lu D ., Schubring J ., "Estimation of land surface -vegetation abundance
relationship for Urban Heat Island studies," 2004.

Xian G.; Crane M ., An analysis of urban thermal characteristics and assoc iated land cover
in Tampa Bay and Las Vegas using Landsat satellite data, 2006.
Xiong Y., Huang, Chen F., Ye H., Wang , Zhu C. , The Im pacts of r apid urbanization o n
the thermal environment, 2012
Yuan F., Bauer E., "Comparison of impervious surface area (ISA ) and normalized
difference vegetation index (NDVI) as indicators of surface Urban Heat Island
(UHI) effects in Landsat Imagery, 2007.
Zha Y., Gao J., Ni S ., "Use of normalized difference built -up index in automatically
mapping urban areas from TM imag ery," 2003.
Zhao H., Chen X ., Use of normalized difference bareness index in quickly mapping bare
areas from TM/ETM, 2005.
* https://en.wikivoyage.org/wiki/Iași
**https://en.wikipedia.org/wiki/Iași