A Satellite Remote Sensing Based Land Surface Temperature Retrieval From Landsat Tm Data.

Transcription

1 Kogi State University, Anyigba From the SelectedWorks of Olarewaju Oluseyi Ifatimehin 2008 A Satellite Remote Sensing Based Land Surface Temperature Retrieval From Landsat Tm Data. Olarewaju Oluseyi Ifatimehin Available at:

2 A SATELLITE REMOTE SENSING BASED LAND SURFACE TEMPERATURE RETRIEVAL FROM LANDSAT TM DATA. Ifatimehin, O.O. 1 and Adeyemi, Segun Department of Geography and Planning, Kogi State University, Anyigba ABSTRACT Land Surface Temperature (LST) of Lokoja town were retrieved from LandSat TM imagery of The NDVI and land emissivity were estimated from visible and near infrared bands, while the effective satellite temperature was estimated from the thermal infrared band. The Qin et al s mono-window algorithm was employed to obtain the land surface temperatures of the different land use/cover types in Lokoja town. The result indicated that vacant land and built-up have the highest LST values of K and K, and vegetation, water bodies and cultivated land having the least with K, K and K respectively. The average retrieved LST for Lokoja town is K. Finally, a cross validation between the LST measured in situ and the retrieved LST shows a 9.60K difference. This method is very effective in the monitoring and studying of both environmental and atmospheric changes witnessed at all levels. Keywords: Lokoja, land use, landsat, surface temperature INTRODUCTION The launched of environmental monitoring satellites i.e. earth observation system (EOS) ushered in a new phase in climate and global change research. These satellites are designated to improve the understanding of the local and regional scale processes occurring on or near the earth s surface and lower atmosphere, including surface temperature interactions (Yamaguchi et al, 1998) and the accurate mapping of land cover parameters such as surface reflection (albedo) and land surface temperature (Wubet, 2003). Sensors onboard such satellites include LandSat TM, LandSat ETM. ASTER, RADASAT and NigeriaSat 1 among others. The clouds, land use, sea and land surface temperature, and exchanges of energy and moisture are processes that vary very rapidly in time and space and are considered to adequately explain the micro, meso and global climate change. Land surface temperatures (LSTs) are most important in global change studies: in estimating radiation budget, heat balance and as control for climate models (Wubet, 2003), and are strongly influenced by the ability of the surface to emit radiation, i.e. surface emissivity. However, the surface emissivity of land use/cover differs; it is relatively uniform for dense vegetated surfaces and highly variable for heterogeneous nature of some surface, such as soils and impervious surfaces (Ifatimehin, 2007 and Weng, 2001). Sobrino and Raissouni (2000), Valor and Caselles (1996), and Van de Griend and Owe (1993) suggested that land surface emissivity is best and easy obtained from Normalised Difference Vegetation Index (NDVI). The chlorophyll present in the leaf tissues reduces the reflectance of red light detected at the sensor as it absorbs red light, thereby suggesting that NDVI is a good indictor of surface radiant temperature (Lo et al, 1997; Wubet, 2003 and Ifatimehin, 2007). The NDVI is estimated using the reflectance of band 3 and band 4 of the imagery. Kidder and Wu (1987), Balling and Brazell (1988), Roth et al (1989), Gallo et al (1993), Weng (2001), Wubet (2003) and Ifatimehin (2007) had in their various works studied the surface temperature characteristics of urban areas using satellite remote sensing techniques on various imageries. These imageries each comes in different bands of wavelength: 3 bands for NigeriaSat 1, 7-8 bands for LandSat and 14 bands for ASTER and so it varies in other imageries. But the thermal 1 Corresponding Author: lanreifa@yahoo.com Ethiopian Journal of Environmental Studies and Management. Vol. 1, No 3 (2008): ISSN:

3 Infrared band (band 6) is the choice band for estimating surface temperature because radiation emitted from the earth surfaces in the form of heat ranges between 3.0 to 100µm (Schmugge et al, 1998) and as indicated in previous studies (Carnahan and Larson, 1990; Nicol, 1994; Weng, 2001 and Ifatimehin, 2007). Remote sensing is becoming a tool to reckon with in the evaluation and monitoring of environmental and ecological processes. Landsat TM is one of the most used data for environmental studies. It is composed of seven bands, of which six of them are in the visible and near and short infrared region and only one band located in the thermal infrared band as shown in Table 1. Table 1: LandSat TM 1987 spectral consideration BAND WAVELENGTH (m) SPECTRAL LOCATION/REGION Blue Green Visible Red Near-Infra-Red Short-Wave-Infra-Red Spatial Resolution (meter) Thermal-Infra-Red 120 Source: Based on the landsat TM 1987 image characteristics. Sobrino et al (2004) summaries the various uses of the Landsat TM bands as follows: Band 1 is used for coastal water studies; Band 2 for crops identification and vegetation stage studies, Band 3 and Band 4 are used to calculate vegetation indexes; Band 5 and Band 7 are used for clouds, ice, snow and geological formations discrimination, and finally Band 6 is used for Land Surface Temperature (LST) retrieval. The objectives of this paper are to: i. Classify Lokoja town into its various land use/cover types; ii. estimate land surface emissivities (LSE) of the different land use/cover of Lokoja town; iii. retrieve Land Surface temperatures of these different land use/cover types. STUDY AREA Lokoja town, the administrative capital of Kogi State lies between 7 o o N and 6 o o E within the lower Niger trough. It has an estimated landmass of sq. km (Fig 1)

4 It is situated in the Guinea savanna belt witnessing the Aw type of climate. Annual Rainfall is between 1016mm and 1524 with its mean annual temperature not falling below 27 o 7 C. The town is sandwiched to the West and East by the Patti ridge and River Niger respectively thereby exhibiting heterogeneity in its topography (altitude) and geology (Basement complex and sedimentary formation). METHODOLOGY Data Used The data used in this study is the Landsat TM image of 10 th December The spectral characteristics are shown in Table 1. Method for land use/cover Classification Idrisi image processing software was used to enhanced and rectified the image to a common UTM coordinate system (WGS84), and then radiometrically corrected. A supervised classification with a maximum likelihood algorithm was conducted to classify the image using three bands of green (2), red (3) and near-infrared (4) (as indicated in Table 1). Training sample sets were collected based on ground truth data gathered during field checks. Following completion it was run on mosaic. The methodology followed is shown in Figure 2. Method in deriving surface temperature. The radiometrically corrected LandSat TM thermal infrared data (band 6) was used for this purpose. The following methods were adopted: i. Digital Number (DN) conversion to radiance: ii. Conversion from radiance to reflectance (Surface albedo): Where t is transmisivity = T o where T o is the near surface temperature. While rp is the broadband reflectance= Where ESUN=mean solar exo atmospheric irradiance rp (λ 1 ) is the planetary reflectance= 65

5 iii. L λ =spectral radiance at the sensor apecture d=earth sun distance CosQ=solar Zenith angle t=one way atmospheric transmittance The NDVI image was computed for 2001 from the band 3 and band 4 reflectance data using the formula below: iv. Emissivity, εo= ( x ln(ndvi)) v. T a = T o, T a is the mean atmospheric temperature vi. Effective satellite temperature T s : vii. The Qin et al s mono-wnidow algorithm developed in 2001 was use to obtain the land surface temperature (T): Where λ = wavelength of emitted radiance = 11.5µm (Markam and Barker, 1985), α=hc/k (1.438 x 10-2 mk), k=stefan Boltzmann s constant (1.38 x JK -1 ), h=planck s constant (6.26x10-34 Js), and c=velocity of light (2.998x10 8 s -1 ), a = and Figure 3 shows the flowchart for the computation of surface temperature (T s ) 66

6 RESULTS AND DISCUSSION The classified LandSat TM imagery of 1987 shows the different classes of land use/cover types of Lokoja town. Figure 2 and Table 2 shows that natural vegetation dominants the entire landscape with per cent, followed by cultivate land with per cent, while the built up area of Lokoja is 1.44 per cent. Table 2: Lokoja town under different land use/cover 1987 Land use type 1987 Area (km 2 ) % Vacant Land Built-up Area Cultivate Land Natural Vegetation Water body Total Source: Based on classified LandSat TM 1987 Imagery Table 3: LandSat ETM spectral radiance ranges (Wm -3 sr -4 µm -4 ) Band number 17 November 2001 Low Gain High Gain Lmin Lmax Lmin Lmax Band Band Band Band Source: Authors computation (2007) Table 4: LandSat ETM spectral Irradiance 67

7 Bands ESUN (Wm -2 µm -1 ) Band Band Band Source: Authors computation (2007) Table 5: LandSat ETM 2001 Thermal band calibration constants Bands K 1 (Wm -2 sr -1 µm -1 ) K 2 (K) Band NDVI, Land Surface Emissivity and satellite temperature were calculated based on the indices in Table 3, 4 and 5. The resulting emissivities and land surface temperatures are shown in Table 6 and in Figure 3 and 4. The LSE ranges from 0.92 to with an average of The highest emissivity is shown were vegetation is very thick and the lowest recorded for vacant land. (0.951) while the LST for each land use/cover recorded vacant land and built-up to have the highest surface temperatures of K and K respectively. It implies that these land use/cover types are devoid evaporating and transpiring objects. While vegetation, water body and cultivated land are having the least surface temperature of K, K and respectively. Table 6: LandSat TM 1987 derived LSE and LST for different land use/cover type Land use/cover LSE LST (K) Min Max Avg Min Max Avg Vacant Land Built-up Area Cultivated land Vegetation Water bodies Total LST 1, Average LST Source: Authors computation,

8 The retrieved land surface temperature from the LandSat TM image of 10 th December 1987 was cross validated with the insitu surface temperature recorded in Lokoja on that day by the Nigeria Meteorological (NIMET) station. The difference of 9.60 K (Table 7) results, indicating that the satellite retrieved land surface temperature is greater than the insitu surface temperature. This difference can be associated partly with the effects of surface roughness on surface temperature and emissivity which was not taken note off (Cassels et al, 1992 and Weng, 2001) and partly because of atmospheric impurities which had obstruct the smooth passage of radiance energy implying the atmospheric correction was not done to the image during processing (Okeke, 2007). Table 7: Cross validation of LST with NIMET insitu data Parameter Unit Satellite Estimation NIMET Data Difference LandSat TM 1987 Surface Temperature K Source: Authors Computation CONCLUSION The study revealed the retrieval of Land Surface Temperature (LST) from the Land Surface Emissivity (LSE) values obtained from the Normalised Difference Vegetation Index (NDVI). It also revealed the differences in the LSE and LST among the five classes of land use/cover in Lokoja. The implication of these differences in the land use/cover LSTs, is that with increase urban expansion in the built-up area, vegetation and cultivated land are likely to decline and water body may silt. All these may likely tamper with the ecosystem services and may result in a phenomenon called the Urban Heat Island. The K LST retrieved and the 9.60K difference recorded from the cross validation shows that satellite imageries when adequately processed, will be more effective in the monitoring of environmental parameters and use in the better understanding of environmental issues at both local and regional level and also global change studies. REFERENCES Balling, R.C. and Brazell, S.W. (1988): High resolution surface temperature patterns in a complex urban terrain; Photogrammetric Engineering and Remote Sensing, Vol, 54, Pg Carnahan, W.H. and Larson, R.C. (1990): An analysis of an urban heat sink; Remote Sensing of Environment, Vol. 33, Pg Gallo, K.P.; McNab, A.L.; Karl, T.R.; Brown, J.F.; Hood, J.J and Tarpley, J.D. (1993): The use of NOAA AVHRR data for assessment of the urban heat Island effect, Journal of Applied Meteorology, Vol.32, Pg Ifatimehin, O.O. (2007) Estimating Surface temperature of Lokoja town using Geoinformatic Technology ; International Journal of Ecology and Environmental Dynamics, Vol. 4 (in press) Kidder, S.Q. and Wu, H.T.(1987): A multispectral study of the St.Louis area under snow covered conditions using NOAA-7 AVHRR data; Remote Sensing of Environment, Vol.22, Pg Lo, C.P; Quattrochi, D.A. and Luvah, J.C. (1997): Application of high-resolution thermal infrared remote sensing and GIS to assess the urban heat island effect; International Journal of Remote Sensing, Vol.18, Pg Nichol, J.E. (1994): A GIS based approach to microclimate monitoring in Singapore s high rise housing estates; Photogrammetric Engineering and Remote Sensing, Vol.601, Pg Qin, Z.; Karnieli, A and Berliner, P. (2001): A mono-window algorithm for retrieving land surface temperature from landsat TM data and its application to the Isreal-Egypt border region, International Journal of remote Sensing, Vol.22(18), pg Roth, M; Oke, T.R. and Emery, W.J. (1989): Satellite derived urban heat islands from three coastal cities and the utilization of such data in urban climatology; International Journal of Remote Sensing, Vol.10, Pg Schmugge, T; Hook, S.J and Coll, C. (1998): Recovering surface temperature and emissivity from thermal infrared multispectral data; Remote Sensing of Environment, Vol.65, Pg Sobrino, J.A.; Jimenez-Munoz and Paolini, L. (2004): Land Surface temperature retrieval from LandSat TM 5, Remote Sensing of Environment, Vol.90, Pg

9 Sobrino, J.A. and Raissouni, N. (2000): Toward remote sensing methods for land cover dynamic monitoring. Application to Morocco, International Journal of Remote sensing, Vol.21, pg Valor, E. and Caselles, V. (1996): Mapping land surface emissivity from NDVI: application to European, African and South American areas, Remote Sensing of Environment, Vol.57, pg Van de Griend, A.A. and Owe, M. (1993): On the relationship between thermal emissivity and the normalized difference vegetation index for natural surfaces, International Journal of remote Sensing, Vol.14(6), pg Weng, Q (2001): A remote Sensing-GIS evaluation of urban expansion and its impacts on surface temperature in the Zhujiang, Delta, China; International Journal of Remote Sensing, Vol.22(10), Pg Wubet, M.T. (2003): Estimation of absolute surface temperature by satellite remote sensing, Unpublished M.Sc Thesis, International Institute for Geoinformation Science and Earth Observation, Netherlands. Yamaguchi, Y., Kahle, A.B., Tsu, H., Kawakami, T. and Pniel, M.: (1998): Overview of Advanced Spaceborne Thermal Emssion and reflection Radiometer (ASTER), IEEE Transactions on geosciences and remote sensing, Vol.36(4), Pg