Modelling and remote sensing of land surface
- 1. 1 23 Journal of the Indian Society of Remote Sensing ISSN 0255-660X Volume 40 Number 3 J Indian Soc Remote Sens (2012) 40:399-409 DOI 10.1007/s12524-011-0158-3 Modelling and Remote Sensing of Land Surface Temperature in Turkey Mehmet Şahin, B. Yiğit Yıldız, Ozan Şenkal & Vedat Peştemalcı
- 2. 1 23 Your article is protected by copyright and all rights are held exclusively by Indian Society of Remote Sensing. This e-offprint is for personal use only and shall not be self-archived in electronic repositories. If you wish to self- archive your work, please use the accepted author’s version for posting to your own website or your institution’s repository. You may further deposit the accepted author’s version on a funder’s repository at a funder’s request, provided it is not made publicly available until 12 months after publication.
- 3. RESEARCH ARTICLE Modelling and Remote Sensing of Land Surface Temperature in Turkey Mehmet Şahin & B. Yiğit Yıldız & Ozan Şenkal & Vedat Peştemalcı Received: 28 September 2010 /Accepted: 15 August 2011 /Published online: 14 September 2011 # Indian Society of Remote Sensing 2011 Abstract This study introduces artificial neural net- works (ANNs) for the estimation of land surface temperature (LST) using meteorological and geo- graphical data in Turkey (26–45°E and 36–42°N). A generalized regression neural network (GRNN) was used in the network. In order to train the neural network, meteorological and geographical data for the period from January 2002 to December 2002 for 10 stations (Adana, Afyon, Ankara, Eskişehir, İstanbul, İzmir, Konya, Malatya, Rize, Sivas) spread over Turkey were used as training (six stations) and testing (four stations) data. Latitude, longitude, elevation and mean air temperature are used in the input layer of the network. Land surface temperature is the output. However, land surface temperature has been estimated as monthly mean by using NOAA-AVHRR satellite data in the thermal range over 10 stations in Turkey. The RMSE between the estimated and ground values for monthly mean with ANN temperature(LSTANN) and Becker and Li temperature(LSTB-L) method values have been found as 0.077 K and 0.091 K (training stations), 0.045 K and 0.003 K (testing stations), respectively. Keywords Generalized regression neural network . Land surface temperature . Satellite data Introduction Land surface temperature (LST) is an important factor controlling most physical, chemical, and biological processes on Earth. Knowledge of land surface temper- ature is necessary for many environmental studies and management activities of the Earth’s resources (Li and Becker 1993). In order to monitor macro-scale spatial changes in surface temperature, scanners designed for sensing in the thermal bands are placed onboard platforms for remote sensing of the Earth’s resources from space (Sabins 1997). The extensive application and significant importance of temperature in environ- mental studies and management is the main force driving the study of LST in remote sensing. With the availability of thermal sensing data, such as channels 4 and 5 of Advanced Very High Resolution Radiometer (AVHRR) data as well as Landsat Thematic Mapper 6 (TM6), the study of LST has become one of the hottest topics in remote sensing during the last two decades (Vogt 1996). Thus, two approaches have been devel- J Indian Soc Remote Sens (September 2012) 40(3):399–409 DOI 10.1007/s12524-011-0158-3 M. Şahin (*) Siirt Vocational School, Siirt University, 56100 Siirt, Turkey e-mail: sahanmehmet2000@yahoo.com B. Y. Yıldız Karaisalı Vocational School, Çukurova University, 01770 Karaisalı, Adana, Turkey O. Şenkal Faculty of Education Department of Computer Education and Instructional Technology, Çukurova University, 01330 Sarıçam, Adana, Turkey V. Peştemalcı Physics Department, Çukurova University, 01330 Sarıçam, Adana, Turkey Author's personal copy
- 4. oped to recover land surface temperature from multi- spectral thermal infrared (TIR) imagery (Schmugge et al. 1998). The first approach utilizes a radiative transfer equation to correct the at-sensor radiance to surface radiance, followed by an emissivity model to separate the surface radiance into temperature and emissivity (Schmugge et al. 1998). The second approach applies the split-window technique for sea surfaces to land surfaces, assuming that the emissivity in the channels used for the split window is similar (Dash et al. 2002). Land surface brightness temperatures are then calcu- lated as a linear combination of the two channels. A study was undertaken to retrieve land (soil– vegetation complex) surface temperature (LST) over a area in India using thermal bands (channel 4 and 5) of AVHRR. The LST values were compared with near synchronous soil and air temperature measurements. The estimated LST of a semi-arid mixed agricultural barren values were near midway between air temper- ature (AT) and soil surface temperature (ST) with mean bias of −2.9 K and 7.0 K respectively in winter. However, in the summer, the LST values were found to be closer to ST, which may be due to lower fractional vegetation cover and NDVI (Bhattacharya and Dadhwal 2003). So this study showed that there is a relationship between land surface temperature, soil temperature and air temperature. Although land surface temperature (LST) can be generally estimated with reasonable accuracy by using satellite data, it is not possible to get same successful in cloudy days. So, a new method is necessary to retrieve land surface temperature. For this purpose, Artificial Neural Networks (ANN) was applied as a new method which used latitude, longitude, elevation and monthly mean air temperature. Artificial Neural Networks (ANN) have been applied to many different environmental sectors, especially in meteorological forecast (Gardner and Dorling 1998; Aires et al. 2002). Govindaraju (2000a, b) reports a number of studies which have used ANNs to forecast rainfall over a short time interval. It is clear from many studies that usage of ANN method is suitable and applicable for estimating global solar radiation especially for regions where very large distances exist between meteorological stations and also having abundant solar energy (Kalogirou 2001; Bechrakis and Sparis 2004). The objective of the present study is to apply the LSTANN, LSTB-L methods in collaboration with each other for the prediction of the land surface tempera- ture of the target station using neighboring measuring stations. This will show that these methods can be applied to predict the land surface temperature for any location around sampled measuring stations. The studies which have used LSTANN, LSTB-L method show a general perspective of land surface temperature in Turkey. Methodology and Data Sources In this study, the ANN is used to model the LST from near-surface air temperature and geographic information. In addition, the LSTB-L method is used to retrieve the LST from National Oceanic and Atmospheric Admin- istration/Advanced Very High Resolution Radiometer (NOAA/AVHRR) data. The LSTB-L method is pre- sented for the determination of monthly global land surface temperature from the NOAA/AVHRR satellite data, which provide wide coverage together with adequate spatial resolution (around 1.1 km at the nadir). For application, ten stations (Adana, Afyon, Ankara, Eskişehir, İstanbul, İzmir, Konya, Malatya, Rize, and Sivas) have been selected from different regions of Turkey. The geographical locations of these land surface temperature stations are shown in Fig. 1. The stations selected can give a general idea about the land surface temperature values in Turkey. The estimation of land surface temperature in Turkey was based on meteorological and geographical data (latitude, longitude, elevation, and monthly mean air temperature). These selected locations in Turkey have different values, as seen in Table 1. Generalized regression neural network (GRNN) is used in the ANN. Meteorological and geographical data are used as input, and land surface temperature is the output. Fig. 1 Land surface temperature measuring stations in Turkey 400 J Indian Soc Remote Sens (September 2012) 40(3):399–409 Author's personal copy
- 5. Split-Window Method The temporal resolution of NOAA-12-14-15 images is at the same location which is scanned by the satellite 5 cloudless days of every month. The satellite images were received in a raw image and zipped data format. They were unzipped and processed with the software ‘Quorum to Level1B’ in order to convert this raw image data to Level1B format, so that remote sensing software can be applied for processing. Radiometric and geometric calibrations were applied first to the images to correct the deficiencies and flaws that could result from the imaging sensor in the platform (satellite). Then, land surface temperature is predicted with LSTB-L method. Usage of these methods is more suitable for large places. Both methods are applica- ble in any region for the very large distances between the stations. Split-window algorithms for the retrieval of LST from NOAA-AVHRR data have been proposed by different scholars in the last decades (Vogt 1996). The algorithm given by Becker and Li (1990) is worthy of detailed description since it has been used in many studies (Franca and Cracknell 1994) on the study of LST. The algorithm requires – brightness temperatures in channels 4 (T4) and 5 (T5) – the mean emissivity in these channels, ε=(ε 4+ ε 5)/ 2=0.975 – the spectral emissivity difference, Δε=ε 4 -ε 5= −0.005 (Caselles et al. 1997; Chrysoulakis et al. 2001). Becker and Li (1990) presented a local split- window algorithm for viewing angles of up to 46° from nadir, given as follows: TBec ker ÀLiÀ1990 ¼ A0 þ P T4 þ T5 2 þ M T4 À T5 2 ð1Þ where A0, P, and M are coefficients influenced by a number of factors in the process of radiance trans- mission from the ground to the sensor. For NOAA/ AVHRR data, coefficient A0=1.274 and A2=0.15616, A3=−0.482, B1=6.26, B2=3.98, B3=38.33. Later, Li and Becker (1993) modified their algorithm into a general one keeping the form of Eq. 1. The only difference in the modified algorithm is that the coefficients are determined in terms of water content calculated from a radiance simulation using the LOW- TRAN 7 program (Caselles et al. 1997). Based on Becker and Li (1990), a generalized split-window algorithm has been developed by Wan and Dozier (1996). The form of the algorithm is the same as Eq. 2, but the coefficients P and M are given as follows: P ¼ A1 þ A2 1 À " " þ A3 $" "2 ð2Þ M ¼ B1 þ B2 1 À " " þ B3 $" "2 ð3Þ where A1~A3 and B1~B3 are parameters estimated by the method given by Li and Becker (1993). The difference is that Wan and Dozier (1996) defined A1 in their model as a variable while Becker and Li (1990) defined it as a constant equal to 1. Similarly, the parameters are determined by regression analysis of the brightness temperature data to be simulated ground data of standard atmosphere given by LOW- Stations Latitude (°N) Longitude (°E) Elevation (m) The mean air temperature (K) The mean land surface temperature (K) Adana 36.59 35.21 27 290.31 291.07 Afyon 38.44 30.35 1001 280.84 282.10 Ankara 39.57 32.53 891 283.60 283.08 Eskişehir 39.47 30.34 786 278.79 282.30 İstanbul 41.01 28.59 0 287.33 287.10 İzmir 38.26 27.10 29 289.83 289.60 Konya 37.58 32.33 1031 282.13 283.57 Malatya 38.21 38.19 948 283.08 284.43 Rize 41.02 40.31 9 286.06 287.45 Sivas 39.45 37.01 1600 277.69 280.53 Table 1 Geographical and meteorological parameters for the stations J Indian Soc Remote Sens (September 2012) 40(3):399–409 401 Author's personal copy
- 6. TRAN 7 program, based on many assumptions on such variables as atmospheric state and viewing angles. (Qin and Karnieli 1999). The mean land surface temperature for ten stations over Turkey has been determined using the LSTB-L method. Artificial Neural Networks Artificial neural networks (ANNs) are information processing systems that are non-algorithmic, non digital, and intensely parallel (Dinçer et al. 1996). The use of ANNs for modeling and prediction purposes has become increasingly popular in recent decades (Çam et al. 2005). Researchers have been applying the ANN method successfully in various fields such as mathe- matics, engineering, medicine, economics, meteorology, psychology and neurology as well as in the prediction of mineral exploration sites, in electrical and thermal load predictions, in adaptive and robotic control. ANNs have been trained to overcome the limitations of conventional approaches to solving complex problems. This method learns from given examples by construct- ing an input–output mapping in order to perform predictions (Kalogirou 2000). Generalized Regression Neural Network (GRNN) proposed by Speckt (1991) does not require an iterative training procedure as in back propagation method. It approximates any arbitrary function between input and output vectors, drawing the function estimate directly from the training data. Furthermore, it is consistent; that is, as the training set size becomes large, the estimation error approaches zero, with only mild restrictions on the function. The GRNN is used for estimation of continuous variables, as in standard regression techniques. It is related to the radial basis function network and is based on a standard statistical technique called kernel regression. By definition, the regression of a dependent variable y on an independent x estimates the most probable value for y, given x and a training set. The regression method will produce the estimated value of y which minimizes the mean-squared error. GRNN is a method for estimating the joint probability density function (pdf) of x and y, given only a training set. Because the pdf is derived from the data with no preconceptions about its form, the system is perfectly general. If f(x,y) represents the known joint continuous probability density function of a vector random variable, x, and a scalar random variable, y, the conditional mean of y given X (also called the regression of y on X) is given by: E y Xj½ Š ¼ R1 À1 yf X; yð Þdy R1 À1 f X; yð Þdy ð4Þ When the density f(x,y) is not known, it must usually be estimated from a sample of observations of x and y. The Probability estimator f(X,Y) is based upon sample values X and Y of the random variables x and y, where n is the number of sample observations and p is the dimension of the vector variable x: f X; Yð Þ ¼ 1 2pð Þðpþ1Þ=2 sðpþ1Þ 1 n  Xn i¼1 exp ÀðX À Xi Þ T X À Xi ð Þ 2s2 " # exp ÀðY À Yi Þ 2 2s2 " # ð5Þ Land surface temperature Pattern layer Output layerInput layer Latitude Longitude Elevation Air temperature Summation layerFig. 2 Topology of a generalized regression neural network (GRNN) artificial neural networks (ANN) 402 J Indian Soc Remote Sens (September 2012) 40(3):399–409 Author's personal copy
- 7. A physical interpretation of the probability esti- mate f(X, Y) is that it assigns sample probability of width σ for each sample Xi and Yi , and the probability estimate is the sum of those sample probabilities (Speckt 1991). Defining the scalar function Di 2 D2 i ¼ X À Xi À ÁT X À Xi À Á ð6Þ and performing the indicated integrations yields the following: Y Xð Þ ¼ Pn i¼1 Yi exp ÀD 2 i 2s2 Pn i¼1 exp ÀD 2 i 2s2 ð7Þ The resulting Eq. 7 is directly applicable to problems involving numerical data. When the smoothing parameter σ is made large, the estimated density is forced to be smooth and in the limit becomes a multivariate Gaussian with covariance σ2 I. On the other hand, a smaller value of σ allows the estimated density to assume non-Gaussian shapes, but with the hazard that wild points may have too great an effect on the estimate (Speckt 1991). The GRNN consists of four layers: input layer, pattern layer, summation layer and output layer. The input units are in the first layer. The second layer has the pattern units and the outputs of this layer are passed on to the summation units in the third layer. The final layer covers the output units (Cigizoğlu and Alp 2006). Mohandes et al. (2004) state that during the training procedure, the weights of the connections between neurons are adjusted in order to achieve the desired input/output relation of the network. This procedure goes on until the difference between the actual output of the network and the desired output is equal with a specified remainder value. Here, the criterion is put forward as the network output which should be closer to the value of desired output. This training procedure has to be repeated for the rest of the input–output pairs existing in the training data. Input variables have been used to validate the ANN. Results and Discussions ANN is used for modeling land surface temperature in Turkey. MATLAB software has been used to train the LSTANN on a personal computer. A GRNN structure GRNN (4, 1.0, 1) corresponds to 4 input nodes, a spread value equal to 1.0 and a single output node. The selected LSTANN structure is shown in Fig. 2. This network consists of input layer, pattern layer, summation layer and output layer. The stations used Fig. 3 Comparison between measured and estimated monthly mean land surface temperature using LSTB-L method in training stations Fig. 4 Comparison between measured and estimated monthly mean land surface temperature using LSTB-L method in testing stations Fig. 5 Comparison between measured and estimated monthly mean land surface temperature using LSTANN method in training stations J Indian Soc Remote Sens (September 2012) 40(3):399–409 403 Author's personal copy
- 8. for training data are at Adana, Afyon, Ankara, Eskişehir, İzmir, Sivas. In test data, stations, İstanbul, Konya, Malatya and Rize are used. In order to train the neural network, meteorological and geographical data measured by the Turkish State Meteorological Service for the period from January 2002 to December 2002 in Turkey from the above ten stations were used as training and testing data. Inputs for the network are latitude, longitude, elevation and monthly mean air temperature; output is land surface temperature. In addition, the monthly mean land surface temperature data were measured with land thermometer from Turkish State Meteorological Service. The LSTANN structure and weights of neurons could help to understand the relationship between air and surface temperature. Separately, the method of LSTB-L was proposed for the estimation of monthly global land surface temper- ature values from meteorological satellite data. Land surface temperature (LST) was estimated as a monthly mean by using the LSTB-L method over ten stations in Turkey. It is suggested in this study that both methods describing the LST throughout a diurnal cycle very well. Data of Turkish State Meteorological Service are deficient throughout a nocturnal cycle than a diurnal cycle in Turkey. In the current study, ground data of the ten national stations were used. These stations were selected in such a way as to represent the widely changing climatic conditions of Turkey. The monthly mean global land surface temperature over Turkey was determined to be a correlation coefficient 94.08% and RMSE 0.091 K (Fig. 3), 91.13% and RMSE 0.003 K (Fig. 4) for LSTB-L values (training and testing stations). In the case of monthly mean correlation coefficient, RMSE was found to be 98.06% and 0.077 K (Fig. 5), 98.01% and 0.045 K (Fig. 6) for LSTANN values (training and testing stations). Estimation of LSTB-L method values and LSTANN method values (training, testing) were evaluated accord- ing to statistical rules by using root mean square error (RMSE) and mean bias error (MBE). The results of evaluation are given in Table 2. The RMSE values, ranging from 0.014 K to 9.318 K, differ from the actual value for all stations. The maximum RMSE was found to be 9.318 K for the Rize station (testing) in the LSTB- L method values, while the best result was found to be 0.014 K for the Afyon station (training) in the LSTANN values. The maximum MBE was found to be −2.690 for LSTB-L method values of Rize station, while the minimum MBE was found as 0.004 for Afyon and Stations (cities) LSTANN LSTB-L RMSE (K) MBE (K) RMSE (K) MBE (K) Stations used in training Adana 0.0346 0.0100 0.2252 −0.0650 Afyon 0.0144 0.0042 1.2038 −0.3475 Ankara 0.0318 0.0092 2.4104 0.6958 Eskişehir 0.0433 −0.0958 0.7130 0.2058 İzmir 0.0144 0.0042 2.8204 −0.8142 Sivas 0.0491 0.0142 0.9873 0.2850 Stations used in testing İstanbul 0.1126 −0.0325 3.7470 1.0817 Konya 0.0895 −0.0258 5.1326 1.4817 Malatya 0.0837 0.0242 1.3221 0.3817 Rize 0.0289 0.0083 9.3184 −2.6900 Table 2 Error values of the LSTANN and LSTB-L method approach Fig. 6 Comparison between measured and estimated monthly mean land surface temperature using LSTANN method in testing stations 404 J Indian Soc Remote Sens (September 2012) 40(3):399–409 Author's personal copy
- 9. İzmir stations. Moreover, another significant point in this table, the performance values of the training are generally better than the performance values of the testing. Figures 7 and 8 shows a comparison between measured, LSTANN and LSTB-L values for the ten stations (training and testing stations). In literature, lots of algorithms have been pro- pounded by Price(1984), Becker(1987), Vidal (1991), Prata(1994, 1993), Sobrino et al. (1996, 1994), Coll et al.(1994), Becker and Li(1995), Coll and Caselles (1997). These studies indicated that it is possible to retrieve LST at a reasonable accuracy (RMSE of 1– 3 K) from current operational and research satellite- borne visible/infrared radiometers. This study has showed that LSTANN and LSTB-L methods are more successful than lots of algorithms which were used in literature to get land surface temperature. Using the LSTANN and LSTB-L methods is a cheap and effective way to estimate monthly global land surface temperature and construct a land surface Adana 260 270 280 290 300 310 41 2 3 5 6 7 8 9 10 11 12 Months 41 2 3 5 6 7 8 9 10 11 12 Months 41 2 3 5 6 7 8 9 10 11 12 Months MeanLandSurfaceTemperature(K) 260 250 270 280 290 300 310 MeanLandSurfaceTemperature(K) 260 250 270 280 290 300 310 MeanLandSurfaceTemperature(K) Measured Backer and L ANN Afyon Ankara Fig. 7 Comparison for the land surface temperature between the LSTANN, LSTB-L and measured values (training stations) J Indian Soc Remote Sens (September 2012) 40(3):399–409 405 Author's personal copy
- 10. temperature database. The LSTANN model, which needs no satellite data, was used to estimate the monthly mean at ground level. On the other hand, since the other method needs no ground data and presents valuable results, it can be applied to any region. The application of these methods is suitable particularly for places where the distances between the stations are very large. Conclusion A generalized regression neural network (GRNN) was used to estimate monthly land surface temperature by using meteorological and geographical data. The generation of a typical land surface temperature is significant for the calculations concerning many land surface temperature methods. By using the LSTANN and LSTB-L methods, a production of land surface temperature was used at over ten stations in Turkey. The monthly mean values were found as 0.077 K and 0.091 K (training stations), 0.045 K and 0.003 K (testing stations), respectively. According to the results of these ten locations, correlation values indicate a relatively good agreement between the observed LSTANN values and the predicted satellite values. So, LSTANN and LSTB-L methods are sug- Eski ehir 250 260 270 280 290 300 MeanLandSurfaceTemperature(K) zmir 250 260 270 280 290 300 310 MeanLandSurface Temperature(K) Sivas 240 250 260 270 280 290 300 MeanLandSurface Temperature(K) 41 2 3 5 6 7 8 9 10 11 12 Months 41 2 3 5 6 7 8 9 10 11 12 Months 41 2 3 5 6 7 8 9 10 11 12 Months Fig. 7 (continued) 406 J Indian Soc Remote Sens (September 2012) 40(3):399–409 Author's personal copy
- 11. gested to researchers who study on topics which are related to land surface processes, combining surface- atmosphere interactions, energy fluxes between the atmosphere and the ground, estimation of radiation budgets in heat balance studies, climatology, human- environment interactions, hydrology systems, ecolo- gy, greenhouse effect, biogeology, agricultural appli- cations such as evaluating water requirements for wheat and determining frost areas for citruses in the future. stanbul 260 270 280 290 300 310 1 2 3 4 5 6 7 8 9 10 11 12 Months MeanLandSurface Temperature(K) Measured Backer and Li ANN Konya 250 260 270 280 290 300 310 1 2 3 4 5 6 7 8 9 10 11 12 Months 1 2 3 4 5 6 7 8 9 10 11 12 Months 1 2 3 4 5 6 7 8 9 10 11 12 Months MeanLandSurface Temperature(K) Malatya 250 260 270 280 290 300 310 MeanLandSurface Temperature(K) 250 260 270 280 290 300 310 MeanLandSurface Temperature(K) Rize Fig. 8 Comparison for the land surface temperature between the LSTANN, LSTB-L and measured values (testing stations) J Indian Soc Remote Sens (September 2012) 40(3):399–409 407 Author's personal copy
- 12. References Aires, F., Chédin, A., Scott, N. A., Rossow, W. B. (2002). A regularized neural net approach for retrieval of atmospheric and surface temperature with the IASI instrument. Journal of Applied Meteorology, 41, 144–159. Bechrakis, D. A., Sparis, P. D. (2004). Correlation of wind speed between neighboring measuring stations. IEEE Transactions on Energy Conversion, 19, 400–406. Becker, F. (1987). The impact of spectral emissivity on the measurement of land surface temperature from a satellite. International Journal of Remote Sensing, 11, 369–394. Becker, F., Li, Z. L. (1990). Towards a local split window method over land surface. International Journal of Remote Sensing, 11, 369–393. Becker, F., Li, Z. L. (1995). Surface temperature and emissivity at various scales: Definition, measurement and related problems. Remote Sensing of Environment, 12, 225–253. Bhattacharya, B. K., Dadhwal, V. K. (2003). Retrieval and validation of land surface temperature (LST) from NOAA AVHRR thermal images of Gujarat, India. International Journal of Remote Sensing, 24, 1197–1206. Çam, E., Arcaklıoğlu, E., Çavusoğlu, A., Akbıyık, B. (2005). A classification mechanism for determining average wind speed and power in several regions of Turkey using artificial neural networks. Renewable Energy, 30, 227–239. Caselles, V., Coll, C., Valor, E. (1997). Land surface emissivity and temperature determination in the whole HAPES- Sahel area from AVHRR data. International Journal of Remote Sensing, 18, 1009–1027. Chrysoulakis, N., Proedrou, M., Cartalis, C. (2001). Variations and trends in annual and seasonal means of precipitable water in Greece as deduced from radiosonde measurements. Environmental Toxicology and Chemistry, 84(1–4), 1–6. Cigizoğlu, H. K., Alp, M. (2006). Generalized regression neural network in modelling river sediment yield. Advances in Engineering Software, 37, 63–68. Coll, C., Caselles, V. (1997). A split-window algorithm for land surfaces temperature from Advanced Very High- Resolution Radiometer data: validation and algorithm comparison. Journal of Geophysical Research, 102, 16697–16713. Coll, C., Sobrino, J. A., Valor, E. (1994). On the atmospheric dependence of the split-window equation for land surface temperature. International Journal of Remote Sensing, 15, 105. Dash, P., Gottsche, F. M., Olesen, F. S., Fischer, H. (2002). Land surface temperature and emissivity estimation from passive sensor data: Theory and practice-current trends. International Journal of Remote Sensing, 23(13), 2563– 2594. Dinçer, I., Dilmaç, Ş., Ture, I. E., Edin, M. (1996). A simple technique for estimating solar radiation parameters and its application for Gebze. Energy Conversion and Management, 37(2), 183–198. Franca, G. B., Cracknell, A. P. (1994). Retrieval of land and sea surface temperature using NOAA-11 AVHRR data in north-eastern Brazil. International Journal of Remote Sensing, 15, 1695–1712. Gardner, M. W., Dorling, S. R. (1998). Artificial neural networks (the multiplayer perceptron)—a review of applications in the atmospheric sciences. Atmospheric Environment, 32, 2627–2636. Govindaraju, R. S. (2000a). Artificial neural network in hydrology, I: preliminary concepts. Journal of Hydrologic Engineering, 5(2), 115–123. Govindaraju, R. S. (2000b). Artificial neural network in hydrology, II: hydrological applications. Journal of Hydrologic Engineering, 5(2), 124–137. Kalogirou, S. A. (2000). Applications of artificial neural- networks for energy systems. Applied Energy, 67, 17–35. Kalogirou, S. A. (2001). Artificial neural networks in renew- able energy systems applications: a review. Renewable Sustainable Energy Reviews, 5, 373–401. Li, Z. L., Becker, F. (1993). Feasibility of land surface temperature and emissivity determination from AVHRR data. Remote Sensing of Environment, 43, 67–85. Mohandes, M. A., Halawani, T. O., Rehman, S., Hussain, A. A. (2004). Support vector machines for wind speed prediction. Renewable Energy, 29, 939–947. Prata, A. J. (1993). Land surface temperatures derived from the advanced very high resolution radiometer and the along-track scanning radiometer. I. Theory. Journal of Geophysical Research, 98, 16689–16702. Prata, A. J. (1994). Land surface temperatures derived from the Advanced Very High Resolution Radiometer and the along-track scanning radiometer. 2. Experimental results and validation of AVHRR algorithms. Journal of Geophysical Research, 99, 13025–13058. Price, J. C. (1984). Land surface temperature measurements from the split window channels of the NOAA-7/AVHRR. Journal of Geophysical Research, 89, 7231–7237. Qin, Z., Karnieli, A. (1999). Progress in the remote sensing of land surface temperature and ground emissivity using NOAA-AVHRR data. International journal of remote sensing, 20(12), 2367–2393. Sabins, F. F. (1997). Remote Sensing: Principles and Interpre- tation (3rd ed.). NewYork: W.H. Freeman. Schmugge, T., Hook, S. J., Coll, C. (1998). Recovering surface temperature and emissivity from thermal infrared multispectral data. Remote Sensing of Environment, 65, 121–131. Sobrino, J. A., Li, Z. L., Stoll, M. P., Becker, F. (1994). Improvements in the split window technique for land surface temperature determination. IEEE Transactions on Geoscience and Remote Sensing, 32, 243–253. Sobrino, J. A., Li, Z. L., Stoll, M. P., Becker, F. (1996). Multi-channel and multi-angle algorithms for estimat- ing sea and land surface temperature with ATSR data. International Journal of Remote Sensing, 17, 2089– 2114. Speckt, D. F. (1991). A generalized regression neural network. IEEE Transactions on Neural Networks, 2, 568–576. 408 J Indian Soc Remote Sens (September 2012) 40(3):399–409 Author's personal copy
- 13. Vidal, A. (1991). Atmospheric and emissivity correction of land surface temperature measured from satellite using ground measurements or satellite data. International Journal of Remote Sensing, 12, 2449–2460. Vogt, J. V. (1996). Land surface temperature retrieval from NOAA-AVHRR data. In G. D’Souza, A. S. Belward, J.-P. Malingreau (Eds.), Advances in the Use of NOAA-AVHRR data for Land Applications (pp. 125– 151). Dordrecht, The Netherlands pp: Kluwer Academic Publishers. Wan, Z., Dozier, J. (1996). A generalized split-windows algorithm for retrieving land surface temperature from space. IEEE Transactions on Geoscience and Remote Sensing, 34, 892–905. J Indian Soc Remote Sens (September 2012) 40(3):399–409 409 Author's personal copy