5.1. Assessment of Influence of Environmental parameters on THI
In the present study, seven important environmental parameters, such as NDVI, NDWI, NBI, LST, brightness, greenness, and wetness parameters, that have a possible influence in estimating the THI are considered. For this purpose, the one-to-one correlation coefficient of these parameters and THI are computed. Details of these coefficients (given in Table 2), the NDVI, NDWI, greenness, and wetness showed an inverse correlation, and LST, NBI, and brightness showed a significant positive correlation. The more the positive (negative) correlation coefficient, the parameter influencing THI toward more discomfort (comfort) conditions. It is noticed that LST has a higher positive correlation with THI, signifying that the higher the LST, the higher the discomfort. This influence of the LST on THI has been reported in (Imran et al., 2021; Taleghani, 2018). Similarly, it is noticed that the wetness parameter has shown a higher negative correlation, and hence it plays an important role in the assessment of more comfortable conditions.
Table 2
The correlation between Environmental parameters and THI
Parameters | NDVI | NDWI | NBI | LST | Brightness | Greenness | Wetness | THI |
NDVI | 1 | 0.763465 | -0.96925 | -0.92427 | -0.96265 | 0.652882 | 0.927594 | -0.75964 |
NDWI | 0.763465 | 1 | -0.87286 | -0.91663 | -0.87656 | 0.124373 | 0.932982 | -0.80089 |
NBI | -0.96925 | -0.87286 | 1 | 0.973764 | 0.997101 | -0.51179 | -0.98338 | 0.838802 |
LST | -0.92427 | -0.91663 | 0.973764 | 1 | 0.980885 | -0.40781 | -0.98715 | 0.808699 |
brightness | -0.96265 | -0.87656 | 0.997101 | 0.980885 | 1 | -0.48444 | -0.98316 | 0.836282 |
greenness | 0.652882 | 0.124373 | -0.51179 | -0.40781 | -0.48444 | 1 | 0.399354 | -0.29449 |
wetness | 0.927594 | 0.932982 | -0.98338 | -0.98715 | -0.98316 | 0.399354 | 1 | -0.83047 |
THI | -0.75964 | -0.80089 | 0.838802 | 0.808699 | 0.836282 | -0.29449 | -0.83047 | 1 |
Table 3 Mean absolute erorr of observed THI and computed THI
| 2018 | 2019 | 2020 |
Summer | 2.1℃ | 0.3℃ | 3.8℃ |
Winter | 2.7℃ | 0.18℃ | 0.66℃ |
Not only LST and wetness parameters but also other environmental parameters also influence the THI. Mijani et al., (2019 & 2020) reported that parameters such as brightness, greenness, and wetness parameters were found to have a high impact on micro-climate. The study of Mijani et al., (2020) revealed that NDVI and NDWI exhibit no significant role due to the study area is Tehran, an arid region; in our study region, they play a significant role in the influence of THI with correlation coefficients of -0.58 and − 0.80, respectively.
5.2. Estimation of THI with the combined influence of environmental parameters using a Machine learning model
After the correlation analysis, the SVM regression model was applied to train the data set, i.e., data of the summer and winter of THI for the year 2019. The THImodel coefficients correspond to NDVI, NDWI, NBI, LST, brightness, greenness, and wetness are 0.0168, -0.0467, 0.7986, 1.2296, 0.8508, 0.0256, -1.0991, respectively. From the coefficients, we noticed that the LST (\({a}_{4}=1.2296)\) has the highest contribution in THI, followed by brightness and NBI. The highest inverse contribution is wetness, followed by NDWI. The remaining parameters, NDVI, and greenness, are influenced reasonably in estimating THI.
The accuracy of the model for three different years and two seasons is given in Table 3. The lowest error was observed in the summer of 2019, and the highest error was in the summer of 2020. The errors are high for summer 2020 and winter 2018 because we assumed the single Landsat imagery data is the mean of a whole month. It is due to the lack of availability of Landsat data in individual day intervals, unlike Modis data.
5.3. Spatial variation of THI and LST over different land covers during summer and winter
The spatial variation of LST for summer and winter during three years, 2018, 2019, and 2020 are presented in Figs. 2a,2b,2c and Figs. 2d, 2e and 2f, respectively. During these three years, the LST varies from 300 − 41 0C for summer and 200 − 30 0C for winter. Clearly, the southern part of Hyderabad has a higher mean LST than the northern part. It is also noticed that higher vegetation cover has been noticed in the northeastern part of the city. Previous studies such as(Sannigrahi et al., 2017; Sultana & Satyanarayana, 2018) reported a similar result of LST over Hyderabad. The spatial variation of THI for summer and winter during the three years 2018, 2019, and 2020 are presented in Figs. 3a,3b,3c and Figs. 3d, 3e and 3f, respectively. The mean THI varies from 200 − 350C for summer and 140 − 280C for winter. In the recent study by Mijani et al., (2020), the mean THI over Tehran city in summer varies from 13.70–28.8 0C, and for winter, 7.30–17.9 0C, and these values show a linear relationship between THI and LST. This is clearly seen that Hussain Sagar and rive Musi are cooler having a lower magnitude in LST and THI exhibits the region of more comfort.
The region over the southern part of the city has shown a slight decrease in LST in the summer months of 2018 to 2020, whereas, during winter, it has shown an increase from 2019 to 2020. Interestingly, the northwestern part of the city has shown no considerable change in LST variation is seen during the summer, whereas during winter, the same region decrease in LST has been noticed (Figs. 2a,2b,2c,2d,2e, and 2f).
After analyzing the spatial variation of LST and THI, it is noticed that the southern part exhibits a discomfort region than the northern part due to the presence of more urban and bare lands. More than urban, the bare land has slightly higher THI magnitudes and causes discomfort for pedestrians. During the study period, variation in minimum and maximum LST was noticed, whereas there was no significant variation in THI over the city. In the summer season of 2019 (Fig. 2b), a maximum LST of > 41 0C is observed, whereas, during the summer of 2018 (Fig. 2a) and 2020 (Fig. 2c), it is 38 0C and 39 0C, respectively. But the maximum thermal comfort value is 32 0C (Figs. 3a, 3b, and 3c) during the study period. Hence, even though the LST influence is more on THI, the influence of the remaining six environmental parameters has been clearly portrayed.
Hence, one can conclude that the environmental parameters combinedly influence the micro-climate of the region. In order to establish more clearly, one can compare the spatial variation of LST and THI where ever the locations of higher LST zones do not exhibit higher THI zones (when compared in terms of area/patches).
5.4. LULC classification and corresponding mean THI over different land covers
To investigate in detail of variation of THI over different regions of the city, we need to classify the LULC. The LULC maps are generated from the maximum likelihood method suggested by Ahmad, (2012). The estimated classification accuracy of LULC maps is noticed to be very reasonable based on kappa statistics (Fleiss & Cohen, 2016). Over the study region, during the study period, the LULC maps ( shown in Fig. 4 ) with four primary classes, i.e., vegetation, water bodies, barren lands, and built-up. It is noticed that the percentage covered with vegetation cover, water bodies, built-up areas, and barren lands are, respectively, 19%, 2%, 51%, and 28%. No significant changes in the LULC are noticed during the study period. So, for analysis purposes, the 2019 winter is taken as the base of LULC. The accuracy of these ranges is estimated with kappa statistics, and the overall accuracy of Kapaa accuracy of 91% in 2019 winter. The estimated kappa hat is 0.84. From Fig. 4, one can clearly see the city of Hyderabad is more dominant in the urbanized area.
Figure 5. a shows the regions of mean THI and mean LST magnitudes over different land covers during all three years together to understand their variation over four different LULLC classes. The mean magnitudes of THI are noticed to be 20.73 0C, 27.28 0C, 24.210C, 26.9 0C, and mean LST is noticed to be 27.39 0C, 31.94 0C, 30.28 0C, 31.89 0C over water bodies, barren lands, vegetation, and built-up regions, respectively. As we noticed, the regions of water bodies, vegetation, and built-up and barren land exhibit more comfortable, neutral, moderate, and higher discomfort conditions, respectively. Figure. 5. b depicts the mean THI for 2018, 2019, and 2020. The mean THI is increasing every year and a 0.96 0C higher from 2018 to 2020 with no notable change in mean THI. The present results are noticed to be similar to that of reported ranges of values by Toy et al., (2007) that a THI value between 15°C and 20°C is perceived as the most comfortable condition and > 30 0C is perceived as the most discomfort.
5.5. Mandal-wise analysis of THI over different land covers
The percentage of the area coming under different classes of THI during summer and winter, along with LULC of each Mandal region, are depicted in Figs. 6a to 6d. During winter, it is noticed that most of the area in each of the Mandal regions has moderately comfortable to neutral conditions (Fig. 6a). But during summer, it is found that about 50 to 70% of the area under each Mandal region is dominated extreme discomfort conditions followed by moderate discomfort conditions as depicted in Fig. 6b, except Saidabad (17.35N, 78.51E) and Charminar (17.36N, 78.47E). The higher percentage of area in the Mandals such as Tirumalgiri (17.47N, 78.51E) and Khairatabad (17.41N, 78.46E) are under discomfort conditions in summer due to the presence of a higher percentage of urban and bare lands. Whereas, the regions of Charminar (17.36N, 78.47E), Golkonda (17.38N, 78.40E) and Himayatnagar (17.40N, 78.48E) have all classes of discomfort conditions due to the presence of moderate urbanization with noticeable vegetation as shown in Fig. 6c.
Figure 6d portrays the percentage of the area having extreme discomfort and high comfort conditions during summer and winter, respectively. It is noticed that the Mandal regions such as Secunderabad (17.43N, 78.49E), Maredpally (17.45N, 78.50E), Tirumalgiri (17.47N, 78.51E) during summer have a percentage of area in each of the Mandal comes under extremely discomfort conditions, respectively 67%, 69%, and 73%. But in winter, the Mandal regions, Nampally (17.38N, 78.46E), Charminar (17.36N, 78.47E), and Saidabad (17.35N, 78.51E) are having highest comfort zones covering 19%, 24%, 33% of the area.
From the spatial distribution of various classes of outdoor thermal comfort, it is noticed that over a few regions, the thermal comfort conditions vary from summer to winter in such a way that a few zones of discomfort conditions become comfort and vice versa. Hence in order to find the areas which are changing from zones of discomfort (summer) to comfort (winter) and vice versa, we have plotted the difference in THI magnitude between summer and winter, as portrayed in Fig. 7. The difference > 7 means the THI class is changing considerably from summer to winter and difference < 3 means the THI class is almost the same in summer and winter. We noticed locations near Bandlaguda (78.46E, 17.30N), Bhaduarapura (78.42E,17.34N), Golkonda (78.39E, 17.38N), and Begumpet airport (78.47E,17.45N) showed a maximum difference in THI from summer to winter becoming more discomfort zones in summer whereas comfort zones in winter. The large gray area in Fig. 7 shows that the variation from summer to winter is between 4–7 0C only, which means only a few regions change discomfort levels largely from season to season.
The variation in the percentage of mean THI among the comfort/discomfort locations is presented in Fig. 8. It is seen that in summer, most of the area above 50% is extreme discomfort, about 20% is moderate discomfort, 16% neutral, 4% moderately comfortable, and only about 1% highly comfortable. In winter, nearly 50% are moderately comfortable and followed by 30% neutral and 11% highly comfortable, 1.5% moderately discomfort, and 0.26% are highly discomfort. In the summer prevailed, extreme discomfort to moderate discomfort conditions, but in the winter, thermal comfort is mostly neutral to moderate comfort. The occurrence of extreme comfort in both summer and winter is less, and the occurrence of extreme discomfort in winter is almost negligible.