Various methods were used in the recent past for LULC classification which includes Maximum likelihood classifier, Indices overlay method, Neural networks, K-means clustering, etc. 32, 33, 34. In this study we applied maximum likelihood classification method for LULC classification. Four LULC classes has been identified i.e. built-up area, agricultural land, water bodies and vegetation cover.

To validate the classified maps kappa coefficient was applied using 350 randomly distributed points. These points then converted to kml to export them on google earth imagery for accuracy assessment. The overall accuracy was more than 89% of each map, literature suggests that accuracy above 80% is acceptable 35. The overall methodology that has been applied for LULC classification is provided in Figure 2.

For the extraction of LST, mono window algorithm has been applied. LST retrieval using mono window algorithm involves five steps 36. Details of methodology used for LST retrieval from Landsat data set is provided in Figure 3.

Step first involves conversion of digital number (DN) to spectral radiance (L_λ) using , respectively.

Spectral radiance calculation for Landsat 5 TM:

Where, L _λ = Spectral radiance

QCAL = Quantized calibrated pixel value.

Lmaxλ = Spectral radiance scaled to QCALMAX, which is taken from meta data.

Lminλ = Spectral radiance scaled to QCALMIN, taken from the metadata

QCALMAX = Maximum Quantized calibrated pixel value

QCALMIN = Minimum Quantized calibrated pixel value

Spectral radiance calculation for Landsat 8 (OLI):

Where:

M_L = multiplicative rescaling aspect of thermal band.

Q_cal = quantized pixel value.

A_L = additive rescaling factor of thermal band.

Second step involves conversion of spectral radiance to at satellite brightness temperature using Equation 3.

Where,

T= effective satellite Temperature in Kelvin.

k₁, and K₂ are Thermal Constants and value is given in Table 5 (from metadata).

Ln = Logarithmic function.

Third step involves the calculation of proportion of vegetation (p_v) by using Equation 4.

Fourth step involves emissivity correction by using Equation 5.

In the last step LST was calculated following Artis & Carnahan (1982) 37 using Equation 6.

Where,

T= LST in kelvin

B_T = brightness temperature

λ = wavelength of radiance

P = h × c l s (h is plank’s constant whose value is 6.624×10^-34 Js, and c is velocity of light, s is Boltzmann constant (1.38×10^-23 J/K)

ε = emissivity corrected

Urban heat island is formed when the temperature of city is higher than its surroundings. UHI is calculated using the Equation 7. 37.

Urban thermal field variance index (UTFVI) is used to evaluate the impact of urban heat island 38. It is determined by using Equation 8.

Urban hotspot in the present study has been calculated using following equation.

Here, μ is the mean LST and δ is the standard deviation of LST.

Land use land cover classification of Srinagar city was carried out using supervised maximum likelihood classification in four categories i.e. water body, agricultural land, settlement and vegetation cover Figure 4. The areal change in different LULC classes is given in Figure 5, Figure 6, which reveals that the area covered by water bodies in 2010 was 16.6 sq.km (6.49%) which reduced to 14.75 sq. km (6%) in 2022 i.e. slightly decreased by 1.85 sq. km (0.49%), whereas the built-up area in 2010 was 48.35 sq. km (18.86%) which increased to 127.89 sq. km (50%) in 2022, shows massive increase in the built -up area by 79.54 sq. km (31.02%) which is expanded into the agricultural and natural vegetation land that led to the reduction of both over the decade. The area under agricultural land in 2010 was 91.45 sq. km (35.66%) which reduced in 2022 to 39.68 sq. km (15%) has seen substantial reduction by 51.77 sq. km (20.18%) which is due to the encroachment by urban expansion in the area. On the other hand, the area under natural vegetation in 2010 was 99.95 sq. (38.99%) km which is also reduced to 74.1 sq. km (29%) in 2022 i.e. decreased by 25.85 sq. km (10.11%) which is attributed to expansion of built-up area and conversion of area under vegetation cover to urbanisation. This transformation is not merely a spatial pattern but is rooted in broader regional and socio-political dynamics. Srinagar, being the summer capital of Jammu and Kashmir, has undergone rapid urban expansion especially post-2010 due major infrastructural development, population pressure and growing influx of tourist. Smart City initiatives, and tourism-driven construction have significantly reshaped the city’s land use, led to the steady conversion of vegetative and agricultural lands into impervious built-up surfaces. Additionally, the valley’s constrained geography and limited availability of plain land further intensify land competition, accelerating this change.

A study conducted by Ahmad et al., (2024)39 recorded similar observation of Srinagar city which shows substantial decrease in water bodies, vegetation cover and increase in built-up area from the year 2000 to 2020. It reflects that urbanisation is growing at a rapid rate which is possibly due to population growth, economic development of city which led to demand of housing and infrastructural development. At the same time this steady increase in the built-up area is at the cost of natural vegetation and agricultural land which shows substantial decrease in area over a decade. This reduction in the natural vegetation and water bodies reflects potential environmental concern including increase runoff, reduction in biodiversity and natural resources.

LST of Srinagar city was calculated for four different seasons i.e. pre-monsoon (April), Monsoon (July), post-monsoon (October) and winter (December) of the year 2010 and 2022 respectively. Figure 7 reveals significant changes in LST across different seasons.

Table 6, Table 7 reveal that in the year 2010, minimum as well as maximum temperature was observed over built-up area i.e. -2.93 ^oC and 34.6 ^oC in the month of December and July respectively. built-up area exhibits higher mean LST than other LULC categories across all the seasons of the year except for the month of December, followed by agricultural land, vegetation and water bodies. Whereas, in the month of December highest temperature observed over agricultural land followed by natural vegetation, built-up area and water bodies.

However, In the year 2022, minimum temperature was -7.78 ^oC over agricultural land in the month of December whereas, maximum temperature 37.19 ^oC observed over built-up area in the month of June, which was 2.59 ^oC higher than temperature reported in 2010. In the year 2022 also, Built-up area showed higher LST across all the months except for the month of December which is followed by agricultural land, natural vegetation and water bodies. While, in the month of December highest temperature observed over natural vegetation followed by built up, water bodies and agricultural land. The overall observation reveals that in both the year 2010 and 2022 built-up area & water bodies showed highest and lowest LST respectively than other LULC category and other LULC categories such as natural vegetation, water bodies and agricultural land also showed consistent increase in LST over a decade. Many researches from different cities have reported similar observation where the average LST of different LULC categories have increased significantly over the years 4, 40.

As shown in the Figure 8, in the year 2010 during pre-monsoon and monsoon season small patches of UHS observed in the middle portion however, during post monsoon and winter season UHS are also confined in the eastern part of the city. In the year 2022, during pre-monsoon the UHS concentrated in central and southern part, during the monsoon period they are expanded to eastern direction, while, during post monsoon and winter season growing trend towards eastern and southern side can be seen. The mean area under UHI in 2010 was 247.22 sq.km which increased to 253.14 sq. km in 2022. These observations of UHI suggest changes in the distribution and intensity of UHIs over time, likely influenced by urbanisation, land use changes and seasonal variations. The shift towards eastern and southern sides is indicative of urban sprawl and reduction in vegetation cover, both of which can affect local temperature and create UHIs.

Urban thermal field variance index (UTFVI) is commonly utilized to assess the UHI effect and quality of life in urban environment 41, 42. To classify the UTFVI this study followed six tier classification proposed by Zhang 2006 (Excellent, good, normal, bad, worse & worst) as shown in Table 8 & Figure 9.

In 2010 majority of the area falls in the “excellent” condition of ecological evaluation index (EEI) with area percentage ranging from 47.27 in April to 57.91 in December. However, in 2022 there is slight decrease in excellent EEI with percentage ranges from 44.83 in April to 47.24 in December. Area under “Good” to “Normal” EEI is negligible with percentage 9.16 in April 2010 which reduced 1.34 in April 2022 thus experiencing weak to moderate UHI phenomenon. However, percentage of area under “Worst” EEI in 2010 was 43.57, 36.49,33.52 & 41.93 for the month of April, July, October & December respectively, which significantly increased in 2022 to 49.8 in April, 44.08 in July, 42.69 in October & 50.06 in December. This indicates gradual deterioration in urban ecological condition from 2010 to 2022 as area under “excellent” and “normal” category reduced from 2010 to 2022 but there is increase in area under “worst” category which reflects an increase in UHI intensity and a corresponding decrease in urban environmental quality.

Built up area shows higher mean LST in all the months of 2010 & 2022 which represents the formation of heat spots due to heat absorption and retention over the regions. This is consistent with other studies which shows positive correlation between LST and built up 21, 23, 43. Agricultural land also exhibits increase in mean LST in April and June, similar results observed by 9, 44 where agricultural land exhibits higher LST compared to natural vegetations due to increase in soil exposure over the regions. Water bodies also show consistent increase in LST. However, waterbodies generally exhibit lower temperature which creates cooling effect in the surrounding, but this cooling effect reduced due to rising temperature of water bodies 45. Natural vegetation shows similar trend to agricultural land with increase in mean LST in April and June but a decrease in October and December.