Introduction

During the COVID-19 epidemic, it is widely believed that nature is “taking a break” from people. Instead, land grabbing, deforestation, illegal mining, and wildlife poaching are all becoming more pressing issues in many tropical rural areas. Job losses in the city are driving people back to the countryside, putting a strain on already strapped natural resources and raising the danger of the spread of COVID-19 (Carella and D’Orazio 2021; Gutiérrez et al. 2021). Climate change is a public health concern because it is contributing to the spread of some diseases and making it harder to combat others (Chen et al. 2022; Nkwunonwo 2013). Season and climate play a significant role in determining the transmission rate of human viruses like the flu. While it is unclear at this time how climate change may affect the spread of COVID-19, research suggests that rising global temperatures will affect the timing, distribution, and severity of future disease outbreaks. The climate shifts caused by this worldwide occurrence will be studied in this specific setting (Kestens et al. 2011; Nkwunonwo 2013). Maps of the urban thermal field variance index (UTFVI) reveal a positive UHI phenomena, with the highest UTFVI zones occurring over developed area and none over the adjacent rural territory. The strongest UTFVI zone in the urban region expands during the forecasted years, especially during the warmer months (Mohammad and Goswami 2022).

The impact of COVID-19 on vegetation cover is a complex issue, as it is influenced by many factors such as changes in human activities, climate conditions, and natural processes. Some potential impacts include:

  1. a)

    Changes in land use: Due to the lockdowns and restrictions on movement, there may have been changes in land use, such as decreased use of agricultural land or decreased deforestation due to reduced logging and clearing activities.

  2. b)

    Changes in air quality: With reduced industrial and transportation activities during the pandemic, there may have been a reduction in air pollution, which can positively impact vegetation growth.

  3. c)

    Changes in water availability: Changes in human activities during the pandemic may have affected water availability in some areas. Reduced water use by industries and people can have positive impacts on vegetation growth. However, in some cases, reduced water flow from rivers can negatively impact vegetation growth.

  4. d)

    Changes in climate: While there may have been localized positive impacts on vegetation growth due to reduced human activities, the pandemic did not significantly alter the long-term trends in global warming or climate change, which can have negative impacts on vegetation growth.

The impacts of COVID-19 on vegetation cover are complex, and there is not enough data to draw clear conclusions about the net impact of the pandemic on vegetation growth. However, some localized positive impacts may have occurred due to changes in human activities and air quality. The lockdown is one of the best ways to prevent the transmission of the COVID-19 virus. As a result of the lockout, pollution levels in a city have dropped, improving the city’s ecological quality (Ahmar 2020; Akhtar and Nawaz 2020; Bashir et al. 2020; Bilal et al. 2020; Chakraborty and Maity 2020; Garg et al. 2020; Gupta et al. 2020; Halder et al. 2021; Halder and Bandyopadhyay 2021; Mandal and Pal 2020; Mohsinul Hoque et al. 2020). Restrictions on industries and factories and vehicle movement during the lockdown have reduced air pollution and improved air quality (Benintendi et al. 2020; Sailor and Dietsch 2007). Vegetation growth has taken place rapidly without human intervention (Fernandez et al. 2022). The decline in land surface temperature (LST) and increase in vegetation cover were one of the most noticeable and positive environmental impacts during this lockdown (Gatti et al. 2021; Kleinschroth et al. 2021; Radu et al. 2013; Song et al. 2022; Wang et al. 2021a, b). Compared to the rural areas, a significant drop in the LST value (by 1–2 °C) was observed in the urban regions during the night, indicating a weak urban heat Isla— phenomenon in the lockdown period due to the sharp reduction in air pollution and absorption of aerosols (Parida et al. 2021). Strategies including as afforestation, reforestation, and urban greening techniques via rooftop gardening are proposed to greatly increase capacity and mitigate the effects of global warming impacts in the most rapidly expanding cities (Rahaman et al. 2022).

Important research has also been carried out in India to examine the improvement in air quality during the lockdown period (Chauhan and Singh 2021; Garg et al. 2020; Kumari and Toshniwal 2022; Singh and Chauhan 2020; Yunus et al. 2020). A study conducted during the lockdown period (i.e., from March 25 to May 31, 2020) for the city of Raipur in India showed that the LST value has decreased and the normalized difference vegetation index (NDVI) value has increased due to the restrictions on transportation and commercial activities. Urban heat islands (UHIs) arise and are exacerbated in large part due to variations in both land use/land cover (LULC) and land surface temperatures (LST). Any city’s urban heat island (UHI) effect can be accurately described using the UTFVI. The purpose of this research is to evaluate the climate of Sylhet, a city in Bangladesh, by assessing and forecasting the seasonal (summer and winter) UTFVI scenario (Kafy et al. 2022b).

The relationship between LST and NDVI has also improved significantly during this period. It has been observed that the natural environment has changed compared to previous years, i.e., before the imposition of the lockdown, was ecologically better off with less pollution (Guha and Govil 2021b). Bera et al. (2021) conducted a study on the significant impact of lockdown on urban air pollution in the city of Kolkata, India, which showed a major decrease in the levels of pollutants such as CO2, NO2, and SO2, while a small increase in O3 levels was observed during the 2020 lockdown due to the closure of industries and transport units during this time. The NDVI is a straightforward graphical indication used in the analysis of remote sensing measurements, in this case, acquired from a space platform, to determine the presence or absence of green vegetation in the observed target. This metric will serve as a proxy because it would not be able to characterize changes in cover directly but will help us make the charts and graphs, we need to learn more about it. Prospective policymakers and city planners can reduce the effects of heat stress and make cities sustainable by evaluating the expected distribution maps of LULC, LST, UHI, and UTFVI (AlDousari et al. 2022).

The urban areas have a higher LST as compared to the surrounding rural regions (Kikon et al. 2022). One of the main reasons for the increase in environmental problems in the world is urbanization, which has a direct impact on climate variability (McCarthy et al. 2010). Vegetation characteristics and changes in soil temperature help determine an area’s LST (Yuan and Bauer 2007). LST is mainly dependent on solar irradiance and land surface properties (Guha and Govil 2020; Peng et al. 2016; Song et al. 2020). The carbon cycle and the robustness of ecosystems are uniquely impacted by human demands and the alteration of hydrology and land resources. Rapid urbanization is a major factor in the reduction of vegetative cover (VC), which in turn increases carbon emissions, land surface temperature (LST), and consequently, global warming. The study estimated LU/LC shifts by GIS and remote sensing methods, with a special emphasis on VC loss and its effect on LST (Kafy et al. 2022a).

High LST is reported in regions occupied by built-up structures, while low LST is reported in areas of vegetation cover (Li et al. 2017; Song et al. 2021). LST has been found to have an impact on land-use management and developmental activities (Gohain et al. 2023; Guha and Govil 2020, 2021a).

Land use and land cover (LULC) and land surface temperature (LST) change are strongly linked to the severity and genesis of urban heat island (UHI) phenomena. Quantifying the impact of UHI is possible with the use of the UTFVI. Increased focus is being placed on LST and UTFVI studies as tools for monitoring urban health and promoting sustainable growth. Using machine learning techniques, this research predicted shifts in LULC, LST by season (summer vs. winter), and UTFVI (Kafy et al. 2021). Rapid urbanization is replacing natural LULC classes, which can increase temperatures and reduce the thermal comfort zone (TCZ), both of which can have harmful effects on the environment and human health (AlDousari et al. 2023).

The NDVI is widely considered to be the main remote-sensing index regulating the change in LST (Chen et al. 2006; Mohammad and Goswami 2022; Saha et al. 2022). The improvement in air pollution and the increase in humidity mainly reinforce the strength of the relationship between LST and NDVI (Govil et al. 2019, 2020). The main purpose of this study was to compare LST, UTFVI, NDVI, normalized difference built-up index (NDBI), and LULC at a kilometer grid scale and to further validate the results using the historical imagery from Google Earth.

The study area and dataset used

Study area

Kohima Sadar is a district located in the north-eastern state of Nagaland in India. The district is one of the 11 administrative districts in Nagaland and is named after the district headquarters, Kohima. Kohima Sadar has a population of approximately 260,000 people and covers an area of 1102 square kilometers. Kohima Sadar is known for its rich cultural heritage and natural beauty. The district is home to several indigenous tribes, including the Angami, Chakhesang, Zeliang, and Rengma tribes. These tribes have their distinct languages, customs, and traditions that have been passed down through generations. The district is also known for its lush green forests and diverse wildlife. Kohima Sadar is home to the Japfu Peak, which is the second-highest peak in Nagaland, standing at an elevation of 3048 m. The Japfu Peak is a popular trekking destination and offers stunning views of the surrounding hills and valleys. Kohima, the district headquarters, is a vibrant city that offers a unique blend of traditional and modern culture. The city is known for its colorful markets, where one can find locally produced handicrafts, textiles, and traditional attire. The Kohima War Cemetery is a significant landmark in the city, which is dedicated to the soldiers who lost their lives during the Second World War.

The district is also known for its traditional festivals, which are celebrated with great enthusiasm and zeal. The Hornbill Festival is one of the most popular festivals in Nagaland, which is celebrated in the month of December every year. The festival is a celebration of the state's rich cultural heritage and is named after the state bird, the hornbill. Kohima Sadar is also known for its efforts to promote sustainable tourism. The district administration has taken several initiatives to promote eco-tourism and responsible tourism practices. The district is home to several community-based tourism initiatives that provide visitors with an opportunity to experience the local culture and traditions first-hand. This district that offers a unique blend of natural beauty, rich culture, and vibrant city life. The district is a testament to Nagaland's diverse cultural heritage and its commitment to promoting sustainable tourism practices. Whether it's trekking to the Japfu Peak, experiencing the local festivals, or exploring the vibrant city life, Kohima Sadar has something to offer for everyone.

Figure 1 shows the geographical location of Kohima Sadar, an administrative block of Kohima district which is also the capital of Nagaland in India. It covers a total area of 69.84 km2 and lies at 94° 7′ 28.606″ E and 25° 41′ 6.962″ N. Kohima Sadar is located in the central part of Kohima district and towards the north, it is bounded by Chiephobozou Block, towards the southeast by Jakhama Block and the west by the Sechu (Zubza) Block. It includes both rural and urban regions and three villages namely Chiede Model Village, Chiedema, and Kohima Village are found in this Block. The population of Kohima Sadar is estimated at 116,870 according to the 2011 census with the male and female ratio being 60,551 and 56,319 respectively and the literacy rate being 89.94% (https://www.censusindia.co.in/subdistrict/kohima-sadar-circle-kohima-nagaland-1849). The maximum and minimum elevation of Kohima Sadar ranges from 717 to 1811 m and consists of hilly topography. The average temperature is recorded throughout the summer season, but occasionally the temperature rises to as high as 30 °C with a minimum of 15 °C. On the other hand, during the winter season, there is extreme cold with a minimum temperature of up to 5 °C and a maximum of 20 °C.

Fig. 1
figure 1

Location map of the study area

Full size image

The dataset and software used

Landsat imagery was acquired from the United States Geological Survey (USGS) to calculate LST along with other land surface parameters such as building and vegetation. The Landsat TM satellite image dated November 29, 2009, and the Landsat 8 satellite image dated November 27, 2020, were used for this study. LST was retrieved from the thermal infrared band (TIRS) using the mono window algorithm. NDVI and NDBI were calculated using the red, near-infrared (NIR), and mid-infrared (MIR) bands of Landsat TM and Landsat 8 data. The red, near-infrared (NIR), and mid-infrared (MIR) bands for Landsat TM and Landsat 8 data each have a spatial resolution of 30 m. While the thermal infrared (TIRS) band has a spatial resolution of 120 m for Landsat TM and 100 m for Landsat 8 satellite images, both resampled by the USGS data center’s cubic convolution resampling method to a spatial resolution of 30 m (Guha and Govil 2021b). ENVI 5.0 software was used to retrieve LST and calculate NDVI, NDBI, and UTFVI. Arc GIS 10.2.2 software was also used for the grid-by-grid analysis and the creation of the maps, with a 1 km2 grid created using the Arc GIS spatial analysis tool. Correlation analysis was also performed using SPSS software.

Methodology

Figure 2 shows the methodology to determine the LST and land surface parameters such as NDVI and NDBI and to assess the ecological status of the study area based on the UTFVI which depends on the value of the LST. LST was determined using the TIR band of Landsat TM and Landsat 8 satellite imagery. In addition, NDVI and NDBI were obtained from the red, near infrared, and mid-infrared bands of Landsat TM and Landsat 8 satellite images. Figure 3 also shows the methodology for performing the grid-wise analysis, where a 1 km2 grid was prepared and the selected grids were further validated using Google Earth's historical imagery. The overall methodology consists of the following sub-items: (1) determination of LST, (2) estimation of UTFVI, (3) calculation of NDVI, (4) calculation of NDBI, (5) grid level analysis.

Fig. 2
figure 2

Flowchart of the methodology for determination of LST and surface parameters

Full size image
Fig. 3
figure 3

Flowchart of the methodology for grid-level analysis

Full size image

Methodology for determination of LST and land surface parameters

Determination of LST

In this study, the LST derivation algorithm is used to calculate the LST (Qin et al. 2001). Emissivity, transmittance, and mean atmospheric temperature are the requisite parameters for this algorithm. This algorithm was applied to the TIR bands (i.e., band 6 and band 10 of Landsat TM and Landsat 8 data, respectively) and Eq. (1) shows the formula of this algorithm:

$$T_{s}= \{a (1-C-D) + [b (1-C-D) +C+D] T_{i}-D*T_{a}\}/C$$
(1)

where

a :

− 67.355351

b :

0.4558606

C:

ε i * τ i ,

D :

(1-τi) [1 + (1-εi)*τi)

ε i :

Emissivity

τ i :

Transmissivity

  1. 1.

    Digital numbers to radiance values

Equation (2) was applied to thermal band 6 and band 10 of Landsat TM and Landsat 8 data, respectively, to obtain spectral radiance values by converting the digital numbers. The metadata folder stores the band-specific information of the satellite data.

$$L_{\lambda}=M_{L} \ast Q_{Cal}+A_{L}$$
(2)

where

L λ :

TOA spectral radiance (W/(m2 × srad × μm))

M L :

Band-specific multiplicative rescaling factor from the metadata (RADIANCE_MULT_BAND_x, where x is the band number)

A L :

Band-specific additive rescaling factor from the metadata (RADIANCE_ADD_BAND_x, where x is the band number)

Q Cal :

Quantized and calibrated standard product pixel values (DN)

  1. 2.

    Radiance values to brightness temperature

The brightness temperature is obtained from the spectral radiance value using the inverse of the Plank function (W.-C. Wang et al. 1990). The formula is in Eq. (3).

$$T=\frac{{K}_{2}}{\mathrm{ln}\left(\frac{{K}_{1}\times \varepsilon }{{CV}_{R1}}+1\right)}$$
(3)

where

T:

Degrees (in Kelvin).

CVR1:

Radiance value.

Also, the K1 and K2 values are obtained from the meta-data file.

  1. 3.

    Atmospheric transmittance

NASA’s Atmospheric Correction Module webpage was used to obtain the atmospheric transmittance values (https://atmcorr.gsfc.nasa.gov/).

  1. 4.

    Emissivity

Van de Griend and Owe (1993) have proposed a method for calculating emissivity. First, NDVI was derived using the reflectance values of the red and near-infrared bands. Equation (4) is used to determine emissivity when the value of NDVI is in the range of 0.157 to 0.727.

$${\varepsilon }_{i}=1.0094+0.0047 ln (NDVI)$$
(4)

Zhang et al. (2006) also proposed a technique for calculating the land surface emissivity which is dependent on the NDVI values as seen in Table 1.

Table 1 Values of NDVI for calculation of land surface emissivity
Full size table

Determination of LST from Landsat 8

Procedure for automated map** of land surface temperature using Landsat 8 satellite data is performed and synthesized in six steps below with the Raster Calculator tool in ArcMap:

  1. 1.

    Calculation of TOA (top of atmospheric) spectral radiance

$$\mathrm{TOA }\left(\mathrm{L}\right)={\mathrm{M}}_{\mathrm{L}}\times {\mathrm{Q}}_{\mathrm{cal}}+{\mathrm{A}}_{\mathrm{L}}$$

where

ML:

Band-specific multiplicative rescaling factor from the metadata (RADIANCE_MULT_BAND_x, where x is the band number)

Qcal:

Corresponds to band 10

AL:

Band-specific additive rescaling factor from the metadata (RADIANCE_ADD_BAND_x, where x is the band number)

$$\mathrm{TOA}=0.0003342\times "\mathrm{Band}" 10+0.1$$
  1. 2.

    TOA to brightness temperature conversion

    $$\mathrm{BT}=\left({\mathrm{K}}_{2}/\left(\mathrm{ln}\left({\mathrm{K}}_{1}/\mathrm{L}\right)+1\right)\right)-273.15$$

    where

K1:

Band-specific thermal conversion constant from the metadata (K1_CONSTANT_BAND_x, where x is the thermal band number)

K2:

Band-specific thermal conversion constant from the metadata (K2_CONSTANT_BAND_x, where x is the thermal band number)

L:

TOA

The radiant temperature is adjusted by adding the absolute zero (approx. − 273.15 °C) in order to obtain the results in Celsius, t

$$\mathrm{BT}=\left(1321.0789/\mathrm{Ln}\left(\left(774.885/"\mathrm{\%TOA\%}"\right)+1\right)\right)-273.15$$
  1. 3.

    Calculate the NDVI

$$\mathrm{NDVI}=\left(\mathrm{Band\ }5-\mathrm{Band\ }4\right)/\left(\mathrm{Band\ }5+\mathrm{Band\ }4\right)$$

Note that the calculation of the NDVI is important because, subsequently, the proportion of vegetation (Pv), which is highly related to the NDVI, and emissivity (ε), which is related to the Pv, must be calculated.

$$\mathrm{NDVI}=\mathrm{Float}\left(\mathrm{Band\ }5-\mathrm{Band\ }4\right)/\mathrm{Float }\left(\mathrm{Band\ }5+\mathrm{Band\ }4\right)$$
  1. 4.

    Calculate the proportion of vegetation Pv

$${\mathrm{P}}_{\mathrm{v}}=\mathrm{Square }\left(\left(\mathrm{NDVI}-{\mathrm{NDVI}}_{\mathrm{min}}\right)/{\mathrm{NDVI}}_{\mathrm{max}}-{\mathrm{NVDI}}_{\mathrm{min}}\right)$$

Usually, the minimum and maximum values of the NDVI image can be displayed directly in the image (both in ArcGIS, QGIS, ENVI, Erdas Imagine); otherwise, you must open the properties of the raster to get those values.

Calculate emissivity ε

$$\upvarepsilon =0.004*{\mathrm{P}}_{\mathrm{v}}+0.986$$

The raster calculator is used to estimate emissivity with a correction value of 0.986 in the equation.

Calculate the land surface temperature

$$\mathrm{LST}=\left(\mathrm{BY}/\left(1+\left(0.00115\times \mathrm{BT}/1.4388\right)\times \mathrm{Ln}\left(\upvarepsilon \right)\right)\right)$$

The LST equation is applied to obtain the surface temperature map.

Estimation of UTFVI

The UTFVI is calculated based on the value of LST. UTFVI has been used to study the urban heat island effect phenomenon on the ecological state of the environment (Kikon et al. 2016). UTFVI is dependent on the land surface temperature i.e., the higher the LST, the higher the heat island effect (Liu and Weng 2012; Liu and Zhang 2011). Equation (5) was used for the calculation of UTFVI.

$$UTFVI= \frac{(T_{S}-T_{mean})}{T_{mean}}$$
(5)

where

T S :

Land surface temperature of a certain point (in Kelvin)

T mean :

Mean LST of the whole study area (in Kelvin)

Calculation of NDVI

Equation (6) was used to calculate the NDVI using the red and near-infrared reflectance bands from Landsat TM and Landsat 8 data to determine the greenness of an area (Liu and Weng 2012).

$$NDVI=\frac{NIR-R}{NIR+R}$$
(6)

where

R:

Band 3 (for Landsat TM) and Band 4 (for Landsat 8)

NIR:

Band 4 (for Landsat TM) and Band 5 (for Landsat 8)

Calculation of NDBI

Equation (7) was used to calculate the normalized difference built-up index (NDVI) using the mid-and near-infrared reflectance bands from Landsat TM and Landsat 8 data to indicate the built-up area of an area (Zha et al. 2003).

$$NDBI=\frac{MIR-NIR}{MIR+NIR}$$
(7)

where

MIR:

Band 5 (for Landsat TM) and Band 6 (for Landsat 8)

NIR:

Band 4 (for Landsat TM) and Band 5 (for Landsat 8)

Grid-level analysis

As suggested in the methodology shown in Fig. 3, a grid-level analysis was performed to examine the micro-level LST for a single grid. The study area was subdivided into a 1 km2 grid using the Arc GIS 10.2.1 fishnet and zonal statistics tool. The main objective of this analysis is to study the main land use category responsible for the formation of urban heat island. The grid-level analysis of the different surface parameters was performed by calculating the mean of LST, UTFVI, NDVI, and NDBI within the area of a 1 km2 grid. At the micro-level, a selection of grids based on the main land use category (i.e., the grids that have built up as the major land feature) was performed for further analysis and validation using the historical imagery from Google Earth, which was also helpful in observing the trend of LST.

Results and analysis

Trends of LULC, LST, NDVI, NDBI, and UTFVI

Spatial–temporal analysis of LULC for 2009 and 2020

Spatial–temporal analysis of land use land cover (LULC) involves the examination of changes in the distribution and patterns of land use and land cover over time. This analysis is essential for monitoring environmental changes, assessing the impacts of human activities, and develo** sustainable land use policies. The process of spatial–temporal analysis of LULC typically involves the following steps:

  1. a)

    Data collection: The first step in spatial–temporal analysis of LULC is to collect data. This data can be collected from various sources, including satellite imagery, aerial photographs, and ground surveys.

  2. b)

    Data pre-processing: Once the data has been collected, it needs to be processed to remove errors and inconsistencies. This may involve correcting image distortions, mosaicking, and filtering the data to remove noise.

  3. c)

    Image classification: Image classification is the process of assigning land use and land cover categories to each pixel in the image. This can be done using various techniques, including supervised and unsupervised classification methods.

  4. d)

    Change detection: Change detection involves comparing two or more LULC maps from different time periods to identify changes in land use and land cover. Change detection techniques can be used to detect land use changes, such as deforestation, urbanization, and agricultural expansion.

  5. e)

    Spatial–temporal analysis: The final step in spatial–temporal analysis of LULC is to analyze the results and interpret the findings. This may involve identifying areas of significant change, map** trends in land use and land cover, and develo** models to predict future changes.

  6. f)

    Spatial–temporal analysis of LULC can provide valuable insights into the dynamics of land use and land cover change. This information can be used to develop policies and strategies for sustainable land use management, biodiversity conservation, and disaster risk reduction. For example, it can help identify areas that are vulnerable to environmental degradation and prioritize conservation efforts.

Spatial–temporal analysis of LULC is an essential tool for monitoring and managing land resources and ensuring sustainable development. Figure 4 shows Kohima Sadar’s LULC map for 2009 and 2020. A comparative study showed that the built-up area was 7.3 km2 in 2009 and increased to 15.9 km2 in 2020. The undeveloped area consisting of barren and open land decreased significantly from 22.8 km2 in 2009 to 16.3 km2 in 2020.

Fig. 4
figure 4

LULC map for 2009 and 2020

Full size image

No noticeable changes were observed in vegetation and green cover over the years, with the area being 39.3 km2 and 37.2 km2 in 2009 and 2020, respectively. This analysis showed that the non-built-up areas have been replaced by the built-up areas and the urbanization is mainly concentrated in the main town area, which is slowly spreading to other directions in the study area. As can be seen in Table 2, it was observed that despite the increase in urbanization, a lower LST was reported in 2020, with the mean LST being 21.5 °C, while in 2009, the mean LST was 24.5 °C. There was a 3 °C decrease in the value of the mean LST in 2020.

Table 2 Table showing the temporal change in LULC categories, LST, NDVI, and NDBI during 2009 and 2020
Full size table

Spatial–temporal analysis of LST for 2009 and 2020

Spatial–temporal analysis of land surface temperature (LST) involves the examination of changes in temperature patterns over space and time. LST is an important variable for studying the Earth’s energy balance, as it affects various environmental processes, including climate, hydrology, and vegetation growth. The process of spatial–temporal analysis of LST typically involves the following steps:

  1. a)

    Data collection: The first step in spatial–temporal analysis of LST is to collect data. This data can be collected from various sources, including satellite images, ground-based observations, and meteorological data.

  2. b)

    Data pre-processing: Once the data has been collected, it needs to be processed to remove errors and inconsistencies. This may involve correcting image distortions, calibration, and filtering the data to remove noise.

  3. c)

    Image processing: Image processing involves the conversion of raw data into LST values. This can be done using various techniques, including thermal band conversion and atmospheric correction.

  4. d)

    Spatial–temporal analysis: The final step in spatial–temporal analysis of LST is to analyze the results and interpret the findings. This may involve identifying areas of significant change in LST patterns, map** trends in temperature over time, and develo** models to predict future changes.

Spatial–temporal analysis of LST can provide valuable insights into the dynamics of temperature changes in different regions. This information can be used to develop policies and strategies for managing the impacts of climate change, such as develo** heatwave warning systems, managing urban heat islands, and mitigating the impacts of droughts. For example, spatial–temporal analysis of LST can help identify areas that are vulnerable to heat stress, such as densely populated urban areas or regions with high agricultural activity. This information can be used to develop strategies to reduce heat exposure and improve public health outcomes.

Spatial–temporal analysis of LST is an essential tool for monitoring and managing temperature patterns over time, which is crucial for understanding the impacts of climate change and develo** effective adaptation strategies. Figure 5 shows the distribution map of LST in 2009 and 2020 for Kohima Sadar. The minimum and maximum values of LST during the pre-COVID scenario (i.e., in 2009) were 16 °C and 33 °C, respectively, with a mean LST of 24.5 °C, while the minimum and maximum values of LST during the post-COVID scenario (i.e., in 2020) was 13 °C and 30 °C, respectively, with a mean LST of 21.5 °C. During the post-COVID scenario, it was observed that the mean LST value decreased by 3 °C within the study region, resulting in a positive change in this period. This is mainly due to various SOPs imposed during the COVID lockdown period whereby movement of the human population and vehicles on roads were restricted to the minimum. A drastic decrease in vehicular movement and industrial pollutants due to the COVID lockdown period plays a vital role in reducing mean LST. Throughout 2020, LST has decreased significantly in the northwest and southwest regions. There was also a significant increase in green cover and a significant decrease in pollution levels, which directly contributed to the reduction in LST.

Fig. 5
figure 5

LST map for 2009 and 2020

Full size image

Spatial–temporal analysis of NDVI for 2009 and 2020

Spatial–temporal analysis of NDVI involves the examination of changes in vegetation cover over time. NDVI is a commonly used remote sensing index to measure vegetation growth, health, and productivity. The process of spatial–temporal analysis of NDVI typically involves the following steps:

  1. a)

    Data collection: The first step in spatial–temporal analysis of NDVI is to collect data. This data can be collected from various sources, including satellite imagery, aerial photographs, and ground surveys.

  2. b)

    Data pre-processing: Once the data has been collected, it needs to be processed to remove errors and inconsistencies. This may involve correcting image distortions, mosaicking, and filtering the data to remove noise.

  3. c)

    Image processing: Image processing involves the calculation of NDVI values from raw data. This can be done using various techniques, including image fusion and atmospheric correction.

  4. d)

    Change detection: Change detection involves comparing two or more NDVI maps from different time periods to identify changes in vegetation cover. Change detection techniques can be used to detect changes in vegetation, such as deforestation, urbanization, and agricultural expansion.

  5. e)

    Spatial–temporal analysis: The final step in spatial–temporal analysis of NDVI is to analyze the results and interpret the findings. This may involve identifying areas of significant change in vegetation cover, map** trends in vegetation growth and productivity, and develo** models to predict future changes.

Spatial–temporal analysis of NDVI can provide valuable insights into the dynamics of vegetation cover and productivity over time. This information can be used to develop policies and strategies for sustainable land use management, biodiversity conservation, and disaster risk reduction. For example, it can help identify areas that are vulnerable to environmental degradation and prioritize conservation efforts. Spatial–temporal analysis of NDVI is an essential tool for monitoring and managing vegetation cover and productivity, which is crucial for understanding the impacts of climate change and develo** effective adaptation strategies. Figure 6 demonstrates the spatio-temporal distribution map of NDVI for 2009 and 2020. The minimum and maximum values of NDVI during the pre-COVID scenario (i.e., in 2009) were − 0.4 and 0.7, respectively, with a mean NDVI value of 0.15, while the minimum and maximum NDVI values during the post-COVID scenario (i.e., in 2020) were found to be 0.04 and 0.8, respectively, with a mean NDVI value of 0.42. The post-Covid scenario had a large impact on the NDVI with a higher mean NDVI value. During this period, vegetation cover health was unaffected by anthropogenic disturbances such as emissions, transport, and human and industrial activities, resulting in low-level pollution that increased NDVI. As shown in the map of NDVI in Fig. 6, NDVI has increased in certain parts of the Northwest and South of the study area during the post-COVID period, indicating a positive change.

Fig. 6
figure 6

NDVI map for 2009 and 2020

Full size image

Spatial–temporal analysis of NDBI for 2009 and 2020

Spatial–temporal analysis of NDBI involves the examination of changes in built-up areas over time. NDBI is a commonly used remote sensing index to measure the degree of urbanization and built-up areas in an image. The process of spatial–temporal analysis of NDBI typically involves the following steps:

  1. a)

    Data collection: The first step in spatial–temporal analysis of NDBI is to collect data. This data can be collected from various sources, including satellite imagery and aerial photographs.

  2. b)

    Data pre-processing: Once the data has been collected, it needs to be processed to remove errors and inconsistencies. This may involve correcting image distortions, mosaicking, and filtering the data to remove noise.

  3. c)

    Image processing: Image processing involves the calculation of NDBI values from raw data. This can be done using various techniques, including band ratio and thresholding.

  4. d)

    Change detection: Change detection involves comparing two or more NDBI maps from different time periods to identify changes in built-up areas. Change detection techniques can be used to detect changes in built-up areas, such as urban expansion and land-use changes.

  5. e)

    Spatial–temporal analysis: The final step in spatial–temporal analysis of NDBI is to analyze the results and interpret the findings. This may involve identifying areas of significant change in built-up areas, map** trends in urbanization and land-use changes, and develo** models to predict future changes.

Spatial–temporal analysis of NDBI can provide valuable insights into the dynamics of urbanization and land-use changes over time. This information can be used to develop policies and strategies for sustainable land use management, urban planning, and disaster risk reduction. For example, it can help identify areas that are vulnerable to urbanization and prioritize measures to mitigate the impacts of urbanization on the environment. Spatial–temporal analysis of NDBI is an essential tool for monitoring and managing urbanization and land-use changes, which is crucial for understanding the impacts of urbanization and develo** effective adaptation strategies.

Figure 7 shows the spatio-temporal distribution map of NDBI for 2009 and 2020. The minimum and maximum values of NDBI during the pre-COVID scenario (i.e., in 2009) were − 0.9 and 0.6, respectively, with a mean NDBI value of − 0.15, while the minimum and maximum NDBI values during the post-COVID scenario (i.e., in 2020) were determined to be − 0.6 and 0.3, respectively, with a mean NDBI value of − 0.15. As can be seen in the NDBI map in Fig. 7, there was a significant increase in NDBI from the north-northwest region to the south-southwest region which is mainly concentrated in the south-west and central part of the study area. A slight increase in NDBI is also observed from the east, spreading to the south-eastern part of the study region. Despite the increase in the built-up area during the post-COVID scenario, the mean LST has fallen by 3 °C over this period, which is a sign of positive environmental change.

Fig. 7
figure 7

NDBI map for 2019 and 2020

Full size image

Spatial–temporal analysis of UTFVI for 2009 and 2020

The UTFVI is a remote sensing index that measures the spatial variation of surface temperature in urban areas. The spatial–temporal analysis of UTFVI involves examining the changes in surface temperature in urban areas over time. The process of spatial–temporal analysis of UTFVI typically involves the following steps:

  1. a)

    Data collection: The first step in spatial–temporal analysis of UTFVI is to collect data. This data can be collected from various sources, including satellite imagery, aerial photographs, and ground surveys.

  2. b)

    Data pre-processing: Once the data has been collected, it needs to be processed to remove errors and inconsistencies. This may involve correcting image distortions, mosaicking, and filtering the data to remove noise.

  3. c)

    Image processing: Image processing involves the calculation of UTFVI values from raw data. This can be done using various techniques, including band ratio and thresholding.

  4. d)

    Change detection: Change detection involves comparing two or more UTFVI maps from different time periods to identify changes in surface temperature. Change detection techniques can be used to detect changes in surface temperature, such as urban heat island effects and changes in land use.

  5. e)

    Spatial–temporal analysis: The final step in spatial–temporal analysis of UTFVI is to analyze the results and interpret the findings. This may involve identifying areas of significant change in surface temperature, map** trends in surface temperature and thermal variability, and develo** models to predict future changes.

Spatial–temporal analysis of UTFVI can provide valuable insights into the dynamics of surface temperature in urban areas over time. This information can be used to develop policies and strategies for sustainable urban planning and management, climate change mitigation and adaptation, and public health interventions. For example, it can help identify areas that are vulnerable to heat stress and prioritize interventions to reduce the impacts of urban heat islands on human health. Spatial–temporal analysis of UTFVI is an essential tool for monitoring and managing surface temperature in urban areas, which is crucial for understanding the impacts of urbanization on the environment and human health and develo** effective adaptation strategies.

Figure 8 shows the spatio-temporal distribution map of UTFVI for 2009 and 2020. To examine the impact of the spatial distribution of urban heat islands in the study area, UTFVI was divided into six grades based on ecological assessment index thresholds, ranging from excellent (< 0) to worst (> 0.020), see Table 3. Kohima Sadar’s topography is mountainous, and most areas are covered by forest and vegetation. It consists of both urban and rural areas, and urbanization is mainly limited to the central part of the study region. The urban heat island phenomenon has been observed to be more concentrated in the central part of Kohima Sadar. The UTFVI range was > 0.020 in certain patches of the central and southern parts of the study area where ecological status was worst. In contrast, a weak urban heat island phenomenon with a UTFVI range < 0 has occurred in the western and southwestern parts of the study region. The normal ecological conditions were observed in small pockets in vegetation and forest areas that were in the UTFVI range (0.005–0.010), while the worst ecological conditions (> 0.020) occurred around the built-up classes. During the post-COVID period, it was observed that the area under the worst ecological conditions had slightly decreased in some places in the central part of the study region, despite increasing urbanization. In addition, Fig. 4 and Table 2 show that during the post-COVID period, the built-up area increased by 8.6 km2 and the area covered with vegetation decreased by 2.1 km2. It must be noted that the study area had a better ecological status during this period due to the drastic decrease in vehicle movements and industrial activities that played a crucial role in reducing the mean LST.

Fig. 8
figure 8

UTFVI map for 2009 and 2020

Full size image
Table 3 Table displaying the threshold values of ecological evaluation index
Full size table

Grid-level analysis

Grid-level analysis is a geographic information system (GIS) technique used to analyze data within a grid or raster format. It involves dividing a study area into smaller grids or cells, where each cell represents a specific area or unit of analysis. This approach allows for the visualization and analysis of data at a higher resolution, providing a more detailed understanding of spatial patterns and relationships. Grid-level analysis can be used in various fields, including environmental science, urban planning, and public health. For example, in environmental science, grid-level analysis can be used to map and analyze changes in land use, vegetation cover, and surface temperature. In urban planning, it can be used to analyze land use patterns, urban density, and transportation infrastructure. In public health, it can be used to identify areas at higher risk for disease outbreaks or environmental hazards. The process of grid-level analysis typically involves the following steps:

  1. a)

    Data collection: The first step in grid-level analysis is to collect data, which can be obtained from various sources, including satellite imagery, aerial photographs, and ground surveys.

  2. b)

    Data processing: Once the data has been collected, it needs to be processed and converted into a grid or raster format. This may involve resampling, georeferencing, and filtering the data to remove noise.

  3. c)

    Grid generation: Grid generation involves dividing the study area into smaller grids or cells, where each cell represents a specific area or unit of analysis. The size of the grid cells can vary depending on the resolution of the data and the scale of analysis.

  4. d)

    Data analysis: Data analysis involves applying various statistical and spatial analysis techniques to the grid data. This may include calculating averages, standard deviations, and spatial autocorrelation, among others.

  5. e)

    Visualization and interpretation: The final step in grid level analysis is to visualize and interpret the results. This may involve creating maps, charts, and graphs to identify patterns and trends, and develo** models to predict future changes.

Grid level analysis is an important GIS technique used to analyze data at a higher resolution, providing a more detailed understanding of spatial patterns and relationships. It can be used in various fields to support decision-making processes and inform policy development.

Figure 9 presents grid-wise map of LULC, Fig. 10 presents grid-wise map of LST, Fig. 11 exhibits grid- wise map of NDVI, Fig. 12 exhibits grid-wise map of NDBI, and Fig. 13 shows grid-wise map of UTFVI for the 2009 and 2020, respectively.

Fig. 9
figure 9

Grid-wise map of LULC for 2009 and 2020

Full size image
Fig. 10
figure 10

Grid-wise map of LST for 2009 and 2020

Full size image
Fig. 11
figure 11

Grid-wise map of NDVI for 2009 and 2020

Full size image
Fig. 12
figure 12

Grid-wise map of NDBI for 2009 and 2020

Full size image
Fig. 13
figure 13

Grid-wise map of UTFVI for 2009 and 2020

Full size image

Discussion

A grid-level analysis was performed to examine the micro-level LST for a single grid region. The Kohima Sadar study area was subdivided into a 1 km2 grid based on zonal statistics. The main purpose of this analysis is to study the main land use category responsible for the formation of urban heat island (Halder and Bandyopadhyay 2021; Pal and Ziaul 2017). The grid-level LST analysis was performed by calculating the mean of the LST within the 1 km2 grid area (Dovì et al. 2009; Kumar 2016; Kumar et al. 2018b). At the micro-level, the main land use category falling within the grid area was shown, which helped observe the trend of LST (Guan 2011; Kumari and Pandey 2020).

As can be seen in Fig. 10, it was observed that the grids with the highest LST in the study area were mainly located in the central part and some grids in the northern parts also showed a high LST (Chen et al. 2021; Kayet et al. 2016; Yang et al. 2020). Comparing each of the high LST grids to the land use grids, it was observed that the grids with the highest LST classes had built-up and non-built-up classes as the main land use category, as can be seen in Fig. 9. This is due to the emission characteristics of the various land surface features, where the ground surface has a low emissivity coefficient value in the range of about 0.65–0.99, for open and barren land, and when the solar radiation falls on the ground, there are no restrictions, and the radiation comes into direct contact with the surface and which is then immediately released, resulting in high land surface temperatures (Lin et al. 2018; Liu et al. 2021). In built-up areas, the various building materials such as brick and concrete have an emissivity of 0.93 and 0.85, respectively, which traps solar radiation and slowly releases it as heat, creating an urban heat island and resulting in increased land surface temperature (Kikon et al. 2022; Kumar 2017; Kumar et al. 2018a; Kumar and Shekhar 2015). As can be seen in Fig. 11, the high LST grids also show the presence of a low NDVI, a high NDBI in the case of grids with built-up areas as the main land use category, and a low NDBI in the grids consisting of open and barren land as the main land use category as shown in Fig. 12. As can be seen in Fig. 13, the UTFVI result also shows that the grids with high LST had the worst ecological status in the study area. In this study, some grids with high LST and grids showing a change in LST were selected and compared to the grid-wise map (Halder and Bandyopadhyay 2021; Najafzadeh et al. 2021). The selected grids are numbered 1 and 2 as shown in the map in Figs. 9, 10, 11, 12, and 13 (Kikon et al. 2022; Najafzadeh et al. 2021). From the results of this analysis, it can be concluded that the lockdown period has a direct impact on the decreasing LST value (Shetty et al. 2021; Shi et al. 2021; Wang et al. 2021a, b). The value of LST has decreased noticeably on all grids during the lockdown period which has reduced the movement of the human population and vehicles on the roads to a minimum. A drastic drop in vehicle movements and industrial pollutants due to restrictions played a crucial role in reducing LST. Below are some example images from Google Earth Historical Imagery of grids number 1 and 2 showing a decrease in LST and an increase in vegetation cover during lockdown period, which represents a positive change.

Pearson’s correlation of different indices

Table 4 shows the summary of correlations between NDBI, NDVI, UTFVI, and LST for grid number 1 in 2009.

Table 4 Summary of correlations for grid number 1 in 2009
Full size table

Table 5 reports a summary of correlations between NDBI, NDVI, UTFVI, and LST for grid number 1 in 2020.

Table 5 Summary of correlations for grid number 1 in 2020
Full size table

Table 6 reports a summary of correlations between NDBI, NDVI, UTFVI, and LST for grid number 2 in 2009.

Table 6 Correlations for grid number 2 in 2009
Full size table

Table 7 provides summary of correlations between NDBI, NDVI, UTFVI, and LST for grid number 2 in 2020.

Table 7 Correlations for grid number 2 in 2020
Full size table

Tables 4, 5, 6, and 7 show the correlation between NDBI, NDVI, UTFVI, and LST during the 2020 lockdown period and the previous year 2009, which was before the start of the pandemic for the selected grid numbers 1 and 2. The result shows that there is a positive correlation between LST and NDBI with a Pearson correlation value of 0.235 and 0.945 in 2009 and 2020, respectively, for grid number 1 and a Pearson correlation value of 0.857 and 0.727 in 2009 and 2020, for grid number 2. This is due to the properties of the various building materials that capture the incident radiation and gradually release it as heat, leading to the formation of urban heat islands that lead to high land surface temperatures. Likewise, LST and NDVI are negatively correlated with a Pearson correlation value of − 0.517 and − 0.736 in 2009 and 2020, respectively, for grid number 1, and a Pearson correlation value of − 0.533 and − 0.519 in 2009 and 2020, respectively, for grid number 2. This is because plants absorb the incoming solar radiation, resulting in low land surface temperature due to their high emissivity (Kikon et al. 2022). A strong inverse correlation was found between NDVI and NDBI, with Pearson correlation values of − 0.771 and − 0.843 in 2009 and 2020, respectively for grid number 1, and Pearson correlation values of − 0.816 and − 0.954 in 2009 and 2020, respectively, for grid number 2. The NDVI-NDBI correlation coefficient was much higher during the lockdown period. LST and UTFVI also share a strong positive correlation with a Pearson correlation value of 0.981 and 0.950 in 2009 and 2020, respectively, for grid number 1 and a Pearson correlation value of 0.956 and 0.975 in 2009 and 2020, for grid number 2. This analysis shows a positive impact on our environment as vegetation cover has increased over this period, which in turn increases the negative correlation between LST and NDVI. This was due to the decrease in the LST value and an increase in the NDVI value. It can be assumed that with pollution control as in the course of the lockdown, an improved state of the environment can be achieved, which reinforces the negativity of the coefficient correlation of LST and NDVI.

Validation of results

Figure 14 shows the historical Google Earth image, LST, and NDVI image of grid number 1. This grid covers the high school colony in Kohima Sadar. As seen in the historical image, in 2009, this grid was mainly occupied by open and barren land with sparse vegetation and scattered built-up areas, while in 2020 the built-up area and green cover in this grid have increased, resulting in a decrease in the LST. The minimum and maximum LST in grid number 1 in 2009, was 20.30 °C and 29.91 °C, respectively, with a mean LST of 25.1 °C. There was a decline in the minimum and maximum LST in grid number 1 in 2020 with a minimum and maximum LST of 17.31 °C and 25.35 °C, respectively, with a mean LST of 21.33 °C. There was a 3.8 °C drop in the LST of this grid. The minimum and maximum NDVI were also − 0.16 and 0.59, respectively, with an average NDVI value of 0.21. Also, the minimum and maximum NDVI were at 0.12 and 0.73, respectively, with an average NDVI value of 0.43. From this micro-level analysis, it can be concluded that the closure period has had a positive impact on the environment, as can be seen in the images of grid number 1, where the vegetation cover has increased and the LST has decreased despite the built-up area, mainly due the decrease in pollution levels due to the various restrictions imposed.

Fig. 14
figure 14

Google Earth historical image, LST, and NDVI image of grid number 1 in 2009 and 2020

Full size image

Figure 15 shows the historical Google Earth image, LST, and NDVI imagery of grid number 2. This grid includes the main town area and locations such as Midland, New Market, Daklane, Naga Bazaar, and D Block colonies fall under this grid. Most of this area has been taken up by development, which has become compact over the years. Green coverage has also been observed to increase. The minimum and maximum LST in grid number 2 in 2009, was 31.08 °C and 20.92 °C, respectively, with a mean LST of 25.5 °C. There was a decrease in the minimum and maximum LST in grid number 2 in 2020 having a minimum and maximum LST of 30.42 °C and 16.51 °C respectively with a mean LST of 23.46 °C. There was a 2.03 °C drop in the LST of this grid. The minimum and maximum NDVI were also − 0.43 and 0.55, respectively, with an average NDVI value of 0.06. Also, the minimum and maximum NDVI were 0.09 and 0.72, respectively, with an average NDVI value of 0.41. This analysis indicates a positive change as low LST was reported despite increased development in this grid with increased vegetation cover.

Fig. 15
figure 15

Google Earth historical image, LST, and NDVI image of grid number 2 in 2009 and 2020

Full size image

Conclusion

This study focuses on comparing the impacts of LST, NDVI, NDBI, UTFVI, and LULC on the environmental condition of Kohima Sadar at a kilometer-grid scale. For this study, the Landsat TM satellite image dated November 29, 2009, and the Landsat 8 satellite image dated November 27, 2020, were used. During the lockdown period, the results showed a decrease in the value of LST and an increase in the value of NDVI due to the suspension of transportation, and industrial including business activities. There was an increase in the negativity of the correlation between LST-NDVI and NDVI-NDBI, while a positive correlation was found between LST-NDBI and LST-UTFVI. It has been observed that after the lockdown, the environment becomes less stressed and ecologically more stable compared to before the lockdown. The result shows the positive effects of the lockdown on the land surface of Kohima Sadar and its nearby areas, which in turn is very useful for town and city planners. Because green cover and water bodies are the largest contributors to reducing the land surface temperature in an area, urban planners need to emphasize the method of land transformation, converting the current barren and open land into green space and water bodies. Planting trees along the streets, including the commercial and residential areas is essential to maintain the ecological balance of an area.