Land Cover Change and the Relationship Between Land Surface Temperature (LST) and Normalized Difference Vegetation Index (NDVI)

-Priyadarshini Acharya,
Student of AGSRT

INTRODUCTION

In recent decades, the Earth's land cover has undergone significant transformations due to various anthropogenic and natural factors. These changes in land cover, such as urbanization, deforestation, agricultural expansion, and climate change, have profound implications for ecosystems, biodiversity, and the climate system. Monitoring and understanding these changes are crucial for sustainable land management and environmental conservation efforts. One of the key indicators used to monitor land cover changes and assess ecosystem health is the Normalized Difference Vegetation Index (NDVI). NDVI is a remote sensing-derived metric that quantifies the density and health of vegetation cover based on the difference in reflectance between near-infrared and red light bands. High NDVI values indicate dense and healthy vegetation, while low values suggest sparse or stressed vegetation. Another important aspect of land surface dynamics is the Land Surface Temperature (LST). LST is a measure of the radiative temperature of the Earth's surface as sensed from satellite or ground-based instruments. It is influenced by various factors such as solar radiation, land cover type, moisture content, and surface properties.

The relationship between NDVI and LST is complex and interdependent. Vegetation cover influences the surface energy balance by absorbing solar radiation for photosynthesis and reducing surface temperatures through evapotranspiration. Consequently, areas with higher vegetation density tend to have lower LST compared to bare or urban surfaces, which absorb more solar radiation and exhibit higher temperatures. Understanding the relationship between NDVI and LST is essential for a range of applications, including land use planning, ecosystem monitoring, climate modeling, and urban heat island studies. Analyzing how changes in land cover affect both NDVI and LST provides insights into ecosystem dynamics, land-atmosphere interactions, and the impacts of human activities on the environment.

In this study, we aim to investigate land cover changes using remote sensing data and analyze the relationship between NDVI and LST in [study area]. We will utilize satellite imagery and geographic information system (GIS) techniques to assess temporal trends in land cover dynamics and quantify the correlations between NDVI and LST. By examining these relationships, we seek to enhance our understanding of ecosystem dynamics and inform sustainable land management strategies in the face of environmental change.

Study area:

Telangana is located on the Deccan Plateau and lies in the Southern region of India. The State is strategically located in the central stretch of the eastern seaboard of Indian Peninsula. The State is bordered by the States of Maharashtra to the North and North-West and Chhattisgarh to the North, Karnataka to the West, and Andhra Pradesh to the South, East and North- East .It falls on co-ordinates like 18.11240 N, 79.01930E (latitude , longitude). In Telangana, land surface temperature (LST) and Normalized Difference Vegetation Index (NDVI) exhibit significant spatial and temporal variations, influenced by the region's diverse geography, climate patterns, land use practices, and socio-economic factors. The relationship between LST and NDVI in Telangana reflects the intricate interplay between land cover dynamics, vegetation health, and surface temperature regulation .

[Map -1 : Study area; Telangana (some Taluk part)]

Stages and Method :

[Fig 2: Model of steps]

Literature review :

Urbanization transforms the natural land surfaces to modern land use and land cover such as buildings, roads and other impervious surfaces, making urban landscapes fragmented and complex and affecting the inhabitability of cities . The enormous changes with land use/cover lead to the urban heat island (UHI). The urban temperatures are 2-5 higher than those in rural surroundings . (Jing Jianga, Guangjin Tiana,2010).

In recent years, the research on the applications of thermal remote sensing of urban areas were mainly in following respects: study on the land surface temperature or the spatial structure of urban thermal patterns and their relationship with surface parameters; urban surface energy balances and fluxes. The relationship between atmospheric temperature and land surface temperature. Some studies examined the effect of land use/land cover change on LST[6-7,15], which was found to be positively correlated with impervious surface.

Some studies estimated the relationship between the land surface temperature and vegetation abundance; different vegetation indices such as normalized vegetation index (NDVI) and fractional vegetation cover were used to indicate vegetation abundance. The results revealed the negative correlation between land surface temperature and NDVI, and the cooling effect of green areas. The changes of Land surface temperature were related to many factors, including changes in land use, land surface parameters, seasonal variation, climatic condition and economic development, etc.(Jing Jianga).

To study the effect of land use/land cover change on land surface temperature, we should choose the Landsat TMˈETM+ images dated on roughly the same season . LST (Land Surface Temperature) is the earth surface temperature which is directly in contact with the measuring instrument (usually measured in Kelvin). LST is the surface temperature of the earth’s crust where the heat and radiation from the sun are absorbed, reflected and refracted. LST changes with a change in climatic condition and other human activities where the exact prediction becomes challenging. Worldwide urbanization has significantly increased in greenhouse gasses and reshaped the landscape, which has important climatic implications across all scales due to the simultaneous transformation of natural land cover and introduction of urban materials i.e. anthropogenic surfaces.

Ground surveys would permit a highly accurate Land Use Land Cover (LULC) classification, but they are time-consuming, burdensome and expensive, which highlights remote sensing as an evident and preferred alternative. Identification and characterization of Urban Heat Island (UHI) is typically based on LST that varies spatially, due to the non-homogeneity of land surface cover and other atmospheric factors. LST is the key factor for calculating the highest and lowest temperature of a particular location.( Anand Babu D, Purushothaman B M, Dr. S. Suresh Babu) . On the other hand, the normalized difference vegetation index (NDVI) has been used to measure the presence and dynamics of vegetation such as the green leaf area index, vegetation cover, green biomass and vegetation productivity. It indicates the vegetation condition and predicts the productivity of plants in several areas of the world. NDVI works on the principle of electromagnetic radiation in which the greenness portion of the vegetation shows less reflectance in the visible spectrum because of the absorption of photosynthetic pigments. Consequently, it has a maximum reflectance in the near-infrared region.( Gbenga F. Akomolafe& Rusly Rosazlina,2022)

Land cover land use :

1.LULC of year 2013 :

LULC stands for Land Use/Land Cover. It refers to the classification and categorization of the Earth's surface based on its physical and human characteristics. Land use refers to how land is utilized by humans, such as for residential, agricultural, industrial, commercial, or recreational purposes . Land cover refers to the physical material or biological cover of the Earth's surface, such as forests, grasslands, wetlands, water bodies, barren land, etc. Here the considered part is LULC changes of the year 2013,it shows vegetation, Agricultural land, Urban, barren land and water body . It was done by ArcGIS Pro software, an unsupervised classification method called Iso Cluster was applied .

Map-2[LULC map of 2013]

Unsupervised and Supervised classifications were adopted together to extract detailed land use and land cover classes, including vegetation, agricultural land, urban, barren land and water body . The classification scheme, with 25 training samples of each class are given. A maximum likelihood classifier was employed for preliminary classifications based on the results of the unsupervised classifications. If you give a look at the area then, vegetation 281914.534 hr, Agricultural land 591156.15 hr., Urban 17091.39, Barren land 431859.61hr and waterbody 40217.19hr covers . so here we can assume that most of the land is agricultural land .

To improve the accuracy of interpretation, some areas which were difficult to distinguish were inspected by ourselves or confirmed with the help of various maps and data. Finally , the overall accuracy of 2013 is 88% and kappa is above 0.8.

These changes in spatiotemporal patterns of different LULC (Land Use Land Cover) categories are depicted using thematic maps (map-2)

[Fig:3;Accuracy assessment of year 2013]

2.LULC of year 2020 :

The year consideration is LULC changes of the year 2020,it shows vegetation, Agricultural land, Urban, barren land and water body . It was done by Arcgis software, an unsupervised classification method called Iso Cluster was applied .( Follow map-3) The classification scheme, with 25 training samples of each class are given. A maximum likelihood classifier was employed for preliminary classifications based on the results of the unsupervised classifications.

If you give a look at the area then, vegetation 448413.16hr, Agricultural land 456208.98hr, Urban 372419.08, Barren land 35114.53hr and water body 51094.77hr covers . so here we can assume that most of the parts are water bodies .

To improve the accuracy of interpretation, some areas which were difficult to distinguish were inspected by ourselves or confirmed with the help of various maps and data.

Finally , the overall accuracy of the 2020 is 84% and kappa is 0.8.(fig-4)

[Map: 3; LULC map of 2020]

[Fig:4; Accuracy assessment of year 2020]

Change detection and map of year 2013-2020:

Change detection in Land Use and Land Cover (LULC) refers to the process of identifying and analyzing changes that occur in the land surface over time. It involves comparing different sets of remotely sensed imagery or other geospatial data acquired at different time periods to detect changes in land use or land cover categories. The main objective of change detection in LULC is to identify and quantify changes such as urban expansion, deforestation, agricultural expansion, and other land use transformations. This information is valuable for various applications including urban planning, environmental monitoring, natural resource management, and climate change assessment.

There are several methods and techniques used for change detection in LULC analysis, including: Image differencing, Classification comparison, Post-classification comparison etc.

The primary objective of this study was to assess the conversion of a specific land use land cover class into other categories. The findings of the change-detection analysis, along with the change matrix table spanning from 2013 to 2020, are provided in the subsequent section .

Here we have taken 2013 to 2020 change detection of class vegetation , agricultural land , urban and water bodies .

First step of change detection was convert the raster file to polygon , then for making all same values together, will go for dissolve and then did intersection for both the area, then we can detect the changes of the classes .

[Fig:5;Changes of classes of 2013-2020]

[Map: 4; LULC Change detection thematic map of 2013 to 2020]

In this map we can see most of the agricultural land changes to urban , Barren land also changes to urban ,and vegetation is the same as well as some parts it changes to barren land and urban itself etc.

Analysis of LST and NDVI & its correlationship

1.Land surface temperature (LST) :

Land cover change refers to the alteration of the physical attributes of the Earth's surface, including changes in vegetation, urbanization, deforestation, and agricultural practices. These changes can have significant impacts on various environmental factors, including land surface temperature (LST). Land surface temperature (LST) is defined as a measure of how hot or cold the surface of the Earth would feel to the touch (Guillevic et al., 2018). So many objects are there which play a significant role in influencing land surface temperature .

Water bodies, such as lakes, rivers, and oceans, play a significant role in influencing land surface temperature (LST) through several mechanisms; like Moderating effect , humidity etc . Barren land, characterized by the absence or minimal presence of vegetation and organic cover, can have significant influences on land surface temperature (LST) due to several factors like high albedo , reduce soil moisture , limited evaporation etc. Vegetation exerts a significant influence on land surface temperature (LST) through various biophysical processes and feedback mechanisms: evapotranspiration , high absorption and release ,etc. Like this all the classes which play significant role influencing LST .

So here the aim is to show the land surface temperature of the study area (taluk part of Telangana) .

1.Top of Atmosphere (TOA) Radiance:
Using the radiance rescaling factor, Thermal Infra-Red Digital Numbers can be converted to TOA spectral radiance.
Lλ = ML * Qcal + AL
Where:
Lλ = TOA spectral radiance (Watts/ (m2 * sr * μm))

ML = Radiance multiplicative Band (No.)
AL = Radiance Add Band (No.)
Qcal = Quantized and calibrated standard product pixel values (DN)
2.Top of Atmosphere (TOA) Brightness Temperature:
Spectral radiance data can be converted to top of atmosphere brightness temperature using the thermal constant Values in Metadata file.
BT = K2 / ln (k1 / Lλ + 1) - 272.15
Where:
BT = Top of atmosphere brightness temperature (°C)

Lλ = TOA spectral radiance (Watts/( m2 * sr * μm))

K1 = K1 Constant Band (No.) K2 = K2 Constant Band (No.)
3.Normalized Differential Vegetation Index (NDVI):
The Normalized Differential Vegetation Index (NDVI) is a standardized vegetation index which Calculated using Near Infra-red (Band 5) and Red (Band 4) bands.
NDVI = (NIR – RED) / (NIR + RED)
Where:
RED= DN values from the RED band

NIR= DN values from Near-Infrared band
4.Land Surface Emissivity (LSE):

Land surface emissivity (LSE) is the average emissivity of an element of the surface of the Earth calculated from NDVI values.
PV = [(NDVI – NDVI min) / (NDVI max + NDVI min)]^2
Where:
PV = Proportion of Vegetation

NDVI = DN values from NDVI Image
NDVI min = Minimum DN values from NDVI Image

NDVI max = Maximum DN values from NDVI Image
E = 0.004 * PV + 0.986
Where:
E = Land Surface Emissivity

PV = Proportion of Vegetation
5.Land Surface Temperature (LST):

The Land Surface Temperature (LST) is the radiative temperature Which calculated using Top of atmosphere brightness temperature, Wavelength of emitted radiance, Land Surface Emissivity.
LST = (BT / 1) + W * (BT / 14380) * ln(E)
Where:
BT = Top of atmosphere brightness temperature (°C)

W = Wavelength of emitted radiance
E = Land Surface Emissivity

2.NDVI(Normalized Differential Vegetation Index:

NDVI, or Normalized Difference Vegetation Index, is a widely used remote sensing vegetation index that quantifies the density and health of vegetation cover on the Earth's surface. NDVI is a valuable tool for monitoring changes in vegetation over time and is often used in agriculture, forestry, environmental science, and land management applications. The NDVI formula is calculated using the reflectance values of near-infrared (NIR) and red (R) bands of electromagnetic radiation obtained from remote sensing instruments such as satellites or drones.

Result and analysis

1.NDVI;

[Map:6;Land Surface temperature (LST) of 2013 and 2020]

In the year of 2013 , the land surface temperature was 13.640c to 37.130c . In the year 2020 , the land surface temperature was 19.670c to 39.210c.
The results of this study showed that LST changes according to the different land use and spatial changes in LST arose from changes in land use; in particular, the extent of bare land increased the land surface temperature of the study region and vegetation cover decreased LST and Urban is also the main reason to increased LST .

Correlation between Lst and NDVI:

The relationship of LST with elevation and NDVI was examined using correlation analysis. The results indicated that LST decreased from North-South and South-East, while increasing from North-East and South-West directions. The correlation coefficient between LST and elevation was negative, with an R-value of −0.51..

The relationship between LST and NDVI can be complex and varies depending on factors such as land cover type, weather conditions, and spatial scale. However, some general relationships include:
Inverse Relationship: In many cases, there is an inverse relationship between LST and NDVI. Higher NDVI values often correspond to lower LST values and vice versa. This is because dense vegetation cover tends to cool the surface by shading and transpiration, whereas bare soil or urban areas absorb more solar radiation, leading to higher temperatures.
Seasonal Trends: LST and NDVI typically show seasonal trends. For example, during the growing season, when vegetation is lush and active, NDVI tends to be higher while LST tends to be lower due to the cooling effect of vegetation. Conversely, during dry or dormant seasons, NDVI decreases and LST increases.
Local Variations: Local variations in land cover, soil moisture, and microclimate can influence the relationship between LST and NDVI. For instance, in urban areas, the urban heat island effect may lead to higher LST values even in areas with vegetation, impacting the usual relationship between LST and NDVI.

[Fig:6;Correlation graph between LST & NDVI of year 2013&2020]

Conclusion

By using Arcmap ,Unsupervised and Supervised classifications were adopted together to extract detailed land use and land cover classes, including vegetation, agricultural land, urban, barren land and water body . The classification scheme, with 25 training samples of each class are given. A maximum likelihood classifier was employed for preliminary classifications based on the results of the unsupervised classifications for the study area (Telangana) of the year 2013 and 2020 .
If you give a look at the area on LULC of 2013 then, vegetation 281914.534 hr, Agricultural land 591156.15 hr, Urban 17091.39, Barren land 431859.61hr and waterbody 40217.19hr covers . so here can assume that most of the part are agricultural land and if we see 2020 years LULC map then, vegetation 448413.16hr, Agricultural land 456208.98hr, Urban 372419.08, Barren land 35114.53hr and water body 51094.77hr covers . So here we can assume that most of the parts are water bodies .The change detection shows an increase of urban areas in the agricultural land and barren land .
By using this data ,NDVI range and LST range is shown i.e. both the years NDVI showing normal land (water body ,tank ) to sparse vegetation .On the same case LST is showing 130C to 370C (2013) and 190C to 390c (2020) because of gradually increasing of urban and for humidity .
Overally the Correlation between LST & NDVI is influenced due to local variations and seasonal trends.

Reference

Jing Jianga , Guangjin Tiana ;Analysis of the impact of Land use/Land cover change on Land Surface Temperature with Remote Sensing 2010;571–575
Owen, T. W., Carlson, T. N., & Gillies, R. R. Assessment of satellite remotely-sensed land cover parameters in quantitatively describing the climatic effect of urbanization. International Journal of Remote sensing 1998;19:1663-1681.
Anand Babu D , Purushothaman B M, & Dr. S. Suresh Babu, Estimation of Land Surface Temperature using LANDSAT 8 ISSN: 2454-132X
Naço P, et al., Automated change detection from remote sensing data: A case study at the Pali Cape - Erzeni river mouth coastal sector, BALWOIS, 2008.
Dr. S. Narayana Reddy, et al., “Land Surface Temperature Retrieval from LANDSAT data using Emissivity Estimation” Vol 12, no 20, pp 9670-9687.
Gbenga F.Akomolafe & Rusly Rosazlina Land use and land cover changes influence the land surface temperature and vegetation in Penang Island, Peninsular Malaysia 2022; 12:21250
Wang, J., Wang, K. & Zhang, C. Z. Impacts of climate change and human activities on vegetation cover in hilly southern China. Ecolog. Eng. 81, 451–461 (2015).
Omran, E. S. E. Detection of land-use and surface temperature change at different resolutions.
J. Geograph. Inform. Syst. 4, 189–203 (2012)