Assessment of LULC Changes for Hesaraghatta Watershed Using GIS Tools and Remote Sensed Data

Hesaraghatta watershed is one of the most vital and environmentally substantial watersheds in the Arkavathi basin. It has a freshwater lake created in the year 1894 across the Arkavathi River to quench the drinking water requirements of Bengaluru city. The watershed is facing significant stresses due to rapid urbanization and other developmental activities. For this study, an attempt is made to assess the distribution of various land use land cover classes and their temporal changes over 18 years using remote sensed data and GIS tools. The watershed is categorized into four land use land cover classifications: settlement, waterbody, vegetation, and bare soil. The maximum likelihood technique is utilized for the image classification and accuracy assessment is carried out to evaluate the accuracy of image classification. The outcome of the study revealed that there is a substantial change in land use land cover classes in the Hesaraghatta watershed.


INTRODUCTION
From the start of human civilization, humankind has been in a close association with nature and its reliance on the environment is greater than any other living thing , Wang et al. 2018. Due to the intensification in population and human intervention, there is a substantial modification in Land use (LU) and Land cover (LC) with time. In general, "land cover refers to the physical land kind such as how much of a region is enclosed by forests, impervious surfaces, agricultural lands, wetlands, and open water, whereas land use documents how people are using the land for development, conservation or mixed uses (Chowdhury et al. 2020). LULC changes greatly affect the regional ecological process (Chowdhury et al., 2020), (Belay & Mengistu 2019). Water quality, runoff and groundwater recharge, etc., gets affected due to changes in LULC in a watershed (Butt et al. 2015, Garg et al. 2019. Thus, it becomes necessary to understand the spatial relationship and mechanisms between ecosystem and LULC , Imran Basha et al. 2018, Salunke & Chavan 2020 and to measure the changes in LULC pattern at the watershed level for planning management strategies (Nayak & Byrne 2019, Preethi et al. 2020, L. Wang et al. 2020). Assessment of LULC change helps to identify the amount of anthropological impact on the natural environment (Chowdhury et al. 2020). Change detection is a method to identify the condition of a component or phenomenon by perceiving it at a different time (Bharath et al. 2020a, Imran Basha et al. 2018. Change detection helps to identify LULC changes like landscape changes, shifting cultivation, land degradation, desertification, urban sprawl, coastal change, deforestation, quarrying activities, and other cumulative changes (Bharath et al. 2020b, Gebeyehu et al. 2019. Traditional approaches are not satisfactory for the analysis of multi-complex environmental investigations because of the difficulty of dealing with the multi-disciplinary dataset; hence we need new advances like remote sensing and Geographical Information Systems (GIS) (Mallupattu & Sreenivasula Reddy 2013). GIS and Remote sensing are the most efficient approaches for observing widespread and rapid LULC changes , Tsai et al. 2019. Remotely sensed data is freely accessible, and developments in spatial, spectral, and temporal resolutions continue to provide tools for detecting changes in LULC at various scales. Landsat satellite systems provide open-access, long-term imageries that are ideal for assessing LULC change (Tsai et al. 2019).
The primary objective of this LULC study is to determine the extent of LULC changes that arisen in the Hesaraghatta Watershed from 2000 to 2018. But the definite objectives are delineating the Hesaraghatta watershed, detecting the changes in LULC classes within the watershed by spatial comparison, and analyzing the capability of integrating GIS with Remote sensing to study the spatial distribution of LULC classes.

Study Area
Hesaraghatta watershed is situated in Bengaluru rural district, and geographically it lies in 77° 20¢ and 77°42¢ E longitude and 13°10¢ and 13° 24¢N latitude as shown in Fig. 1. Hesaraghatta watershed is one of the most vital and environmentally substantial watersheds. It has a freshwater lake created in the year 1894 across the Arkavathi River Arkavathi River to quench the drinking water requirements of Bengaluru city. The watershed spreads over an area of about 605.2 km 2 with a width and length of 27.28 km and 33.36 km, respectively (Somashekar et al. 2011).

Data Acquisition
In this study, Cartosat-1: DEM -Version-3R1 collected from the Bhuvan website (https://bhuvan.nrsc.gov.in/) is used for watershed delineation. Satellite images of two different years comprising of multi-spectral data procured via Landsat provided by the United States Geological Survey (USGS) Earth Explorer (EarthExplorer n.d.) are used for visual image interpretation and detecting LULC classes. Images collected are of the same months to have minimal seasonal variation (Chowdhury et al. 2020 The primary objective of this LULC study is to determine the extent of LULC changes that arisen in the Hesaraghatta Watershed from 2000 to 2018. But the definite objectives are delineating the Hesaraghatta watershed, detecting the changes in LULC classes within the watershed by spatial comparison, and analyzing the capability of integrating GIS with Remote sensing to study the spatial distribution of LULC classes.

Study Area
Hesaraghatta watershed is situated in Bengaluru rural district, and geographically it lies in 77° 20′ and 77°42′ E longitude and 13°10′ and 13° 24′N latitude as shown in Fig

Data Acquisition
In this study, Cartosat-1: DEM -Version-3R1 collected from the Bhuvan website (https://bhuvan.nrsc.gov.in/) is used for watershed delineation. Satellite images of two different years comprising of multi-spectral data procured via Landsat provided by the United States Geological Survey (USGS) Earth Explorer (EarthExplorer n.d.) are used for visual image interpretation and detecting LULC classes. Images collected are of the same months to have minimal seasonal variation (Chowdhury et al. 2020

Watershed Delineation
Watershed delineation is a process of marking a boundary that signifies the area contributing to an outlet (Chowdhury et al. 2020, Rimal et al. 2019. In this research, Cartosat-1 DEM of the study area with a spatial resolution of 30m and QGIS software is used for delineating the watershed. Fig.  3 highlights the process flow diagram for delineating the boundary of the watershed.

Image Pre-Processing & Image Classification
Before image classification, satellite images are enhanced using histogram equalization to increase accuracy (Mallupattu & Sreenivasula Reddy 2013). The resultant satellite images are clipped using the Hesaraghatta watershed boundary. A supervised image classification technique is used for image classification. In supervised classification, training sites are defined by the user based on the pixel signature (Chowdhury watershed. Fig. 3 highlights the process flow diagram for delineating the boundary of the watershed.

Accuracy Assessment
It is a process that compares the classified pixels with the ground-truth points to assess how well the classified image signifies the real world (Chowdhury et al. 2020

LULC Change Assessment
After image classification, LULC maps of the years 2000 and 2018 are obtained, which gives information about LULC distribution in the study area. Further, the classified images of two time periods are compared to estimate LULC changes over time (Chowdhury et al. 2020).

RESULTS AND DISCUSSION
Hesaraghatta watershed is delineated using Cartosat-1 DEM of 30m resolution by the QGIS tool. The watershed spreads over an area of 605.2 km 2 . By applying the supervised image classification technique, Landsat data for the years 2000 and 2018 are classified under four LULC classes. The outcomes show that in the year 2000, an area of 98.3 km 2 was covered with vegetation, 1.1 km 2 with water, 431.1 km 2 with bare soil, and 74.6 km 2 with settlement, whereas in the year 2018, 125.6 km 2 of the area was covered with vegetation, 2.5 km 2 with water, 153.3 km 2 with settlement, and 125.6 km 2 with bare soil (Table 2)  LULC classes in the Hesaraghatta watershed for the years 2000 and 2018. In the results, bare soil includes both barren lands as well as agricultural fallow land because the study area falls under the semi-arid zone. Most of the agricultural activities depend on rainfall only whereas the satellite data which is considered is of the summer season..  waterbody increased from 1.1 km 2 to 2.5 km 2, and bare soil reduced from 431.1 km 2 to 323.8 km 2 . In between these years, urbanization and industrialization have happened at a very greater rate i.e., 105.54% in the study area due to population growth and developmental activities as it is very nearer to Bengaluru city. There has been a significant decline in the bare soil i.e., 24.89% because most of the land was converted into plantation and settlement. Table 3   and Kappa co-efficient for those years are 0.909 and 0.822 respectively. The Kappa coefficient values obtained portray that the accuracy of the image classification is acceptable.

CONCLUSION
The current study aimed at assessing the change in LULC classes in the Hesaraghatta watershed. This study discloses that the LULC pattern in the Hesaraghatta watershed has transformed significantly over 18 years. It is detected that there has been an escalation in the settlement in the study area due to population growth and developmental activities and it is also noticed that there is an increase in vegetation due to an increase in plantation, thereby reducing the amount of bare land. This study has also revealed that integration of GIS and Remote sensing can be a powerful tool for mapping and assessment of LULC changes in an area. Though, the accuracy of results is greatly influenced by interpretation.

CONCLUSION
The current study aimed at assessing the change in LULC classes in the Hesaraghatta watershed. This study discloses that the LULC pattern in the Hesaraghatta watershed has transformed significantly over 18 years. It is detected that there has been an escalation in the settlement in the study area due to population growth and developmental activities and it is also noticed that there is an increase in vegetation due to an increase in plantation, thereby reducing the amount of bare land. This study has also revealed that integration of GIS and Remote sensing can be a powerful tool for mapping and assessment of LULC changes in an area.
Though, the accuracy of results is greatly influenced by interpretation.