p-ISSN: 0972-6268 ature nvironment and ollution echnology N E P T (Print copies up to 2016) Vol. 20 No. 2 pp. 875-880 2021 An International Quarterly Scientific Journal e-ISSN: 2395-3454

Original Research Paper Originalhttps://doi.org/10.46488/NEPT.2021.v20i02.050 Research Paper Open Access Journal A New Approach for Environmental Modelling of LULC Changes in Semi- arid Regions of , , Using Geospatial Techniques

B. Pradeep Kumar*, K. Raghu Babu*†, P. Padma Sree**, M. Rajasekhar* and M. Ramachandra* *Department of Geology, Yogi Vemana University, Kadapa, Andhra Pradesh, India **Department of Geology, Government College (A), Anantapur, Andhra Pradesh, India †Corresponding author: K. Raghu Babu; [email protected]

ABSTRACT Nat. Env. & Poll. Tech. Website: www.neptjournal.com This study aims to define the changing pattern of land use and its geo-environmental impacts on the semi-arid region of Anantapur district of AP state, India. Satellite imageries were analysed to 04-05-2020 Received: perceive the variations in land use and land cover in the past 9 years from 2010 to 2019. RS and GIS 23-07-2020 Revised: modelling has helped in the mapping of land use and land cover changes. The study has assumed five Accepted: 27-08-2020 characteristic features, they are (i) Waterbodies, (ii) Vegetation, (iii) Fallow land, (iv) Cultivation lands, and (v) Degraded lands. The results reveal that, from 2010 to 2019, there is a decrease in water bodies, Key Words: 2 2 2 LULC vegetation and fallow lands of 6.75 km , 42.96 km and 105.45 km respectively. While cultivation lands 2 2 Satellite data and degraded lands increased to 4.7 km and 105.45 km respectively. The environmental ecosystem NDVI is disturbed due to the increase in degraded lands, thus making the study area turn into a desert. Ecosystem degradation Normalized Differential Vegetation Index (NDVI) and Soil Adjusted Vegetation Index (SAVI) are very Remote Sensing and GIS useful for the accuracy assessment of vegetation, cultivation land and waterbodies in this LULC change detection studies.

INTRODUCTION wet and water bodies etc. in simple terms land cover is the result of observation. Land use denotes land cover but differs Geo-environment, not only from the environmental point of in relations to its socio-economic persistence and global use view but also in relative social pecuniary aspect has a specific (Harshika & Sopan 2012). This is in pure contrast with land perception in the human life. The role of the geo-environment cover as stated above which is most expressive and deals with in the aforesaid aspects has been undervalued and we are still physical observations. Land use may differ in nature and far from exploiting it to its full extent. An empathetic of the intensity, with the purpose it dealt with during the physical possessions of the Earth’s resources and their undercurrents observations (Dewan et al. 2009, Anees et al. 2014, Kumar is vital for unravelling difficulties in fluctuating fields, such as et al. 2020). Land use varies from land cover for the reason soil, air, water contamination, soil destruction, waste disposal intentional on and role of people to familiarize the natural and construction provisions and foundations (Senthil 2013, land cover to their assistances. Badapalli et al. 2019). Empathetic geological and geomor- phological processes are also important to other parts of the Geographical Information System (GIS) epitomizes a study including the progression and alleviation of natural commanding set of tools for gathering, filament, recovering and man persuaded threats and jeopardizes, land protection and presenting spatial data from the real world. Remote Sens- and renovation, landscape and urban planning, ecosystem ing (RS) from airborne and spaceborne platforms delivers inventories and natural heritage assessments (Chao et al. appreciated data for mapping, geo-environmental monitoring 2004, Rawat et al. 2013). (Babu et al. 2012). The incorporation of RS and GIS mod- elling technologies in the field of environmental protection Land Use and Land Cover (LULC) are often blended. is inevitable and assists in decision making. This is somewhat logical for the reason that both terms are scarcely related and to more or less scope even overlap. In STUDY AREA this normal state, land cover systematizes a flawless arrival of the environmental equilibrium among parent rock, climatic The present study region falls under semi-arid provinces ailment, soil, and vegetation. Landcover eminent in various of Anantapur district (13°40’ and 15°15’ Northern latitude categories i.e., area of vegetation, bare soil, rock outcrops, and 76°50’ and 78°30’ Eastern longitude) which falls within 876 B. Pradeep Kumar et al. the rain shadow range of , in the interior of Methodology Deccan Plateau (Fig. 1). The district is circumscribed by The data sets were improved in ERDAS Imagine 2014, Kurnool District in the north Kadapa District in the north- satellite image processing software to create a false colour east, Chittoor District in the south-east, and State composite (FCC). The layer stack option in the image in- in the West. Two main rivers flows over the study region, one terpreter toolbox was used to generate FCCs for the study is Hagari/Vedavathi and another one is . There regions. The sub-setting of satellite images were performed are two seasons in the study region, they are South-west for extracting study region from both images by taking monsoon (June-September), and the North-east monsoon geo-referenced outline boundary for better classification (October to December). Hot weather from March to May of LULC through Unsupervised option, NDVI and SAVI and the cold period from January to February. were also adopted to assess the vegetation and water bodies in the classification. DATA AND METHODOLOGY RESULTS AND DISCUSSION Data Acquired LULC Change Detection and Analysis 1. SOI (Survey of India) topo sheet no. 57F/1, 57F/2, 57F/5, 57F/6, 57B/13 and 57B/14. For the execution of LULC change detection, a supervised classification method was engaged. A pixel-based evaluation 2. Landsat imageries of three dissimilar periods i.e.,2010 was cast-off to collect the transformation info on pixel base and 2019 have taken freely from USGS earth explorer and hence, infer the variations more precisely. Classified site (http://earthexplorer.usgs.gov/) and from NRSC site image pairs of the two-fold dissimilar period (2010 to 2019) (http://bhuvan.nrsc.gov.in). with the resolution of 30m information are linked by means of cross-tabulation in order and 15m, datum WGS 1984 and UTM zone 44N. to determine qualitative and quantifiable variation features. A change matrix (Babu et al. 2012, Kumar et al. 2018 & Software Used 2019, Rajasekhar et al. 2019) was shaped with the help of ERDAS imagine 2014 software. Measurable extent data of 1. ERDAS imagine 2014 the complete LULC variations as well as increases and upset 2. ArcGIS in apiece category between 2010 and 2019 were compiled.

FigFig.. 1: Location map. map.

Vol. 20, No. 2, 2021 • DATA AND Nature METHODOLOGY Environment and Pollution Technology Data Acquired

1. SOI (Survey of India) topo sheet no. 57F/1, 57F/2, 57F/5, 57F/6, 57B/13 and 57B/14. 2. Landsat imageries of three dissimilar periods i.e.,2010 and 2019 have taken freely from USGS earth explorer site (http://earthexplorer.usgs.gov/) and from NRSC site (http://bhuvan.nrsc.gov.in). with the resolution of 30m and 15m, datum WGS 1984 and UTM zone 44N.

Software Used

1. ERDAS imagine 2014 2. ArcGIS

Methodology The data sets were improved in ERDAS Imagine 2014, satellite image processing software to create a false colour composite (FCC). The layer stack option in the image interpreter toolbox was used to generate FCCs for the study regions. The sub-setting of satellite images were performed for extracting study region from both images by taking geo-referenced outline boundary for better classification of LULC through UnsupervisedUnsupervised option, NDVIoption ,and NDVI SAVI and were SAVI also were adopted also adopted to assess to theassess vegetation the vegetation and water and bodieswater bodies in the in the classification.classification. RESULTSRESULTS AND DISCUSSION AND DISCUSSION LULC CLULChange C Dhangeetection Detection and Analysis and A nalysis For the executionFor the execution of LULC of changeLULC changedetection, detection, a supervised a supervised classification classification method method was engaged was engaged. A pixel. A- pixel- based evaluationbased evaluation was cast was-off castto -collectoff to collectthe transformation the transformation info on info pixel on basepixel and base hence and , henceinfer, theinfer the variationsvariations more precisely more precisely. Classified. Classified image pairsimage of pairs the oftwo the-fold two dissimilar-fold dissimilar period period(2010 (2010to 2019) to 2019) informationinformation are link edare bylink meansed by ofmeans cross of-tabulation cross-tabulation in order in to order determine to determine qualitative qualitative and quantifiable and quantifiable variationvariation features .features A change. A mchangeatrix (mBabuatrix et (Babu al. 2012 et al., Kumar 2012, Kumaret al. 2018 et al. & 2018 2019 &, Rajasekhar2019, Rajasekhar et al. 201 et al.9) 2019) was shapedwas withshaped the with help the of helpERDAS of ERDAS imagine imagine 2014 software. 2014 software. Measur Measurable extentable extentdata of data the ofcomplete the complete

LULC variationsLULCENVIRONMENTAL variations as well asas increaseswell MODELLINGas increases and upset and OF in LULCupset apiece inCHANGES categorapiece category IN between SEMI-ARIDy between 2010 REGIONS and 2010 2019 and were 2019 compiled. were877 compiled.

Fig. 2: 2010 LULC Fig. 3: 2019 LULC Fig. 2: 2010 LULC Fig. 3: 2019 LULC Accuracy Assessment Fig.of LULC 2: 2010 Using LULC NDVI and SAVI water bodies assessment. Fig. 3: 2019 Both theLULC techniques were used to assess, whether vegetation, cultivation land and waterbodies NDVI and SAVI techniques have been adopted for the extended in the unsupervised classification are identical to accuracy assessment of vegetation, cultivation land, and AccuracyAccuracy Assessment Assessment of LULC of ULULCsing NDVI Using andNDVI SAVIthe and NDVI SAVI and SAVI categorized areas or not. Fig. 4 and waterbodies existing in the study region examination for Fig. 5 depict the NDVI and SAVI respectively. The results NDVIthe LULC andNDVI change SAVI and detection. techniques SAVI NDVI techniques have is involved been have adoptedin beenthe ex adopted- forrevealed the foraccuracy thatthe bothaccuracy assessment the unsupervised assessment of v andegetation, of NDVI, vegetation, SAVIcultivation are c ultivation amination of vegetation and cultivation land, and SAVI for almost the same. land, andland waterbodies, and waterbodies existing existing in the study in the region study examinationregion examination for the forLULC the LULCchange change detection. detection. NDVI NDVIis is involved in the examination of vegetation and cultivation land, and SAVI for water bodies assessment. involved in the examination of vegetation and cultivation land, and SAVI for water bodies assessment. Both the techniques were used to assess, whether vegetation, cultivation land and waterbodies extended Both the techniques were used to assess, whether vegetation, cultivation land and waterbodies extended in the unsupervised classification are identical to the NDVI and SAVI categorized areas or not. Fig. 4 and in the unsupervised classification are identical to the NDVI and SAVI categorized areas or not. Fig. 4 and Fig. 5 depict the NDVI and SAVI respectively. The results revealed that both the unsupervised and NDVI, Fig. 5 depict the NDVI and SAVI respectively. The results revealed that both the unsupervised and NDVI, SAVI are almost the same. SAVI are almost the same.

Fig. 4: NDVI Fig. 5: SAVI Fig. 4: Fig.NDVI 4: NDVI Fig. 5: Fig.SAVI 5: SAVI

Nature Environment and Pollution Technology • Vol. 20, No.2, 2021 NDVI NDVI EstablishedEstablished on the valueson the valuesof NDVI of NDVIand in relationand in relation to the Google to the Google Earth imageries Earth imageries to classify to classify and enumerate and enumerate the vegetationthe vegetation change change during duringthe period the periodof 2010 of and 2010 2019 and ( Fig.2019 2 ( andFig. Fig.2 and 3) Fig., have 3) ,been have applied been applied to five to five classesclasses i.e., Waterbody, i.e., Waterbody, Vegetation, Vegetation, Cultivation Cultivation land, Fallow land, Fallow land, and land, Degraded and Degraded lands. Thislands. was This carried was carried out in ERDASout in ERDAS imagine imagine 2014, with2014, the with formula the formula or combination or combination of bands of givenbands belowgiven. below. NDVI N= DVI(near =-infrared (near-infrared (NIR) band(NIR) – bandRED – band RED)/( bandnear-)/(infrarednear-infrared (NIR) band(NIR) + band RED + band RED) band) ..(1) ..(1) near-infrarednear-infrared (NIR) band(NIR) and band the and RED the band RED are band the arereflectance the reflectance radiated radiated in the NIRin the band NIR 7 bandand the 7 and visible the visible red 5 (0.8red- 1.15 (0.8 μm-1.1 and μ 0.6m and-0.7 0.6 μm)-0.7 waveband μm) waveband of the Landof the sat Land 8. The sat 8.procedur The procedure of thee NDVIof the NDVIin the ERDAS in the ERDAS imagineimagine software software is shown is shownin Fig. in6. Fig. 6.

878 B. Pradeep Kumar et al.

NDVI 2019. Usually, SAVI specifies waterbody, vegetation, and cultivation exposure and health with respect to saturation, Established on the values of NDVI and in relation to the soil colour, and moisture therefore accounts for the high Google Earth imageries to classify and enumerate the veg- inconsistency areas. SAVI correspondingly controls the in- etation change during the period of 2010 and 2019 (Fig. 2 fluence of soil brightness in NDVI and thus, minimizes soil and Fig. 3), have been applied to five classes i.e., Waterbody, brightness related noise in vegetation coverage estimation. Vegetation, Cultivation land, Fallow land, and Degraded Since coverage, brightness, and health of vegetation are lands. This was carried out in ERDAS imagine 2014, with strappingly related to surface permeability, SAVI delivers the formula or combination of bands given below. a significant proxy for the identification of impermeable NDVI = (near-infrared (NIR) band – RED band) surfaces. Calculations of SAVI could be done through the /(near-infrared (NIR) band + RED band) ...(1) formula. near-infrared (NIR) band and the RED band are the reflec- SAVI = ((NIR – RED)/(NIR+RED+L)) * (1+L) …(2) tance radiated in the NIR band 7 and the visible red 5 (0.8- 1.1 μm and 0.6-0.7 μm) waveband of the Land sat 8. The Where RED is the reflectance of band 3 and NIR is the procedure of the NDVI in the ERDAS imagine software is reflectance value of the near-infrared band (Band 4). L is the shown in Fig. 6. soil brightness correction factor. For dense vegetation and highly permeable surface areas, L = 0 and for vegetation SAVI scarce and impermeable surface areas, L = 1. The procedure In the present study, the computed SAVI is used to evalu- of the SAVI in the ERDAS imagine software is shown in ate the variations in surface porousness between 2010 and Fig. 7.

Fig. 6: NDVI bands in ERDAS imagine software. Fig. 6: NDVI bands in ERDAS imagine software.

SAVI

In the present study, the computed SAVI is used to evaluate the variations in surface porousness between 2010 and 2019. Usually, SAVI specifies waterbody, vegetation, and cultivation exposure and health with respect to saturation, soil colour, and moisture therefore accounts for the high inconsistency areas. SAVI correspondingly controls the influence of soil brightness in NDVI and thus, minimizes soil brightness related noise in vegetation coverage estimation. Since coverage, brightness, and health of vegetation are strappingly related to surface permeability, SAVI delivers a significant proxy for the identification of impermeable surfaces. CalculationsFig. 7: of SAVI SAVI bands could in ERDAS be imagine done software.through the formula. SAVI = ((NIR – RED)/(NIR+RED+L)) * (1+L) …(2) Vol. 20, No. 2, 2021 • Nature Environment and Pollution Technology Where RED is the reflectance of band 3 and NIR is the reflectance value of the near-infrared band (Band 4). L is the soil brightness correction factor. For dense vegetation and highly permeable surface areas, L = 0 and for vegetation scarce and impermeable surface areas, L = 1. The procedure of the SAVI in the ERDAS imagine software is shown in Fig. 7. Fig. 7: SAVI bands in ERDAS imagine software. LULC Status

LULC classification results obtained from 2010 to 2019 were shown in Table 1. Waterbody has calculated as 207.06 km2 in the year 2010 and it is decreased to 200.31 km2 in the year 2019. Vegetation is calculated as 793.92 km2 in the year 2010 and it is decreased to 750.96 km2 in the year 2019. There is little increase in Cultivation land, i.e., 671.07 km2 in the year 2010 and in the year 2019 it is increased to 675.77 km2. Fallow land and is calculated as 589.24 km2 in the year 2010 and it is decreased to 483.79 km2 in the year 2019. Degraded land is increased at alarming rates and it is calculated as 524.94 km2 in the year 2010 and it is increased to 675.40 km2 in the year 2019. Fig. 8 shows the graphical representation of LULC and its impact on geo-environmental changes in the percentage during the decade i.e., 2000 to 2019. Table 1: statistical calculation.

2010 2019 Changes from 2000-2019

LULC categories km2 Percentage% km2 Percentage km2 Percentage ENVIRONMENTAL MODELLING OF LULC CHANGES IN SEMI-ARID REGIONS 879 waterbody 207.06 7.43 200.31 7.18 - 6.75 - 0.25 VegetationTable 1: statistical calculation. 793.92 28.49 750.96 26.95 - 42.96 - 1.54 LULC categories 2010 2019 Changes from 2000-2019 Agri land 671.07 24.08 675.77 24.25 4.70 0.17 km2 Percentage% km2 Percentage km2 Percentage Fallowwaterbody land 207.06 589.24 7.43 21.14 200.31483.79 7.18 17.36 - 6.75 -105.45 - 0.25 - 3.78 Vegetation 793.92 28.49 750.96 26.95 - 42.96 - 1.54 DegradedAgri land land 671.07 524.94 24.08 18.84 675.77675.40 24.25 24.24 4.70 150.46 0.17 5.40 Fallow land 589.242,786.23 21.14 100 483.792,786.23 17.36 100 -105.45 - 3.78 Degraded land 524.94 18.84 675.40 24.24 150.46 5.40 2,786.23 100 2,786.23 100

Change detection from 2010 to 2019 in percentage

6 5 4 3 2 1 0 waterbody Vegetation Agri land Fallow land Degraded -1 land -2 -3 -4

Fig. 8: Change detection from 2010 to 2019. Fig. 8: Change detection from 2010 to 2019. LULC Status activities and environmental influences of land-use change in the study region over the past 9 years using an integrated CONCLUSIONLULC classification results obtained from 2010 to 2019 method of Remote Sensing and GIS modelling. Fallouts evi- were shown in Table 1. Waterbody has calculated as 207.06 dently reveal that LULC variations were significant through- km2 in the year 2010 and it is decreased to 200.31 km2 in This study evaluated land use patterns, major improvementout the period activities from 2010 and toenvironmental 2019. There is a influencesnegative sign of the year 2019. Vegetation is calculated as 793.92 km2 in in certain factors like the decrease in waterbodies is noticed the year 2010 and it is decreased to 750.96 km2 in the year land-use change in the study region over the past 9 years(0.25 using %) there an also integrated a prominent method decrease of in vegetation,Remote Sensing fallow 2019. There is little increase in Cultivation land, i.e., 671.07 land (1.54 % and 3.78 % respectively). Positive sign observed andkm GIS2 in the modelling. year 2010 andFallouts in the yearevidently 2019 it revealis increased that toLULC variations were significant throughout the period in Cultivation land with an increase of 0.17%. Degraded land 675.77 km2. Fallow land and is calculated as 589.24 km2 is increased at an alarming rate (5.4%) and it is a negative fromin the 2010 year to2010 2019. and Thereit is decreased is a negative to 483.79 sign km 2in in certain the factors like the decrease in waterbodies is noticed sign for the environmental ecosystem immanence. Most of year 2019. Degraded land is increased at alarming rates (0.25 %) there also a prominent decrease in vegetation,the peoplefallow living land in (1.54 these %villages and 3.78are getting % respectively migrated to ). and it is calculated as 524.94 km2 in the year 2010 and it other villages or towns for their livelihood. This may lead to is increased to 675.40 km2 in the year 2019. Fig. 8 shows Positive sign observed in Cultivation land with an increasean environmental of 0.17% ecosystem. Degraded imbalance. land isOur increased results clearly at an the graphical representation of LULC and its impact on reveal this severe condition through the NDVI and SAVI alarminggeo-environmental rate (5.4%) changes and itin isthe a percentagenegative signduring for the the environmental ecosystem immanence. Most of the in relation to that of the unsupervised classification studies. decade, i.e. 2000 to 2019. people living in these villages are getting migrated toGeospatial other villages techniques or townslike RS for and their GIS livelihood.provide pointers This on tools to monitor, approximation, appraise and achieve CONCLUSION may lead to an environmental ecosystem imbalance.supervisory Our results factors clearly on the reveenvironmentalal this severe imperils condition to save This study evaluated land use patterns, major improvement the life and the society.

Nature Environment and Pollution Technology • Vol. 20, No.2, 2021 880 B. Pradeep Kumar et al.

ACKNOWLEDGEMENT change in Greater Dhaka, Bangladesh using remote sensing to promote sustainable urbanization. Applied Geography Journal, 29: 390-401. B. Pradeep Kumar, greatly thankful to the Department of Harshika and Sopan I. 2012. Land use land cover classification and change Science and Technology (DST), Government of India, for detection using high resolution temporal satellite data. Journal of Environment, 1(4): 146-15. financial support through Inspire programme (Sanction order Kumar, B. P., Babu, K. R., Rajasekhar, M. and Ramachandra, M. 2019. No. DST/INSPIRE Fellowship/2017/IF170114). Also thank- Assessment of land degradation and desertification due to migration of ful to USGS for remote sensing data utilization, Department sand and sand dunes in Beluguppa Mandal of Anantapur district (AP, of Geology, Yogi Vemana University, for necessary facilities India), using remote sensing and GIS techniques. J. Ind. Geophys. Union, 23(2): 173-180. for carrying out the research work. Kumar, B. P., Raghu Babu, K., Rajasekhar, M., Narayana Swamy, B. and Ramachandra, M. 2019. Landuse/landcover changes and REFERENCES geo-environmental impacts on Beluguppa Mandal of Anantapur District of Andhra Pradesh, India, using remote sensing and GIS Anees, M. T., Javed, A. and Khanday, M. Y. 2014. Spatio-temporal land modelling. Research & Reviews: Journal of Space Science & cover analysis in Makhawan Watershed (MP), India through remote Technology, 8(2): 6-15. sensing and GIS techniques. Journal of Geographic Information Kumar, B.P., Babu, K.R., Rajasekhar, M. et al. 2020. Identification of land System, 6(4): 298-306. degradation hotspots in semiarid region of Anantapur district, Southern Babu, K. R. and Raju, G. S. 2012. Application of remote sensing for India, using geospatial modelling approaches. Model. Earth Syst. delineation of uranium bearing Vempalli dolomites in and around Environ. 6: 1841-1852. https://doi.org/10.1007/s40808-020-00794-x. Tummalapalli area, Cuddapah Basin, India. International Journal of Rajasekhar, M., Raju, G.S. and Raju, R.S. 2019. Data on artificial recharge Geomatics and Geosciences, 2(3): 842-852. sites identified by geospatial tools in semi-arid region of Anantapur Badapalli, P. K., Raghu Babu, K., Rajasekhar, M. and Ramachandra, M. District, Andhra Pradesh, India. Data Brief, 19: 462-474. 2019. Assessment of aeolian desertification near Vedavathi river Rawat, J. S., Biswas, V. and Kumar, M. 2013. Changes in land use/cover channel in Central part of Andhra Pradesh: Remote Sensing Approach, using geospatial techniques: A case study of Ramnagar town area, Remote Sensing of Land, 3(1): 39-49. district Nainital, Uttarakhand, India. The Egyptian Journal of Remote Chao, H.A.N. 2004. Some research advances and methods on detecting land Sensing and Space Science, 16(1): 111-117. cover change by remote sensing. Journal of Arid Meteorology, 22(2): Senthil, S., Arivazhagan, S. and Rangarajan, N. 2013. Remote sensing and 76-81. GIS applications in environmental science. J. Environ. Nanotechnol, Dewan, M. Ashraf, Yamaguchi Yasushi, 2009. Land use and land cover 2: 292-101. doi:10.13074/jent.2013.06.132025.

Vol. 20, No. 2, 2021 • Nature Environment and Pollution Technology