Monitoring Land Use and Land Cover Change of Forest Ecosystems of Shendurney Wildlife Sanctuary, Western Ghats, India

Asian Journal of Environment & Ecology

15(4): 20-27, 2021; Article no.AJEE.70507 ISSN: 2456-690X

H. Bilyaminu1,2*, P. Radhakrishnan1, K. Vidyasagaran1 and K. Srinivasan1

1Kerala Agricultural University, India. 2Kano University of Science and Technology, Wudil, Nigeria.

Authors’ contributions

This work was carried out in collaboration among all authors. All authors read and approved the final manuscript.

Article Information

DOI: 10.9734/AJEE/2021/v15i430234 Editor(s): (1) Dr. Ravi Kant Chaturvedi, Chinese Academy of Sciences, P. R. China. Reviewers: (1) Ranjita Sawaiker, PES’S RSN College of Arts and Science, India. (2) Romanus Udegbunam Ayadiuno, University of Nigeria, Nigeria. Complete Peer review History: http://www.sdiarticle4.com/review-history/70507

Received 01 April 2021 Original Research Article Accepted 06 June 2021 Published 09 July 2021

ABSTRACT

Understanding forest degradation due to human and natural phenomena is crucial to conserving and managing remnant forest resources. However, forest ecosystem assessment over a large and remote area is usually complex and arduous. The present study on land use and land cover change detection of the Shendurney Wildlife Sanctuary forest ecosystems was carried out to utilize the potential application of remote sensing (RS) and geographic information system (GIS). Moreover, to understand the trend in the forest ecosystem changes. The supervised classification with Maximum Likelihood Algorithm and change detection comparison approach was employed to study the land use and land cover changes, using the Landsat Enhanced Thematic Mapper (ETM±) and Landsat 8 OLI-TIRS using data captured on July 01, 2001, and January 14, 2018. The study indicated the rigorous land cover changes. It showed a significant increase in the proportion of degraded forest with negligible gain in the proportion of evergreen forest from 21.31% in 2001 to 22.97% in 2018. A substantial loss was also observed in moist deciduous from 27.11 % in 2001 to 17.23 % in 2018. The result of the current study indicated the degree of impacts on forests from the various activities of their surroundings. This study provides baseline information for planning and sustainable management decisions. ______

*Corresponding author: E-mail: [email protected];

Bilyaminu et al.; AJEE, 15(4): 20-27, 2021; Article no.AJEE.70507

Keywords: Land use and land cover; change detection; forest ecosystem; forest degradation.

1. INTRODUCTION forest ecosystems and the water body of the shendurney wildlife sanctuary were exposed. Climatic and anthropogenic factors are nowadays considered the most critical factors 2. MATERIALS AND METHODS that lead to the degradation and fragmentation of forest ecosystems. Fragmentation is a dynamic 2.1 Study Area process that gradually reduces habitat into smaller patches that became increasingly This study was conducted in Shendurney Wildlife isolated and vulnerable to edge effects [1]. This Sanctuary, located between the geographical phenomenon leads to the separation of habitat locations of 76º 59’ 30” and 77º 16’ 30” East and impairing the potential ecosystem function, Longitude and 8º 44 ‘ and 9º 14’ North Latitude in which imperils plant species, mammals, and Thenmala, Kollam district of Kerala state. The birds. Assessing the forest ecosystem structure sanctuary is part of the Agasthyamalai Biosphere over a large and remote area is usually complex Reserve, one of the Western Ghats' most and arduous. Still, GIS provides essential biodiverse areas. The sanctuary has notified information to model multiple-use forest area of 171sq.km with well-defined natural management decisions. The knowledge of boundaries. The vegetation was classified into remote sensing and geographic information West Coast Tropical Evergreen Forest, West systems (GIS) are the modern tools for the Coast Tropical Semi-Evergreen Forest, Southern assemblage and manipulation of remotely Hilltop Tropical Forest, and Secondary Moist sensed data. The remote sensing imagery of a Mixed Deciduous Forest. All the forest types large variety of space-borne and airborne differ significantly in species composition with a sensors provides vast data about the earth's change in elevations. The study area was surface for detailed global analysis, change classified using selected clear and cloud-free detection, and monitoring [2]. Landsat images: July 01, 2001, and January 14, 2018. The Shendurney Wildlife Sanctuary area is The vegetation classification and mapping entirely contained within the Landsat Path 143 commonly generate a stable descriptive view of and Raw 054. The overall image was rectified, the vegetation resource. Therefore, it is geo-referenced to the Universal Transverse considered significant in driving baseline Mercator (UTM) projection zone 43 and WGS 84 information in ecosystem conservation [3,4]. datum using at least 205 and 94 well-distributed Nevertheless, vegetation is dynamic, and its ground control points and nearest neighbor changes over time are the most crucial resampling. The root means square errors were information for management decisions. The 4.21 m for the 2001 image and 8.41 m for the knowledge of specific vegetation changes helps 2018 image. The image was processed using identify and quantify challenges, set targets, and ARCGIS 10.1 and QGIS 2.18. The land use and assess responses to management actions [4]. land cover mapping were successful by The knowledge of remote sensing and interpreting Landsat Enhanced Thematic Mapper geographic information systems has been (ETM±) satellite images, 2001 generated and utilized to detect and monitor the changes in the Landsat 8 OLI-TIRS images, 2018 generated. forest ecosystem globally. This study examined the land use and land cover change in the 2001- 2.2 Methodology 2018 period using satellite imagery and a geographic information system (GIS) in the 2.2.1 Image processing Shendurney Wildlife sanctuary, Kollam, Kerala. Landsat Enhanced Thematic Mapper (ETM±) 2.2.1.1 Training captured on July 01, 2001 and Landsat 8 OLI- TIRS captured on January 14, 2018, was used to The present study adopted the land cover and achieve the current investigation. The land use classification developed by Anderson Supervised classification with a Maximum [5] to interpret remote sensor data at different Likelihood Algorithm and change detection scales and resolutions. According to Anderson comparison approach was utilized to study the land use and land cover classification scheme, land use and land cover changes. The land-use land use and land cover are categorized as changes that occurred during the period in the different forest land, waterbody, open forest, and

Bilyaminu et al.; AJEE, 15(4): 20-27, 2021; Article no.AJEE.70507

degraded area. An unsupervised image 2.2.3 Change detection classification system was used to assess strata for ground truth before the field visit. Fieldwork The changes in land use and land cover for the was conducted to gather data for training and time intervals were analyzed following the validating land-use and land-cover analyses categorization of imagery from specific years. A based on the 2001 satellite image and for multi-date post-classification comparison change qualitative descriptions of each land-use and detection technique was utilized, which is the cover class's features. A random selection of most extensively used method of detecting testing points was used to create a testing changes. [8]. sample set. 3. RESULTS AND DISCUSSION

2.2.1.2 Allocation 3.1 Forest Ecosystems Change Detection

The ARCGIS 10.1 and QGIS 2.18 software was The land use and land cover (LULC) of the forest used for carrying out the image processing. ecosystems of the Shendurney was studied. The Firstly, the supervised classification with a six land use and land cover classes (LULC) were Maximum Likelihood Algorithm based on the 168 established as evergreen forest, semi-evergreen, training samples, the 2001 image, and 168 moist deciduous forest, hilltop forest, degraded samples for the 2018 images was employed. forest, and open forest (Table 1). The spatial Secondly, the supervised image classification distribution and the estimated change of all the techniques appropriate to Maximum Likelihood forest cover changes were represented in Figs. 2 Classifier (MLC) and 168 training samples were and 3. The present study mapped all the forest applied to produce the land use and cover maps ecosystems and figured out their change using of 2001 and 2018 [6]. Lastly, a 3*3 majority filter Landsat Enhanced Thematic Mapper (ETM±) was utilized to each classification to recode and Landsat-8 OLI-TIRS. These findings isolated pixels classified differently other than the demonstrated that the west coast tropical majority class of the window. evergreen forest occupied the most significant land, followed by the secondary moist deciduous 2.2.2 Accuracy assessment forest (Table 1 and Fig. 1). However, the result indicated that a significant area of Shendurney For accuracy assessments, an error matrix was wildlife sanctuary is occupied by a degraded developed to ensure the consistency of forest, followed by evergreen, semi-evergreen, information obtained from remotely sensed data. and moist deciduous forest (Table 1 and Fig. 1). The sample points were collected and confirmed The insignificant gain in the evergreen forest was by comparing the remote sensing study results to noticed from 3722.02 hectares in 2001 to reference or ground truth data [7]. An 4011.61 ha in 2018, indicating an increased area independent sample of 168 polygons with of 289.53 hectares between the time intervals approximately 100 pixels per polygon was (Fig. 1). On the other hand, a significant loss was randomly selected from each classification to observed in the moist deciduous forest from assess the classification accuracy. Classification 4735.13 hectares in 2001 to 3008 hectares in accuracy was assessed using error metrics such 2018. The semi-evergreen forest showed a as cross-tabulation of mapped class vs. decline from 4699 ha in 2001 to 3313.9 hectares reference class [7]. in 2018.

Table 1. Summary of Land use land cover changes of Shendurney Wildlife Sanctuary for the period of 2001 and 2018

Forests Type 2001 (ha) % 2018 (ha) % Changes West Coast Trop. Evergreen 3722.02 21.31 4011.61 22.97 289.53 West Coast Trop. Semi-Evergreen 4699.02 26.91 3313.9 18.98 -1385.12 Southern Moist Deciduous 4735.13 27.11 3008.82 17.23 -1726.3 Southern Trop. Hilltop 253.48 1.45 341.44 1.96 87.96 Open Forest 1028.50 5.89 2306.96 13.21 1278.46 Degraded Forest 2607.97 14.93 4124.51 23.62 1516.55 Water Body 418.05 2.39 356.93 2.04 -61.11 Total 17464.17 100 17464.17 100 -

Bilyaminu et al.; AJEE, 15(4): 20-27, 2021; Article no.AJEE.70507

2001 (ha) 2018 (ha) Changes

6000.00

5000.00 4699.02 4735.13 4011.61 4124.51 3722.02 4000.00 3313.9 3008.82

e 3000.00 2607.97 r 2306.96 a t c

e 2000.00 1278.46 1516.55 H

r 341.44 1028.50 356.93 e

P 1000.00 87.96 418.05 a 289.53 253.48 e r

A 0.00 Trop. Evergreen Semi-Evergreen Moist Trop. Hilltop Open Forest Degraded Water Body -1000.00 Deciduous Forest -61.11 -1385.12 -2000.00 -1726.3

-3000.00

Fig. 1. Summary of Land use and land cover changes of Shendurney Wildlife Sanctuary for the period of 2001 and 2018

Bilyaminu et al.; AJEE, 15(4): 20-27, 2021; Article no.AJEE.70507

The notable increases in the open forest from increased from 14.93 % in 2001 to 23.62 % in 1028.5 hectares to 2306.36 hectares and hilltop 2018. However, the water body shrunk forest increased from 235.48 hectares to 341.44 considerably from 418.05 hectares in hectares in 2018 were also discovered. 2001 to 356.93 hectares in 2018 (Tab. 1 and Moreover, degraded forests substantially Fig. 1).

Fig. 2. Spatial distribution of the land use and land cover classification 2001

Fig. 3. Spatial distribution of the land use and land cover classification 2018

Bilyaminu et al.; AJEE, 15(4): 20-27, 2021; Article no.AJEE.70507

3.2 Forest Land Cover Changes by the inhabitants. The rapid growth of the human population close to forest ecosystems Land use and cover change have become has increased the risk of degradation and paramount for understanding and proper fragmentation [17]. In Lombok eastern Indonesia, planning of productive ecosystems and Kim [18] reported a significant decrease in the biodiversity, environmental degradation, wetland extent of the forest land for the 20 years interval. deterioration, loss of aquatic organisms, and He presumably attributed the loss by timber wildlife habitat [9]. The Earth Resources extraction, the pressure of land for agriculture Technology Satellite, which was later named and urban development, and poor governance Landsat-1, was launched in early 1972. Satellite institutions [19]. remote sensing provided its potential benefit in assessing, planning, and monitoring natural The changes in land use and land cover are not resources [10,11]. Its ability to provide real-time a random process. Lira et al. [20] studied the data with contemporaneous and repetitive effect of land use and land cover changes coverage provides distinct advantages over (LULC) on the size, shape, and degree of forest conventional methods. There is a common belief patch isolation. Kushwaha [16] reported a that forests in the Western Ghats are gradually marginal increase in the water body area. shrinking because of increasing biotic pressure Contrary to the present study, despite the on the resources. Therefore, the study of land ongoing rehabilitation of the Thenmala Dam, the use and land cover changes (LULC) is significant study reported a decreasing trend in the water for proper planning, managing, and utilization of body. Due to varying precipitation and forest resources. temperature, the size of water bodies can change from year to year. Torahi and Rai [21] Similar to the studies in the Western Ghats have a similar view on water bodies fluctuation in regions [12,13,14,15]. The current study of his study in study. Poor forest management Shendurney wildlife sanctuary revealed that the practices like forest fire management may forest is going through a gradual decrease with increase open and degraded forests in the time. The tremendous decline in the major two wildlife sanctuary. The strata formation of forest forest ecosystems, and a significant increase in ecosystem types of Shendurney rendered it the degraded forest, and the relatively proximate to fragmentation and susceptible to insignificant gain in the tropical evergreen and anthropogenic disturbances. Most of the forest hilltop forest clearly showed the intensity of ecosystems of Shendurney appeared to be in anthropogenic and climatic pressure on most of irregular patches, exceptionally moist deciduous, the forests. Kushwaha [16], in his study, reported semi-evergreen, and the myristica swamp. the 5.66 % overall decrease with zero gain in the According to Ewers and Didham [22], edge forest ecosystem types over twelve years. effects are more common in forest patches with However, the present study report on the irregular shapes than in patches with more considerable increase in the degraded forest, compact shapes. They have been reported to which predicted the entire future of the forest negatively impact many species [22]. ecosystems and its proximity to anthropogenic disturbance and other impacts associated with 4. CONCLUSION human and climatic change. The evergreen and hilltop forests are virtually secured and relatively Monitoring the protected areas for sustainable stable and showed significant increases in their management and conservation is considered extent (Figs. 2 and 3). The stability and crucial. The study of land use and land cover increased in the higher elevation forests could be change is paramount to understand the shift in attributed to the fact that these forests are the forest ecosystems for setting up monitoring located in the core zone areas far from human and planning tools for effective management settlements. Unlike the semi-evergreen and decisions. The land cover of Shendurney Wildlife moist deciduous forests, which are mainly low to Sanctuary is reported changing. The main medium elevation forests and close to human change observed in the Shendurney wildlife settlements, making them easily accessible. sanctuary was the significant increase in Therefore, they faced different types of degraded forests, the decline in the extent of disturbances. The substantial increase in moist deciduous forests, semi-evergreen forests, degraded forests could be due to the expansion and insignificant gain in evergreen forests. The in the human settlements in the nearby locations apparent changes reported in most of the forest and pressure on the demand for agricultural land ecosystems in the present investigation

Bilyaminu et al.; AJEE, 15(4): 20-27, 2021; Article no.AJEE.70507

emphasized the need for compelling managerial 8. Jensen JR. "Digital Change Detection” action to sustainably monitor various human Introductory digital image processing: a activities, which are considered the major change remote sensing perspective. Prentice-Hall, actors. Additionally, improving the living standard New Jersey; 2004. of the forest fringe communities should be 9. Mallupattu PK, Sreenivasula Reddy JR. underscored. Analysis of land use/land cover changes using remote sensing data and GIS at an ACKNOWLEDGEMENT Urban Area, Tirupati, India. Sci. World J. 2013;6. We humbly acknowledge the Kerala forest 10. Roy PS, Kaul RN, Sharma Roy MR, department for the approval to conduct this Garbyal SS. Forest-type stratification and research and all the management of the delineation of shifting cultivation areas in Shendurney Wildlife Sanctuary. We also the eastern part of Arunachal Pradesh acknowledge the Department of Natural using LANDSAT MSS data. International Resource Management for providing all the Journal of Remote Sensing. 1985;6(3-4): facilities for the research. 411-418. 11. Kushwaha SPS, Madhavan Unni NV. COMPETING INTERESTS Hybrid interpretation for tropical forest classification. Asian-Pacific Remote Authors have declared that no competing Sensing J. 1989;1(2):69-75. interests exist. 12. Panigrahy RK, Kale MP, Dutta U, Mishra A, Banerjee B, Singh S. Forest cover REFERENCES change detection of Western Ghats of Maharashtra using satellite remote sensing based visual interpretation technique. 1. Echeverría C, Newton AC, Lara A, Current Science. 2010;98(5):657-664. Benayas JMR, Coomes DA. Impacts of 13. Dutta K, Reddy CS, Sharma S, Jha CS. forest fragmentation on species Quantification and monitoring of forest composition and forest structure in the cover changes in Agasthyamalai temperate landscape of southern Biosphere Reserve, Western Ghats, India Chile. Glob. Ecol. Biogeogr. 2007;16(4): (1920–2012). Current Science. 2016; 426-439. 110(4):508-520. 2. Benz UC, Hofmann P, Willhauck G, 14. Kale MP, Chavan M, Pardeshi S, Joshi C, Lingenfelder I, Heynen M. Multi-resolution, Verma PA, Roy PS, Srivastav SK, object-oriented fuzzy analysis of remote Srivastava VK, Jha AK, Chaudhari S, Giri sensing data for GIS-ready Y. Land-use and land-cover change in information. ISPRS J. Photogramm. Western Ghats of India. Environ. Monit. Remote Sens. 2004;58(3-4):239- A ssess. 2016;188(7):1-23. 258. 15. Reddy CS, Jha CS, Dadhwal VK. 3. Ellenberg D, Mueller-Dombois D. Aims and Assessment and monitoring of long-term methods of vegetation ecology. New York: forest cover changes (1920–2013) in Wiley; 1974. Western Ghats biodiversity hotspot. 4. Wallace J, Behn G, Furby S. Vegetation Journal of Earth System Science. 2016; condition assessment and monitoring from 125(1):103-114. sequences of satellite imagery. Ecol. 16. Kushwaha SPS. Forest-type mapping and Manag. Restor. 2006;7:31-S36. change detection from satellite 5. Anderson JR. A land use and land cover imagery. ISPRS J. Photogramm. Remote classification system for use with remote Sens. 1990;45(3):175-181. sensor data. US Government Printing 17. FAO. Global forest resources assessment Office. 1976;964. 2000: main report. FAO forestry paper 140, 6. Richards JA, Richards JA. Remote sensing Food and Agriculture Organization, Rome, digital image analysis Berlin: Springer. Italy; 2001. 1999;3:10-38. 18. Kim M, Madden M, Warner TA. Forest type 7. Congalton RG, Green K. Assessing the mapping using object-specific texture Accuracy of Remotely Sensed Data: measures from multispectral Ikonos Principles and Practicesboca Rotan. Lewis imagery. Photogramm. Eng. Remote Publishers, Florida; 1999. Sensing. 2009;75(7):819-829.

Bilyaminu et al.; AJEE, 15(4): 20-27, 2021; Article no.AJEE.70507

19. Curran LM, Trigg SN, McDonald AK, 21. Torahi AA, Rai SC. Land cover Astiani D, Hardiono YM, Siregar P, classification and forest change analysis, Caniago I, Kasischke E. Lowland forest using satellite imagery-a case study in loss in protected areas of Indonesian Dehdez area of Zagros Mountain in Iran. J. Borneo. Science. 2004;303(5660):1000- Geogr. Inf. Syst. 2011;3(01):1. 1003. 22. Ewers RM, Didham RK. Confounding 20. Lira PK, Tambosi LR, Ewers RM, Metzger factors in the detection of species JP. Land-use and land-cover change in responses to habitat fragmentation. Biol. Atlantic Forest landscapes. For. Ecol. Rev. 2006;81:117–142. Manag. 2012;278:80-89. ______© 2021 Bilyaminu et al.; This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Peer-review history: The peer review history for this paper can be accessed here: http://www.sdiarticle4.com/review-history/70507