See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/384261276

Using Neural Networks for Sustainable Land Use Prediction in Sumbawa

Regency, Indonesia

Article in Applied Environmental Research · September 2024

DOI: 10.35762/AER.2024045

CITATIONS READS

0 66

12 authors, including:

Muhammad Ramdhan Rudhy Akhwady

National Research and Innovation Agency Ministry of Marine Affairs and Fisheries
122 PUBLICATIONS 288 CITATIONS 37 PUBLICATIONS 63 CITATIONS

SEE PROFILE SEE PROFILE

Taslim Arifin Dini Purbani

National Research and Innovation Agency Ministry of Marine Affairs and Fisheries The Republic of Indonesia
59 PUBLICATIONS 219 CITATIONS 39 PUBLICATIONS 100 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Muhammad Ramdhan on 25 September 2024.

The user has requested enhancement of the downloaded file.

 App. Envi. Res. 46(3) (2024): 045

Applied Environmental Research

- RESEARCH FOR SUSTAINABLE PLANET -

Research Article

Using Neural Networks for Sustainable Land Use Prediction in Sumbawa Regency,
Indonesia

Muhammad Ramdhan1,*, Rudhy Akhwady2, Taslim Arifin2, Dini Purbani2, Yulius2, Dino G. Pryambodo3,
Rinny Rahmania4, Olivia Maftukhaturrizqoh1, Abdul Asyiri5, Syamsul Hidayat6, Arya Ningsih7, Sadad7

1 Research Center for Geoinformatics, Research Organization for Electronics and Informatics, Bandung, Indonesia
2 Research Center for Conservation of Marine and Inland Water Resources, Research Organization for Earth and Maritime, Cibinong, Indonesia
3 Research Center for Lymnology and Water Resources, Research Organization for Earth and Maritime, Cibinong, Indonesia
4 Research Center for Ecology and Etnobotany, Research Organization for Life Science, Cibinong, Indonesia
5 Center for Data and Information, Bearau of Organization and Human Resources, Cibinong, Indonesia
6 Faculty of Environmental and Mineral Technology, Sumbawa University of Technology, Sumbawa, Indonesia
7 Agency for Local Research and Development, Government of Sumbawa, Sumbawa, Indonesia

*Correspondence Email: muha307@brin.go.id

Abstract ARTICLE HISTORY

Agriculture is vital to Sumbawa Regency's economy, with key activities such as Received: 2 Aug. 2024
rice cultivation, corn production, onion farming, and cattle rearing. This study Accepted: 16 Sep. 2024
applies artificial neural networks (ANN) to predict land cover changes, focusing on Published: 23 Sep. 2024
agricultural land expansion. Using land cover datasets from ESRI, digital elevation
model, and topographical maps, we analyzed land cover changes from 2017 to KEYWORDS
2023 and generated future projections for 2050 with the MOLUSCE plugin in
Artificial neural network;
qGIS. The predictive model achieved an 85% accuracy rate when comparing 2023
Land cover change;
actual data with predictions. Results indicate a significant increase in agricultural
Sumbawa regency;
land cover by 2050. The key finding is that over a long-term period, the simulation
Sustainable land
of land use and land cover (LULC) change in Sumbawa reveals an increase of crop
management;
areas in the Lunyuk and Labangka Districts. This study highlights the effectiveness
MOLUSCE
of ANN in land cover prediction and emphasizes the need for sustainable practices
to balance agricultural expansion. AI-driven insights can aid policymakers in opti-
mizing resource allocation and ensuring long-term environmental and economic
stability in Sumbawa Regency. Future research should refine models and incorporate
additional factors for improved accuracy.

Introduction which collectively bolster productivity and resilience

Agriculture constitutes a fundamental economic against climatic fluctuations [1]. The integration of AI-
activity in Sumbawa Regency, significantly supported driven solutions is essential for ensuring long-term food
by rice cultivation, corn production, onion farming, security and economic stability in Sumbawa Regency.
and cattle rearing, all of which benefit from the region's Agriculture necessitates land as a critical component
favorable tropical climate and diverse geographical for its production processes, providing the essential space
features. The application of artificial intelligence (AI) and natural resources required for cultivating crops and
technologies has the potential to optimize both irrigated raising livestock. The availability and quality of arable
and dryland agricultural practices, enabling more efficient land directly influence agricultural productivity, deter-
adaptation to seasonal variations and thereby enhancing mining the types of crops that can be grown and the
crop yields. Scholarly research underscores the pivotal methods of cultivation that can be employed. Furthermore,
role of AI in advancing sustainable farming practices land management practices, including soil conservation,
and the adoption of precision agriculture technologies, irrigation, and crop rotation, play a vital role in sustaining

https://doi.org/10.35762/AER.2024045
 App. Envi. Res. 46(3) (2024): 045

agricultural output and mitigating the impacts of envi- provide objective guidance for local government
ronmental challenges. Effective land use planning and authorities regarding the allocation of suitable areas for
sustainable agricultural practices are imperative to optimize agricultural expansion. By utilizing these predictive insights,
land resources and ensure long-term food security and policy-makers can make informed decisions to optimize
environmental health [2]. resource allocation and promote sustainable agricultural
Monitoring and controlling land cover change is development in the region.
crucial due to the limited available area in the regency.
Land cover change, including deforestation, urbanization, Materials and methods
and agricultural expansion, can significantly impact local 1) Area description
ecosystems, biodiversity, and agricultural productivity. Sumbawa Regency (Figure 1), is one of the ten adminis-
Effective land use management is essential to balance trative regions in West Nusa Tenggara Province, is located
development needs with environmental conservation, at the western tip of Sumbawa Island, Indonesia [6]. Following
ensuring sustainable use of the finite land resources. the division of Sumbawa Regency into two separate
Advanced geospatial technologies, such as remote sensing regions in 2003, Sumbawa Regency itself comprises 24
and GIS, play a vital role in tracking land cover changes sub-districts. Geographically, it is bordered by West
and informing policy decisions aimed at sustainable Sumbawa Regency to the west, Dompu Regency to the
land management [3–4]. east, the Flores Sea to the north, and the Indian Ocean
Predicting land use and land cover (LULC) is to the south. Notably, Sumbawa Regency stands out in
essential for sustainable development planning because Indonesia for including Saleh Bay, an enclosed ecosystem
it provides a detailed understanding of how land is utilized within its territory.
and how it changes over time. Before the advent of neural This unique geographical positioning of Sumbawa
networks and artificial intelligence, several widely-used Regency, encompassing diverse ecosystems and bordering
methods for predicting land use change were employed, significant marine environments, underlines its ecological
such as Markov model [5]. These methods generally involved and economic importance. Saleh Bay, in particular, offers
statistical, econometric, and spatial modeling approaches. a unique enclosed marine ecosystem that supports various
This paper aims to investigate the application of arti- aquatic species and contributes to the local economy
ficial neural networks in predicting land cover changes through fishing and tourism. The distinct sub-districts
within Sumbawa Regency. It will particularly focus on within the regency display varied land use and cover
scenarios projecting an increase in agricultural land cover dynamics, reflecting broader trends in agricultural
in the future. The resulting predictions are intended to expansion, urbanization, and environmental conservation.

Figure 1 Sumbawa Regency map.

 App. Envi. Res. 46(3) (2024): 045

Sumbawa Regency experiences a tropical climate

characterized by distinct rainy and dry seasons. In 2023, 2) Data and materials
the maximum temperature peaked at 38°C in September, To analyze land use and land cover change using
while the minimum temperature dropped to 19.8°C, MOLUSCE plug-in qGIS version 2.8, the first step was
also in September. The average humidity in 2023 reached to ensure that all relevant data layers were imported
its highest point at 95 percent in February. Throughout (Table 1). This included the ESRI LULC dataset,
the year, there were a total of 102 recorded rainy days. providing detailed land use and land cover information
The highest rainfall of 263.3 mm was observed in January in Sumbawa Regency for year 2017–2023 (Figure 3),
[5]. digital elevation model at National Level (DEMNAS)
According to data from BPS-Statistics Sumbawa dataset sourced from the Indonesian Geospatial Infor-
Regency, in September 2022, the population of Sumbawa mation Agency (BIG) for elevation data (Figure 4a), the
was reported as 527,607 thousand people, with a slightly road network layer obtained from the Topographical
higher number of males than females. The population Map of Indonesia (RBI Map) (Figure 4b), and the
density of Sumbawa Regency in the same year was recorded administrative boundaries layer, also from the RBI Map,
at 79 people per square kilometer. Among the 24 sub- delineating the study area's administrative divisions.
districts within the regency, there was significant variation The ESRI LULC dataset is derived from ESA Sentinel-2
in population density. Notably, Sumbawa Sub-district satellite imagery with 10-meter spatial resolution and
had the highest density at 1499 people per square kilometer, produced from 6 spectral bands. The annual maps from
while Orong Telu Sub-district had the lowest at 12 people this dataset in the range of 5 years from 2017–2023
per square kilometer. The population growth from 2011 which has 8 classes of LULC classification were used as
to 2023 can be seen in Figure 2a. one of the inputs, classes details can be seen in Table 2.
In 2023, Sumbawa saw a notable increase in onion Once all datasets were loaded into qGIS, the
production across all its 24 subdistricts (Figure 2b), with a MOLUSCE plugin was activated. Within the MOLUSCE
collective rise of 325,951 tons compared to the previous interface, the input layers were specified, including the
year. Livestock numbers also witnessed significant figures, ESRI LULC data (Table 2), DEMNAS, road network, and
notably native chickens, Bali cattle, and broiler administrative boundaries layers. Configuration settings
chickens, with respective totals of 808,215, 562,013, and and parameters were adjusted as per the analysis require-
288,300. Fisheries production in 2023 amounted to ments. Subsequently, the analysis was initiated, with
688,033 tons, marking a substantial uptick from the MOLUSCE processing the input data to generate outputs
preceding year. The rich potential of Sumbawa's fishery such as land use/land cover change maps, statistical
resources is evident, spanning marine fisheries, public analyses, and comprehensive reports. These outputs
waters, freshwater, brackish water aquatic resources, and facilitated a detailed assessment of land use dynamics
marine aquaculture, indicating diverse opportunities and and patterns within the study area, providing valuable
sectors for development and sustainability initiatives. insights for land management and planning purposes.

Figure 2 (a) Population growth in Sumbawa Regency and (b) sub-districts in Sumbawa Regency.
 App. Envi. Res. 46(3) (2024): 045

Table 1 data that used for this study

Data type Data source Resolution/scale Time
Sumbawa land cover https://livingatlas.arcgis.com/landcover/, ESRI 10 m 2017– 2023
Digital elevation model https://tanahair.indonesia.go.id/demnas/, DEMNAS-BIG 5–10 m 2018
Road network https://tanahair.indonesia.go.id, RBI Map 1:25.000 2017
Administrative boundaries https://tanahair.indonesia.go.id, RBI Map 1:25.000 2017

Figure 3 Sumbawa land cover.

Figure 4 (a) Digital elevation model and (b) proximity map from road network.
 App. Envi. Res. 46(3) (2024): 045

Table 2 LULC classification based on ESRI [7]

No. Class Description
1 Water Areas with permanent bodies of water; do not encompass locations with fleeting or ephemeral
water presence; scarce vegetation, rocks, or human-constructed structures such as docks are
absent; examples include rivers, lakes, oceans, and salt plains that are inundated.
2 Trees Area with a high concentration of tall and dense vegetation (around 15 m or more). A closely
knit canopy typically characterizes these; examples: are forests, areas with densely packed tall
plants in savannas, plantations, swamps, and mangroves (locations with abundant vegetation
and water that is obscured by the thickness of the canopy).
3 Flood vegetation Areas where vegetation and water coexist throughout the year. Areas that experience periodic
flooding have a combination of grass, shrubs, trees, and bare ground. Examples are mangroves
inundated with water, plants growing in water, rice paddies, and other agricultural lands with
heavy water management and inundation.
4 Crops Agricultural lands with crops grown by humans, not reaching the height of trees; examples:
corn, wheat, soy, structured land without crops.
5 Built Area Man-made structures; major roadways and railway networks; large surfaces that don’t absorb water,
like parking lots, office buildings, and houses; examples: houses, urban areas, paved roads, asphalt.
6 Bare ground Areas with rocks or soil and minimal to no vegetation yearround; vast regions of sand and
deserts with scarce vegetation; examples: exposed rock or soil, deserts, sand dunes, dry salt flats,
dried lake beds, mines.
7 Clouds Areas where persistent cloud cover makes it impossible to determine the land cover.
8 Rangeland Open areas covered with uniform grass with limited tall vegetation; wild cereals and grasses
without human cultivation (not a farmed field); examples: natural meadows and fields with
scarce tree cover, open savannas with few trees, parks, golf courses, lawns, pastures.

3) LULC change analysis LULC maps of 2017 and 2020 and the spatial variables
To calculate the spatiotemporal change data from were used to predict the map of 2023. MOLUSCE plugin
the LULC classification maps, the area analysis tool of the also offers validation of the simulated and actual map to
Modules for Land-Use Change Simulation (MOLUSCE), approve the accuracy of the model using % of correctness
a free and most widely used plugin for urban modeling and Kappa validation coefficients. In the transition
and future scenario simulations through qGIS was used potential modeling module of MOLUSCE [19], based
[8–13]. The data was calculated for the study intervals on satisfactory results it was concluded that the ANN
of (2017 to 2020) to generate change maps on which model for the study data training was better with a
further ANN modelling was done. neighborhood value of 1×1 pixels, the learning rate of
0.001, maximum iterations set to 100, with 10 hidden
4) Artificial neural network modeling and validation layers, and the momentum value of 0.001 [20] as shown
measures in Figure 5. Then in the cellular automata simulation
The ANN model is a reliable tool that has been used module, using an iteration value of 1, the LULC map of
in numerous research studies for future LULC pre- 2023 was simulated. Afterward, in the validation module,
dictions [14-16]. The purpose of the model is to generate the simulated LULC map of 2023 was validated with the
a transitional potential map using different compu- reference ESRI-based LULC map of 2023. The performance
tational intelligence aspects. The strategy of the model of the algorithms was also assessed by confusion matrix
is to handle enormous amounts of uncertain data. ANN derivatives such as classification accuracy and Kappa
incorporates fuzzy logic, by describing the terrain on a statistics. In addition, the accuracy of the prediction was
continuous range from 0 to 1. The alteration of the weight assessed by root mean square error (RMSE) [20]. After
connections between geographically linked neurons is obtaining satisfactory results, as shown in Figure 6, the
an important element of ANN [17]. They are dependent process was repeated for simulation prediction of the
on the computation power offered [18]. Nevertheless, LULC map of 2050 from the LULC maps 2017 along
it is highly suitable for urban growth modeling as it can with the thematic layers. The transition map generated
connect well between the complex relationship of huge by MOLUSCE for this prediction of future LULC can
data fed and extracted. be seen in Figure 7. While the map of Sumbawa land
cover prediction for the year 2050 is presented in Figure
8.
 App. Envi. Res. 46(3) (2024): 045

Figure 5 Artificial neural network parameters and process using MOLUSCE plug-in qGIS.

Figure 6 Validation parameters and process using MOLUSCE plug-in qGIS.

Figure 7 Transition map generated by MOLUSCE for prediction of future LULC.

 App. Envi. Res. 46(3) (2024): 045

Figure 8 Land cover prediction in year 2050.

Results and discussion a broader context for understanding the local changes
1) Data analysis observed in your dataset, emphasizing the interconnected
The analysis of land use changes between 2017 and nature of land use change and its environmental impacts.
2020 in the given dataset (Table 3) highlights significant The provided matrix on Table 4 represents the land
transformations in various land cover types. Water bodies cover transition probabilities between different land
increased by 5.88 km2 (0.09%), while areas covered by cover types within a specified area. Each cell in the matrix
trees slightly decreased by 2.40 km2 (-0.04%). There was a represents the likelihood of transitioning from one land
notable reduction in flood vegetation (-3.40 km2, -0.05%) cover type (row) to another (column) over a given time
and bare ground (-4.47 km2, -0.07%). Conversely, crop period. For instance, the value at row "Water" and column
areas expanded significantly by 153.00 km2 (2.30%), "Trees" (0.016762) indicates the probability of transi-
indicating a shift towards agricultural development. tioning from water to trees. Similarly, the value at row
Built-up areas also grew by 21.73 km2 (0.33%), reflecting "Crops" and column "Built Area" (0.030102) represents
urbanization trends. The rangeland experienced a decline the probability of transitioning from crops to built areas.
of 161.63 km2 (-2.43%), highlighting a transition from These transition probabilities provide valuable insights
natural landscapes to other land uses. into the dynamics of land cover change within the study
These changes align with global trends observed in area, informing land management strategies, environ-
land use and cover, where urban expansion and agricul- mental planning, and conservation efforts. Analyzing
tural intensification are prominent. Studies like the one the matrix can help identify trends, hotspots of change,
by Achu et al. [22] using Landsat data reveal similar and potential areas for intervention or protection. With
patterns of increased cropland and built-up areas globally, using this transition matrix, the MOLUSCE qGIS give a
often at the expense of natural vegetation and water prediction of LULC of Sumbawa District in the year of
bodies. Additionally, research on the impacts of land use 2023. The prediction land use of 2023 then compare
change on water and carbon budgets, such as that by with the ESRI LULC in 2023, the result shown in Figure
Sun et al. [23], underscores the environmental implications 5. that the kappa validation number is quite satisfying
of these shifts, including changes in water consumption with number of 0.82244. Then we use the parameters
and carbon storage dynamics. These references provide to predict the LULC of Sumbawa District in 2050.
 App. Envi. Res. 46(3) (2024): 045

Table 3 Land cover change from 2017- 2020

2017 (km2) 2020 (km2) Δ (km2) 2017% 2020% Δ%
Water 51,97 57,86 5,88 0,78 0,87 0,09
Trees 4263,21 4260,82 -2,40 64,20 64,16 -0,04
Flood vegetation 25,25 21,85 -3,40 0,38 0,33 -0,05
Crops 801,05 954,05 153,00 12,06 14,37 2,30
Built area 99,34 121,07 21,73 1,50 1,82 0,33
Bare ground 7,87 3,40 -4,47 0,12 0,05 -0,07
Clouds 10,05 1,33 -8,72 0,15 0,02 -0,13
Rangeland 1382,21 1220,58 -161,63 20,81 18,38 -2,43

Table 4 Transition matrix, land cover change from 2017 – 2020

Water Trees Flood Crops Built Area Bare Clouds Rangeland
vegetation ground
Water 0,863793 0,016762 0,015297 0,050950 0,014229 0,005629 0,000000 0,033339
Trees 0,000184 0,951976 0,000295 0,004985 0,001168 0,000002 0,000281 0,041110
Flood vegetation 0,127389 0,151842 0,668771 0,030684 0,000930 0,001453 0,000000 0,018931
Crops 0,004504 0,003079 0,000707 0,898510 0,018231 0,000035 0,000000 0,074934
Built area 0,011137 0,003369 0,000942 0,030102 0,941752 0,002420 0,000003 0,010275
Bare ground 0,214641 0,001371 0,131873 0,103098 0,046070 0,346590 0,000114 0,156241
Clouds 0,000000 0,975769 0,000000 0,004587 0,002159 0,000000 0,011951 0,005533
Rangeland 0,001888 0,133861 0,000873 0,148857 0,004910 0,000049 0,000006 0,709556

The future of crops area appears relatively stable soil improvement. These transitions emphasize the
based on graph on Figure 9. With an 80% probability importance of sustainable land reclamation projects
that existing cropland will remain as such in 2050. and environmental impact assessments to ensure the
However, notable transitions include an 8.18% like- productive and ecological viability of new croplands. As
lihood of cropland converting to water, likely due to for Foley et al. [25] discussed solutions for cultivated planet
issues such as flooding or irrigation practices. Additionally, scenarios, emphasizing the balance needed between
there is a 2.73% chance of crops transitioning to trees, expansion and sustainability of croplands.
which may result from reforestation or natural succession, The analysis of land use and cover across various
and a 2.84% probability of urban expansion encroaching sub-districts reveals distinct trends (Figure 10). Water
on agricultural land. These transitions highlight the coverage remains relatively stable in most areas, with
need for effective water management, careful urban notable increases in some sub-districts, such as Alas
planning, and the promotion of agroforestry to mitigate Barat, where it peaked at 16.64% in 2023, and Plampang,
the loss of cropland. According to Pete et al. [24], com- which reached 7.74% the same year. Conversely, areas
petition for land resources is intensifying, making such like Sumbawa and Rhee experienced a decreasing trend
interventions crucial to ensure sustainable agricultural in water coverage. Tree coverage is mostly on a decline,
practices. with Lunyuk showing a decrease from 676.51 to 648.94
Conversely, the matrix indicates that some land by 2023, and Moyo Hilir witnessing a significant reduction.
currently not used for crops may be converted into crop- Despite this, some areas like Batu Lanteh have stable
land in the future. Specifically, there is a 10% probability of tree coverage with only minor fluctuations. Crop areas
flood vegetation being transformed into cropland through generally show an upward trend, reflecting agricultural
drainage and land reclamation efforts. Furthermore, expansion, with significant increases seen in Empang,
water areas have a 5.33% chance of becoming cropland, Moyo Hulu, and Plampang by 2023. This pattern aligns
likely through agricultural development and irrigation with findings by Pete et al. [24] which emphasize the
initiatives. Although bare lands show a minor 1.14% global trend of agricultural expansion at the cost of
probability of conversion to crops, this suggests potential other land covers.
opportunities for expanding agricultural activities through
 App. Envi. Res. 46(3) (2024): 045

Figure 9 Graphical change between each land cover: (a) from 2017-2020 and (b) from 2017-2050.

Figure 10 Crops land cover change in each sub-district.

Urbanization is evident from the gradual increase in Conclusions

built areas across all sub-districts, peaking in places like The study demonstrated the utility of the ANN-CA
Sumbawa and Moyo Hilir in 2023. Rangeland usage model in monitoring and predicting LULC changes in
fluctuates significantly, with some sub-districts like Sumbawa Regency from 2017 to 2050. By incorporating
Lunyuk and Lenangguar reaching peaks in 2023, while satellite-based LULC maps and thematic layers, the
others such as Sumbawa and Empang show a decreasing research validates the model's accuracy through the
trend. The minimum water coverage is consistently MOLUSCE plug-in, providing a reliable basis for future
reported as minimal and stable in sub-districts like projections. The visual assessment of agricultural land
Ropang, Unter Iwes, and Batu Lanteh. These trends further elucidates the patterns of area sprawl, offering
indicate ongoing shifts in land use, driven by factors crucial insights into the dynamics of agricultural deve-
such as deforestation, agricultural expansion, and urbani- lopment. The findings underscore the significant role
zation, highlighting the dynamic nature of these regions' of advanced modeling techniques in understanding
landscapes. Pete et al. [24] discuss similar dynamics, and managing LULC changes, which are critical for
noting the increasing competition for land resources as sustainable agricultural planning and development.
urban and agricultural needs grow.
 App. Envi. Res. 46(3) (2024): 045

The research highlights the urgent need for revising Geographic Information System, 2016, 8(3), 412
agricultural development plans to incorporate predicted –427.
trends and prevent potential chaotic scenarios. Recom- [6] BPS-Statistics Sumbawa Regency, Sumbawa Regency
mendations for promoting sustainable agricultural in figures 2024, Volume 16, ISSN: 0215-5834,
development through societal planning, environmental Publication Number: 52040.24001, Catalog: 1102001.
considerations, and controlled growth policies are crucial 5204, Sumbawa Regency, Indonesia, 2024, 529 p.
for guiding future agricultural expansion. The study serves [7] Karra, K., Kontgis, C., Statman-Weil, Z.,
as a valuable resource for policymakers, agricultural plan- Mazzariello, J.C., Mathis, M., Brumby, S. Global
ners, and administrative authorities, providing a compre- land use/land cover with Sentinel-2 and deep
hensive view of current and future agricultural sprawl. learning, 2021 IEEE International Geoscience and
For future research, the incorporation of satellite data at Remote Sensing Symposium, 2021, 4704–4707.
shorter intervals and additional variables is suggested to [8] Abbas, Z., Yang, G., Zhong, Y., Zhao, Y.,
enhance the accuracy and depth of LULC change analysis, Spatiotemporal change analysis and future scenario
offering new avenues for exploration and improvement. of LULC using the CA-ANN approach: A case
study of the Greater Bay Area, China. Land, 2021,
Acknowledgements 10(6), 584.
This research was supported by the Rumah Program [9] Bathe, D.K., Patil, N.S., Assessment of land use-
(Purwarupa Sistem Pendukung Keputusan Berdasarkan land cover dynamics and its future projection
Analisis Citra Satelit) and RIIM-2 - LPDP Grant, grant through Google Earth Engine, machine learning
number (82/II.7/HK/2022). We also thanks to the and QGIS-MOLUSCE: A case study in
Director of Research Center for Geoinformatics and Jagatsinghpur District, Odisha. Journal of Earth
Director of Research Center for Conservation of System Science, 2024, 133, 111.
Marine and Inland Water Resources, BRIN and [10] El-Tantawi, A.M., Bao, A., Chang, C., Liu, Y.,
Sumbawa Regency Government (PEMDA). Monitoring and predicting land use/cover changes
in the Aksu-Tarim River Basin, Xinjiang-China
References (1990–2030). Environmental Monitoring and
[1] Mana, A.A., Allouhi, A., Hamrani, A., Rehman, Assessment, 2019, 191, 480.
S., el Jamaoui, I., Jayachandran, K., Sustainable [11] Al Kafy, A., Al-Faisal, A., Rahman, S., Islam, M.,
AI-based production agriculture: Exploring AI Al Rakib, A., Islam, A., …, Sattar, G.S., Prediction
applications and implications in agricultural of seasonal urban thermal field variance index
practices. Smart Agricultural Technology, 2024, using machine learning algorithms in Cumilla,
7, 100416. Bangladesh. Sustainable Cities and Society, 2021,
[2] Mpanga I.K., Tronstad R., Guo J., LeBauer D.S., 64, 102542.
Idowu O.J., On-farm land management strategies [12] Khan, A., Sudheer, M., Machine learning-based
and production challenges in United States organic monitoring and modeling for spatio-temporal urban
agricultural systems, Current Research in Envi- growth of Islamabad. The Egyptian Journal of Remote
ronmental Sustainability, 2021, 3(1), 100097. Sensing and Space Sciences, 2022, 25, 541–550.
[3] Haj, F.A.E., Ouadif, L., Akhssas, A., Monitoring [13] Reddy, C.S., Pasha, S.V., Satish, K.V., Unnikrishnan,
land use and land cover changes using remote A., Chavan, S.B., Jha, C.S., …, Dadhwal, V.K.,
sensing techniques and the precipitation- Quantifying and predicting multi-decadal forest
vegetation indexes in Morocco. Ecological Engi- cover changes in Myanmar: A biodiversity hotspot
neering & Environmental Technology, 2023, 24 under threat. Biodiversity and Conservation, 2019,
(1), 272–286. 28(5), 1129–1149.
[4] Okoduwa, A.K., Amaechi, C.F. Exploring Google [14] Arifin, T., Amri, S.N., Rahmania, R., Yulius,
Earth engine, machine learning, and GIS for land Ramdhan, M., Chandra, H., Adrianto, L., …,
use land cover change detection in the federal Kurnia, R. Forecasting land-use changes due to
capital territory, Abuja, between 2014 and 2023, coastal city development on the peri-urban area
2024, Applied Environmental Research, 2024, in Makassar City, Indonesia. The Egyptian Journal
46(2), 029. of Remote Sensing and Space Sciences, 2023,
[5] Khawaldah, H.A., A prediction of future land 26(1), 197–206.
use/land cover in Amman area using GIS-based [15] Rahman, M.T., Esha, E.J. Prediction of land
Markov model and remote sensing. Journal of cover change based on CA-ANN model to assess
 App. Envi. Res. 46(3) (2024): 045

its local impacts on Bagerhat, southwestern Oplenac wine-producing area, Serbia). Environ-
coastal Bangladesh. Geocarto International, 2020. mental Monitoring and Assessment, 2018, 190
[16] Saputra, M.H., Lee, H.S. Prediction of land use 675.
and land cover changes for North Sumatra, Indonesia, [21] Achu, A., Thomas, J., Aju, C., Gopinath, G., Kumar,
using an artificial-neural-network-based cellular S., Reghunath, R., Machine learning modelling
automaton. Sustainability, 2019, 11(11), 3024. of fire susceptibility in a forest-agriculture Mosaic
[17] Bhattacharya, R.K., Chatterjee, N.D., Das, K., landscape of Southern India. Ecological Informatics,
Land use and land cover change and its resultant 2021, 64.
erosion susceptible level: An appraisal using RUSLE [22] Potapov, P., Hansen, M.C., Pickens, A., Serna,
and Logistic Regression in a tropical plateau basin A.H., Tyukavina, A., Turubanova, S., …,
of West Bengal, India. Environment, Development Kommareddy, A., The global 2000-2020 land
and Sustainability, 2021, 23, 1411–1446. cover and land use change dataset derived from
[18] MohanRajan, S.N., Loganathan, A., Manoharan, the Landsat archive: First results. Frontiers in
P., Survey on land use/land cover (LU/LC) change Remote Sensing, 2022, 3.
analysis in remote sensing and GIS environment: [23] Sun, J., Twine, T.E., Hill, J., Noe, R., Shi, J., Li,
techniques and challenges. Environmental Science M. Effects of land use change for crops on water
and Pollution Research, 2020, 27, 29900–29926. and carbon budgets in the Midwest USA, Sus-
[19] NextGIS and Asia Air Survey, MOLUSCE: tainability, 2017, 9(2), 225.
Modules for land use change evaluation, Quick [24] Pete S., Peter, J.G., Detlef, V., Michael, O., Petr
Help. 2024. [Online] Available from: https://github. H., Mark, R., …, Jessica, B., Competition for land,
com/nextgis/qgis_molusce/tree/master/src/molusce philosophical transactions of the Royal Society,
/doc/en 2010, B3652941–2957.
[20] Perovic, V., Jakšic, D., Jaramaz, D., Kokovic, N., [25] Foley, J., Ramankutty, N., Brauman, K., Cassidy,
Cakmak, D., Mitrovic, M., Pavlovic, P., Spatio- E.S., Gerber, J.S., Johnston, M., …, Zaks, D.P.M.,
temporal analysis of land use/land cover change Solutions for a cultivated planet, Nature, 2011,
and its effects on soil erosion (Case study in the 478, 337–342.

View publication stats