Abstract It is important to understand the land-use/land-cover (LULC) transitions of the past and future to manage coastal zones sustainably. This paper uses a machine learning-based geospatial method to examine LULC change in the Nagapattinam Coastal District of Tamil Nadu, between 1990 and 2030. Using remote sensing (RS), geographic information systems (GIS), and Artificial Neural Network (ANN) modelling, LULC change was historical and projected under a business-as-usual approach. The analysis examined seven major LULC classes: Waterbody, built-up land, agricultural land, forest, mudflat, saltpan, and barren land. Changes to these LULC classes were examined using classified satellite data to show both changes and predict the future trends. The analysis showed a substantial increase in built-up land from 29.58 km2 in 1990 to 59.19 km2 in 2030, indicating the rapid urbanisation of the coastal area. Agricultural land is predicted to decrease substantially from 726.10 to 603.88 km2 during the same time period, mainly due to converted land usage. Mudflat and saltpan showed a dramatic increase in LULC change that suggests increased gross salt production and sedimentation. Waterbody and forest, by contrast, showed slight decreases in area, which raises ecological concerns about land degradation. The mudflat area grew from 89.59 km2 in 1990 to 169.55 km2 in 2030. Barren land continually decreased in area, meaning it transitioned to other use types. These expected changes show possible environmental concerns like habitat loss, reduced agricultural production, and increased exposure to coastal hazards. The research illustrates the utility of ANN-based geospatial modelling in documenting complex patterns of land use change and provides valuable knowledge to policymakers and planners on how to promote adaptive and resilient coastal development. Highlights <UnorderedList Mark="Bullet"> <ItemContent> Integrated RS–GIS–ANN modelling reconstructs historical (1990–2020) and predicts future (2030) LULC transitions in the Nagapattinam Coastal District. </ItemContent> <ItemContent> Findings show sharp urban expansion, major agricultural decline, and significant growth in mudflat and saltpan areas, indicating intensified sedimentation and salt production. </ItemContent> <ItemContent> Projected LULC shifts reveal rising ecological risks – habitat loss, reduced crop productivity, and increased coastal vulnerability – supporting evidence-based sustainable coastal planning. </ItemContent> </UnorderedList>

Machine learning-based assessment of past and future LULC transitions through geospatial modelling from 1990 to 2030 for Nagapattinam Coastal District, Tamil Nadu

期刊论文
发表日期：

2025年12月10日

查看全文

D Deva Raja Subha,
D Jayganesh

摘要

Abstract

It is important to understand the land-use/land-cover (LULC) transitions of the past and future to manage coastal zones sustainably. This paper uses a machine learning-based geospatial method to examine LULC change in the Nagapattinam Coastal District of Tamil Nadu, between 1990 and 2030. Using remote sensing (RS), geographic information systems (GIS), and Artificial Neural Network (ANN) modelling, LULC change was historical and projected under a business-as-usual approach. The analysis examined seven major LULC classes: Waterbody, built-up land, agricultural land, forest, mudflat, saltpan, and barren land. Changes to these LULC classes were examined using classified satellite data to show both changes and predict the future trends. The analysis showed a substantial increase in built-up land from 29.58 km² in 1990 to 59.19 km² in 2030, indicating the rapid urbanisation of the coastal area. Agricultural land is predicted to decrease substantially from 726.10 to 603.88 km² during the same time period, mainly due to converted land usage. Mudflat and saltpan showed a dramatic increase in LULC change that suggests increased gross salt production and sedimentation. Waterbody and forest, by contrast, showed slight decreases in area, which raises ecological concerns about land degradation. The mudflat area grew from 89.59 km² in 1990 to 169.55 km² in 2030. Barren land continually decreased in area, meaning it transitioned to other use types. These expected changes show possible environmental concerns like habitat loss, reduced agricultural production, and increased exposure to coastal hazards. The research illustrates the utility of ANN-based geospatial modelling in documenting complex patterns of land use change and provides valuable knowledge to policymakers and planners on how to promote adaptive and resilient coastal development.

Highlights

Integrated RS–GIS–ANN modelling reconstructs historical (1990–2020) and predicts future (2030) LULC transitions in the Nagapattinam Coastal District.

Findings show sharp urban expansion, major agricultural decline, and significant growth in mudflat and saltpan areas, indicating intensified sedimentation and salt production.

Projected LULC shifts reveal rising ecological risks – habitat loss, reduced crop productivity, and increased coastal vulnerability – supporting evidence-based sustainable coastal planning.