Browsing by Author "Nayomi, MPA"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • Thumbnail Image
    item: Conference-Abstract
    Determining the Invasive Plant Dynamics in Bolgoda Lake Using Open-source Data
    (Department of Earth Resources Engineering, University of Moratuwa, Sri Lanka, 2022-12-23) Kannangara, KATT; Shoukie, MB; Nayomi, MPA; Dassanayake, SM; Dassanyake, ABN; Jayawardena, CL; Jayawardena, CL
    Identifying invasive plants (IP) and monitoring their dynamics is essential to minimize potential adverse effects on natural resources. Remote sensing (RS) could effectively cater to such requirements by acquiring data in many critical domains. Limitations of spatial resolution, spectral information, and large imagery files usually hinder retrieving, managing, and analyzing remotely sensed data. The cloud-based computational capabilities of Google Earth Engine (GEE) provide the amenities for geospatial data analysis, retrieval, and processing with access to a majority of freely available, public, multi-temporal RS data. Integrating machine learning algorithms into GEE generates a promising path toward operationalizing automated RS-based IP monitoring by overcoming traditional challenges. Use of Classification and Regression Trees (CART) classifier to generate water-vegetation classification over six years (2016-2021) with Landsat 8 and Sentinel 2 images enabled mapping the invasive plants and their dominant component of Water Hyacinth (Pontederia crassipes) across a heterogeneous landscape in Bolgoda Lake, Sri Lanka. Also, the study could develop a relatively accurate classification of the water-vegetation dynamics over the time of interest. The classified time series data indicates the annual variation of the water, vegetation, and non-vegetation classes with rapidly fluctuating seasonal cycles for the vegetation cover. These results could benefit regulatory authorities and institutions to optimize environmental resource management and prioritize eco-preservation attempts. Moreover, the findings reflect the capabilities of deep learning models to identify invasive plant behaviors even with modest spatial and spectral resolution imagery.
  • Thumbnail Image
    item: Conference-Full-text
    Spatiotemporal growth dynamics of invasive plant distribution in Bolgoda lake, Sri lanka: a gis based approach
    (IEEE, 2022-07) Kannangara, KATT; Shoukie, MB; Nayomi, MPA; Dassanayake, SM; Jayawardane, C; Anjula, ABN; Rathnayake, M; Adhikariwatte, V; Hemachandra, K
    Inland water bodies in urban areas, such as Bolgoda lake, host vegetation covers that exhibit significant spatio-temporal variations throughout the year. Seasonal weather patterns, anthropogenic activities, such as surface mining, wastewater discharge and invasive plant growth typically govern these dynamics. Measuring, and monitoring, these factors over the spatial extent of these waterbodies require significant efforts. Yet, remote sensing and earth observation data can effectively minimize these efforts. This study employs both Landsat and Sentinel satellite data to estimate the normalized difference vegetation index (NDVI) and the normalized difference water index (NDWI) to develop a relatively accurate classification of the water-vegetation dynamics over the time of interest. The Google Earth Engine and ArcGIS software were used to download and generate the classifications over six years (2016-2021) for four different seasons (i.e., 24 processed images). The classified time series data show that the vegetation cover varies at two temporal frequencies. The annual variation of the water, vegetation, and non-vegetation (other) classes is consistent and cyclic. However, at a finer temporal resolution (i.e., on seasonal cycles), vegetation dynamics fluctuate rapidly. The study offers scope for using the results to support policymakers in optimizing environmental resource management strategies for urban surface water bodies.