Accuracy assessment of determining the areas of small lakes based on remote sensing data

Abstract

Small lakes are important elements of the ecosystem, especially under conditions of climate change and active water use. Studying their condition is of key importance for understanding the water balance and biodiversity of the region. This study compares remote sensing techniques for small lakes using Landsat-8, Sentinel-2, Planet and unmanned aerial vehicle (UAV) data.

The aim of the work is to determine the most effective approach for monitoring lake conditions by analyzing the accuracy of different methods and indices. The methods used include the analysis of multispectral data such as NDWI, MNDWI and AWEI, which allow the distinction between aquatic and non-aquatic objects. The scientific significance of the study lies in assessing the potential of modern remote sensing technologies for detailed monitoring of ecosystem changes. Practical significance of the work consists in providing recommendations for sustainable water resources management and development of adaptation strategies.

The research methodology included processing satellite data of different resolutions, accuracy verification using UAV reference data, and application of spectral indices for water surface mapping. Results showed that high-resolution imagery (PlanetLab data) most closely matched field observations. The spectral indices NDWI, MNDWI and AWEI showed different accuracy depending on data characteristics and lake features. The analysis confirms the effectiveness of an integrated approach using high-resolution data and spectral indices for monitoring the condition of small lakes. The practical value of the study lies in optimizing the monitoring of water bodies and maintaining their ecological sustainability.