Estimating Lake Volumes with AI Techniques

Strategic Insights -

Research from the University of Florida has demonstrated the efficacy of utilizing artificial intelligence (AI) in estimating lake volumes without the need for bathymetric data. This innovative approach employs the Google Earth Engine platform, AI, and genetic algorithms to analyze water surface areas and estimate lake volume. The method was applied to Lake Okeechobee in Florida, achieving a high degree of accuracy with a root mean square error of 273 million cubic meters, a mean absolute percentage error of 28.85%, and a percent bias of 21.8%.

Key Takeaways:

  • The research introduced an innovative approach that utilizes the Google Earth Engine platform, AI, and genetic algorithms to estimate lake volumes without bathymetric data.
  • The method was applied to Lake Okeechobee in Florida, utilizing Landsat-8 imagery and the normalized difference water index to calculate the lake's surface area.
  • AI techniques, including image segmentation and thresholding, were employed to refine the images, which were then analyzed using a genetic algorithm to estimate the lake's volume.
  • The estimated volume was compared with calculations derived from the lake's bathymetry, achieving a high degree of accuracy with a root mean square error of 273 million cubic meters (Mm3).
  • The research suggests that AI techniques can be highly useful for estimating lake volumes and further exploration of their application is recommended in future studies.
  • The study was conducted by researchers at the University of Florida, led by Golmar Golmohammadi, Department of Soil, Water and Ecosystem Sciences, IFAS, RCREC.
  • Additional authors for this research include Hamid Reza Eslami, Rashid Rasouli Harouni, Kamyab Eslami, Mahin Jamali, Seyed Mostafa Biazar, and Kourosh Mohammadi.

Statistics:

  • Root mean square error: 273 million cubic meters (Mm3)
  • Mean absolute percentage error: 28.85%
  • Percent bias: 21.8%
  • Accuracy achieved in estimating lake volume using AI techniques and genetic algorithms
  • Number of authors involved in the research: 7

Sources:

  • Estimating lake okeechobee's water volume using NDWI-derived remote sensing and genetic algorithm optimization on Google Earth Engine. Discover Water, 2025,5(1):1-26 (open access at https://doi-org.sdpl.idm.oclc.org/10.1007/s43832-025-00255-5)
  • University of Florida, Department of Soil, Water and Ecosystem Sciences, IFAS, RCREC
  • Life Science Weekly. September 2, 2025; p 7498 (citation: NewsRx. University of Florida Researchers Illuminate Research in Algorithms (Estimating lake okeechobee's water volume using NDWI-derived remote sensing and genetic algorithm optimization on Google Earth Engine).