Distribución espacio-temporal de Eichhornia crassipes (Mart.) Solms a través de teledetección en laguna La Turbina, Cuba

Abstract

The early detection of invasive exotic plant species is essential for planning management and mitigation strategies. The primary objective of this study is to evaluate the spatio-temporal distribution of Eichhornia crassipes (Mart.) Solms, in La Turbina lagoon, using Sentinel-2 satellite images (period 2021-2023). The spectral bands of the visible, red edge, near infrared and shortwave infrared regions were used as input data in three machine learning algorithms during the classification process. The relationship between plant coverage and air temperature and precipitation have been analyzed. The model combining spectral bands and the selected spectral indices using the Random Forest recursive elimination method was the most efficient in the detection of E. crassipes. The months with the least coverage of E. crassipes were December 2022, January and March 2023, due to mechanized cleaning actions. E. crassipes is concentrated in the Northeast and Southwest of the lagoon, as an emerging macrophyte. The direction of the wind and human intervention has determined the spatio-temporal distribution of this plant. Based on these results, it is recommended to confine E. crassipes, in its condition as a floating macrophyte, to the deepest areas of La Turbina.

The early detection of invasive exotic plant species is essential for planning management and mitigation strategies. The primary objective of this study is to evaluate the spatio-temporal distribution of Eichhornia crassipes (Mart.) Solms, in La Turbina lagoon, using Sentinel-2 satellite images (period 2021-2023). The spectral bands of the visible, red edge, near infrared and shortwave infrared regions were used as input data in three machine learning algorithms during the classification process. The relationship between plant coverage and air temperature and precipitation have been analyzed. The model combining spectral bands and the selected spectral indices using the Random Forest recursive elimination method was the most efficient in the detection of E. crassipes. The months with the least coverage of E. crassipes were December 2022, January and March 2023, due to mechanized cleaning actions. E. crassipes is concentrated in the Northeast and Southwest of the lagoon, as an emerging macrophyte. The direction of the wind and human intervention has determined the spatio-temporal distribution of this plant. Based on these results, it is recommended to confine E. crassipes, in its condition as a floating macrophyte, to the deepest areas of La Turbina.