Spatio-Temporal Modeling and Projection of Changes in the Lucre-Huacarpay Wetland Against Land-Use Transformation and Climate Variability

Resumen

The Lucre-Huacarpay wetland, located in Cusco, is a representative case of the progressive threat faced by high- Andean wetlands due to land use and land cover transformation and the intensification of climate variability. This study analyzed the spatiotemporal changes of the wetland during the period 2000–2024, with projections for the years 2030 and 2050, with the aim of generating scientific evidence to contribute to its conservation. Landsat 5 and 8 satellite images were used, applying supervised classification using the Random Forest (RF) algorithm, complemented by spectral indices. An LSTM-type artificial neural network was used for the prediction model, along with input data on the area of the conserved and degraded wetland. The results show that the wetland lost 277.42% of its surface area during the study period, decreasing from 296.4 ha to 206.6 ha. Most of this reduction was due to the increase in degraded wetland, which represented water stress in various areas of the wetland. This category increased from 9.6 to 77.1 ha. The increase in pasture and cropland also contributed to the change. Projections indicate a reduction of conserved wetland to 176,54 ha in 2030 and 168,53 ha in 2050, considering conservation measures implemented in the final years of the analyzed period. These results indicate the need to strengthen land management, continuous monitoring through remote sensing, and integrated decision-making to ensure ecosystem conservation.