Answers are grounded in this paper and cite their sources.
Demonstrates playa inundation changes in west Texas, indicating crucial environmental implications.
Playas are shallow surface-water features that intermittently store precipitation and snow, contribute to groundwater recharge of the Ogallala Aquifer, and provide critical habitat for migratory birds and other wildlife. However, existing playa inventories are largely derived from manual digitization over the past two decades and lack systematic updates to capture boundary dynamics, active wetland status, and inundation variability, limiting their usefulness for large-scale monitoring and water resource management. This study presents a novel AI-driven framework that integrates multi-temporal remote sensing datasets with an advanced deep learning architecture combining a ResNet50 (50-layer residual neural network) backbone, UPerNet (unified perceptual parsing network) decoder, and graph convolutional network (GCN) fusion to demonstrate and delineate playa boundaries across west Texas. Playa activity was quantified using Normalized Difference Water Index (NDWI) time-series criteria to assess long-term changes in inundation extent, frequency and the number of active playas from 1995 to 2025. The proposed framework achieved robust segmentation performance, with an intersection-over-union (IoU) of 0.79 and a Dice validation coefficient of 0.82 and demonstrated strong spatial agreement with reference data (R² = 0.99; RMSE = 828 ha). County-level trend analysis using the Mann–Kendall test revealed statistically significant declines (p ≤ 0.05) in both playa inundation extent and the number of active playas, with the most pronounced decreases occurring in the drier western regions of west Texas, consistent with the east-to-west rainfall gradient across the region. Moreover, the annual frequency of playa inundation exhibited a proportion of approximately 20% within five- to ten-year intervals. Additionally, larger playas exhibit stronger declines in inundation area across all percentile groups (>75th, 25–75th, <25th percentiles), indicating pronounced spatial variability in hydrological dynamics despite interannual fluctuations. Overall, this framework enables consistent long-term monitoring of playa dynamics and provides timely decision support for sustaining surface-water availability, groundwater recharge, and wetland ecosystem services under increasing hydroclimatic stress.
No takes yet. Share an insight, caveat, or question.
Chaiyana et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: