RESUMO
We present a novel method for detecting red tide (Karenia brevis) blooms off the west coast of Florida, driven by a neural network classifier that combines remote sensing data with spatiotemporally distributed in situ sample data. The network detects blooms over a 1-km grid, using seven ocean color features from the MODIS-Aqua satellite platform (2002-2021) and in situ sample data collected by the Florida Fish and Wildlife Conservation Commission and its partners. Model performance was demonstrably enhanced by two key innovations: depth normalization of satellite features and encoding of an in situ feature. The satellite features were normalized to adjust for depth-dependent bottom reflection effects in shallow coastal waters. The in situ data were used to engineer a feature that contextualizes recent nearby ground truth of K. brevis concentrations through a K-nearest neighbor spatiotemporal proximity weighting scheme. A rigorous experimental comparison revealed that our model outperforms existing remote detection methods presented in the literature and applied in practice. This classifier has strong potential to be operationalized to support more efficient monitoring and mitigation of future blooms, more accurate communication about their spatial extent and distribution, and a deeper scientific understanding of bloom dynamics, transport, drivers, and impacts in the region. This approach also has the potential to be adapted for the detection of other algal blooms in coastal waters. Integr Environ Assess Manag 2024;20:1432-1446. © 2024 SETAC.
Assuntos
Monitoramento Ambiental , Proliferação Nociva de Algas , Tecnologia de Sensoriamento Remoto , Monitoramento Ambiental/métodos , Florida , DinoflagellidaRESUMO
Transformers are widely used deep learning architectures. Existing transformers are mostly designed for sequences (texts or time series), images or videos, and graphs. This paper proposes a novel transformer model for massive (up to a million) point samples in continuous space. Such data are ubiquitous in environment sciences (e.g., sensor observations), numerical simulations (e.g., particle-laden flow, astrophysics), and location-based services (e.g., POIs and trajectories). However, designing a transformer for massive spatial points is non-trivial due to several challenges, including implicit long-range and multi-scale dependency on irregular points in continuous space, a non-uniform point distribution, the potential high computational costs of calculating all-pair attention across massive points, and the risks of over-confident predictions due to varying point density. To address these challenges, we propose a new hierarchical spatial transformer model, which includes multi-resolution representation learning within a quad-tree hierarchy and efficient spatial attention via coarse approximation. We also design an uncertainty quantification branch to estimate prediction confidence related to input feature noise and point sparsity. We provide a theoretical analysis of computational time complexity and memory costs. Extensive experiments on both real-world and synthetic datasets show that our method outperforms multiple baselines in prediction accuracy and our model can scale up to one million points on one NVIDIA A100 GPU. The code is available at https://github.com/spatialdatasciencegroup/HST.