Search for: All records

Introduction:Traditional methods to estimate exposure to PM_2.5(particulate matter with less than 2.5 µm in diameter) have typically relied on limited regulatory monitors and do not consider human mobility and travel. However, the limited spatial coverage of regulatory monitors and the lack of consideration of mobility limit the ability to capture actual air pollution exposure. Methods:This study aims to improve traditional exposure assessment methods for PM_2.5by incorporating the measurements from a low-cost sensor network (PurpleAir) and regulatory monitors, an automated machine learning modeling framework, and big human mobility data. We develop a monthly-aggregated hourly land use regression (LUR) model based on automated machine learning (AutoML) and assess the model performance across eight metropolitan areas within the US. Results:Our results show that integrating low-cost sensor with regulatory monitor measurements generally improves the AutoML-LUR model accuracy and produces higher spatial variation in PM_2.5concentration maps compared to using regulatory monitor measurements alone. Feature importance analysis shows factors highly correlated with PM_2.5concentrations, including satellite aerosol optical depth, meteorological variables, vegetation, and land use. In addition, we incorporate human mobility data on exposure estimates regarding where people visit to identify spatiotemporal hotspots of places with higher risks of exposure, emphasizing the need to consider both visitor numbers and PM_2.5concentrations when developing exposure reduction strategies. Discussion:This research provides important insights for further public health studies on air pollution by comprehensively assessing the performance of AutoML-LUR models and incorporating human mobility into considering human exposure to air pollution. 

Natural phenomena are intrinsically spatiotemporal and often highly dynamic. The increasing availability of simulation and observation datasets has provided us a great opportunity to better capture and understand the complexity and dynamics of natural phenomena. Challenges are posed by the formalization of the representation of such phenomena in terms of their non-rigid boundaries and the quantification of event dynamics over space and time. The objectives of this research are to (1) conceptually represent the natural phenomenon as an event, and (2) quantify the dynamic movements and evolutions of events using a graph-based approach. This proposed data framework is applied to a dust simulation dataset to represent the 4D dynamic dust events. Dust events are identified, and movements are tracked to reconstruct dust events in the Northern Africa region from December 2013 to November 2014. Quantified dynamics of different dust events are demonstrated and verified to be in alignment with observations. 

Climate and weather data such as precipitation derived from Global Climate Models (GCMs) and satellite observations are essential for the global and local hydrological assessment. However, most climatic popular precipitation products (with spatial resolutions coarser than 10km) are too coarse for local impact studies and require “downscaling” to obtain higher resolutions. Traditional precipitation downscaling methods such as statistical and dynamic downscaling require an input of additional meteorological variables, and very few are applicable for downscaling hourly precipitation for higher spatial resolution. Based on dynamic dictionary learning, we propose a new downscaling method, PreciPatch, to address this challenge by producing spatially distributed higher resolution precipitation fields with only precipitation input from GCMs at hourly temporal resolution and a large geographical extent. Using aggregated Integrated Multi-satellitE Retrievals for GPM (IMERG) data, an experiment was conducted to evaluate the performance of PreciPatch, in comparison with bicubic interpolation using RainFARM—a stochastic downscaling method, and DeepSD—a Super-Resolution Convolutional Neural Network (SRCNN) based downscaling method. PreciPatch demonstrates better performance than other methods for downscaling short-duration precipitation events (used historical data from 2014 to 2017 as the training set to estimate high-resolution hourly events in 2018). 

Precipitation, especially convective precipitation, is highly associated with hydrological disasters (e.g., floods and drought) that have negative impacts on agricultural productivity, society, and the environment. To mitigate these negative impacts, it is crucial to monitor the precipitation status in real time. The new Advanced Baseline Imager (ABI) onboard the GOES-16 satellite provides such a precipitation product in higher spatiotemporal and spectral resolutions, especially during the daytime. This research proposes a deep neural network (DNN) method to classify rainy and non-rainy clouds based on the brightness temperature differences (BTDs) and reflectances (Ref) derived from ABI. Convective and stratiform rain clouds are also separated using similar spectral parameters expressing the characteristics of cloud properties. The precipitation events used for training and validation are obtained from the IMERG V05B data, covering the southeastern coast of the U.S. during the 2018 rainy season. The performance of the proposed method is compared with traditional machine learning methods, including support vector machines (SVMs) and random forest (RF). For rainy area detection, the DNN method outperformed the other methods, with a critical success index (CSI) of 0.71 and a probability of detection (POD) of 0.86. For convective precipitation delineation, the DNN models also show a better performance, with a CSI of 0.58 and POD of 0.72. This automatic cloud classification system could be deployed for extreme rainfall event detection, real-time forecasting, and decision-making support in rainfall-related disasters.