Search

Abstract The increasing knowledge in the capabilities of satellite imagery to hazard applications is especially useful in emergency situations where timing and ability to cover large areas are of the essence. For optical imagery, cloud coverage can corrupt an image rendering it unusable for intended emergency analyses. This study proposes the use of Synthetic Aperture Radar (SAR) imagery for burn severity analysis for western United States sites, as an alternative to its optical based counterpart, differenced normalized burn ratio (dNBR). Unlike optical sensors, the radar sensor is an active sensor that is able to penetrate clouds and smoke, an attribute that is crucial in emergency situations where immediate burn severity data are needed to assess the vulnerability of fire affected areas to post-fire hazards. Using C5 decision tree algorithm we developed a SAR-based metric that attempts to classify burn severities of fire affected locations in the western USA. We then compared the performance of this developed metric to that obtained by the existing dNBR metric, to determine if there is any merit to its adoption as an alternative for the western USA landscape. The results showed the SAR approach to produce higher validation metrics in comparison to the dNBR. It had an overall accuracy and kappa of 60% and 0.35, respectively, in comparison to the 35% and 0.1 of the dNBR approach. This shows an improved ability to quickly obtain burn severity data and make better informed decisions in emergency situations. [ABSTRACT FROM AUTHOR]

Existing precipitation-type algorithms have difficulty discerning the occurrence of freezing rain and ice pellets. These inherent biases are not only problematic in operational forecasting but also complicate the development of model-based precipitation-type climatologies. To address these issues, this paper introduces a novel light gradient-boosting machine (LightGBM)-based machine learning precipitation-type algorithm that utilizes reanalysis and surface observations. By comparing it with the Bourgouin precipitation-type algorithm as a baseline, we demonstrate that our algorithm improves the critical success index (CSI) for all examined precipitation types. Moreover, when compared with the precipitation-type diagnosis in reanalysis, our algorithm exhibits increased F1 scores for snow, freezing rain, and ice pellets. Subsequently, we utilize the algorithm to compute a freezing-rain climatology over the eastern United States. The resulting climatology pattern aligns well with observations; however, a significant mean bias is observed. We interpret this bias to be influenced by both the algorithm itself and assumptions regarding precipitation processes, which include biases associated with freezing drizzle, precipitation occurrence, and regional synoptic weather patterns. To mitigate the overall bias, we propose increasing the precipitation cutoff from 0.04 to 0.25 mm h−1, as it better reflects the precision of precipitation observations. This adjustment yields a substantial reduction in the overall bias. Finally, given the strong performance of LightGBM in predicting mixed precipitation episodes, we anticipate that the algorithm can be effectively utilized in operational settings and for diagnosing precipitation types in climate model outputs. Significance Statement: Freezing rain can have significant impacts on transportation and infrastructure, making accurate prediction of precipitation types crucial. In this study, we use a machine learning method known as LightGBM to predict precipitation types. We show that the new algorithm performs better than the existing methods for all precipitation types examined. Additionally, we compute a freezing-rain climatology over the eastern United States. Although the resulting climatology pattern corresponds well to observations, the algorithm overpredicts freezing-rain occurrence. We argue that this bias can be substantially reduced by increasing the precipitation cutoff from 0.04 to 0.25 mm h−1. Overall, this work highlights the potential of the LightGBM algorithm for both weather forecasting and diagnosing precipitation types in climate models. [ABSTRACT FROM AUTHOR]

Seasonal snow-derived water is a critical component of the water supply in the mountains and downstream regions, and the accurate characterization of available water in the form of snow-water-equivalent (SWE), peak SWE, and snowmelt onset are essential inputs for water management efforts. Arising from recent advancements in artificial intelligence (AI) and machine learning (ML), we introduce a large-scale ML SWE model leveraging publicly available data sources and open-source software. The model demonstrates the application of a limited feature space in a relatively simple ML architecture without the need for process-based formulations to effectively estimate spatially continuous SWE at a daily temporal resolution. Beginning with in situ SWE measurements (i.e., SNOTEL), lidar-derived terrain features, and temporal variables, we employ localized feature engineering and optimization via gradient-boosting decision trees to identify regionally unique drivers of snowpack dynamics and use the optimal features to train regionally independent artificial neural networks to estimate regional SWE at a 1 km spatial resolution. The model results yield respectable skill in reconstructed 1 km gridded SWE magnitudes in a hindcast simulation of the 2019 water year that is independent of the training and testing data. Comparing model estimates to over 6200 observations, the model demonstrates a weighted RMSE of 15.4 cm, Kling-Gupta Efficiency metric of 0.86, and a percent bias of 0.71% across 23 snow-influenced regions in the western U.S. The model simulation produces peak SWE estimates within 10 cm for twenty of the twenty-three regions, demonstrating capability in effectively capturing regional snow accumulation processes. The demonstration of low-error ML workflows capable of providing near-real-time, spatially continuous SWE estimates at a high spatial resolution provides proof-of-concept and a foundation to effectively update snow state variables that drive water supply forecasts in snow-dominated regions. • A new Machine Learning-based SWE estimation model for the Western United States. • Utilization of open-source tools and up-to-date in-situ observations to estimate SWE. • Reconstructed SWE magnitudes exhibit respectable skill in a hindcast simulation. • Output may serve as a supplemental state-of-snowpack input to hydrologic models. [ABSTRACT FROM AUTHOR]

Approximately seven years of daily initializations from the convection-allowing National Severe Storms Laboratory Weather Research and Forecasting Model are used as inputs to train random forest (RF) machine learning models to probabilistically predict instances of excessive rainfall. Unlike other hazards, excessive rainfall does not have an accepted definition, so multiple definitions of excessive rainfall and flash flooding—including flash flood reports and 24-h average recurrence intervals (ARIs)—are used to explore RF configuration forecast sensitivities. RF forecasts are analogous to operational Weather Prediction Center (WPC) day-1 Excessive Rainfall Outlooks (EROs) and their resolution, reliability, and skill are strongly influenced by rainfall definitions and how inputs are assembled for training. Models trained with 1-yr ARI exceedances defined by the Stage-IV (ST4) precipitation analysis perform poorly in the northern Great Plains and Southwest United States, in part due to a high bias in the number of training events in these regions. Increasing the ARI threshold to 2 years or removing ST4 data from training, optimizing forecast skill geographically, and spatially averaging meteorological inputs for training generally results in improved CONUS-wide RF forecast skill. Both EROs and RF forecasts have seasonal skill—–poor forecasts in the late fall and winter and skillful forecasts in the summer and early fall. However, the EROs are consistently and significantly better than their RF counterparts, regardless of RF configuration, particularly in the summer months. The results suggest careful consideration should be made when developing ML-based probabilistic precipitation forecasts with convection-allowing model inputs, and further development is necessary to consider these forecast products for operational implementation. Significance Statement: Machine learning (ML) models can deduce statistical relationships between a set of predictors and meteorological events. In this work, ML models are developed to predict excessive rainfall events. Since excessive rainfall is difficult to uniformly define across the United States, multiple ML models are built from a variety of rainfall datasets with predictors gathered from output of a high-resolution numerical weather prediction model and forecasts are made from each model. Forecasts made from these models are highly sensitive to both the definitions of excessive rainfall (e.g., 100 mm of rain in a day may cause flooding in a usually dry area, but not in a wet area) and the predictors used. Forecast skill can increase when excessive rainfall events are rarer and when predictors synthesize the surrounding environment rather than characterize specific geographical points. ML-based models have great potential for excessive rainfall prediction, but careful attention to the configuration of these models is required. [ABSTRACT FROM AUTHOR]