Hybrid physics-AI outperforms numerical weather prediction for extreme precipitation nowcasting
Not provided.
Meteorology & Atmospheric Sciences↗
Engineering topics
Singh, Debjani
Publications and source records attributed to Singh, Debjani.
Streamflow in the United States: Characteristics, trends, regime shifts, and extremes
Long-term streamflow observations contain essential information for understanding hydrological changes and managing water resources. A continental-scale dataset or analysis of temporal streamflow change is still lacking across hydrologic gauges in the Conterminous United States (CONUS). Here, we compiled 70 years of streamflow records from 1951 to 2021 at ~ 8000 hydrologic stations across the CONUS and characterized temporal trends, regime shifts, and extreme events using a Bayesian time series analysis algorithm. We found that the occurrences of sudden streamflow changes (e.g., regime shifts and extreme events) have been increasing with time across the CONUS. In addition, we derived 181 streamflow indicators that are valuable for hydrological and biological applications, such as the duration and frequency of high or low streamflow events. The Mississippi River Basin, especially the middle and lower parts, was a hot spot of high-frequency high-flow events. Overall, we anticipate the dataset generated here offers valuable information for understanding and quantifying changes in water resources across the CONUS.
54 ENVIRONMENTAL SCIENCES↗
Hydropower Capacity Factor Trends & Analytics for the United States
This data repository contains all code, input data, and data generated for Turner et al. (2024)—“Hydropower capacity factors trending down in the United States”. File descriptions: – hydro-cf-trends-inputs.zip: Full set of input data used in this study, organized for direct entry into “/data” directory of hydro-cf-trends data processing pipeline. – hydro-cf-trends.zip: Full data processing pipeline, coded using the R {targets} framework. This is a snapshot release (v1.0) of the code repository stored at https://code.ornl.gov/turnersw/hydro-cf-trends/. – hydro-cf-trends-results.zip: Provides all dam level results required to reproduce results and graphics in Turner et al. (2024). Dams are identified by the “complxID” (root of the hydropower plant ID in the Existing Hydropower Assets Database, inherited from HILARRI). Results include: • dam_CF_trends.csv: Table of long-term trends in annualized capacity factors for 610 dams and modeled annualized capacity factors for 362 modeled dams (naturalized and assimilated flows). • dam_annualized_CF_gen.csv: Annualized time series of the following variables for each of 610 hydropower dams with nameplate > 5MW – Reported nameplate capacity (MW) – Implied maximum annual generation (MWh) – Reported net generation (MWh) – Computed annual capacity factor – Modeled annual capacity factor (362 modeled plants only)
13 HYDRO ENERGY↗
The 2024 National Hydropower Map
A visualization of the geospatial distribution and characteristics of operational hydropower plants in the United States in 2024.
13 HYDRO ENERGY↗
Improving Text Classification with Large Language Model-Based Data Augmentation
Large Language Models (LLMs) such as ChatGPT possess advanced capabilities in understanding and generating text. These capabilities enable ChatGPT to create text based on specific instructions, which can serve as augmented data for text classification tasks. Previous studies have approached data augmentation (DA) by either rewriting the existing dataset with ChatGPT or generating entirely new data from scratch. However, it is unclear which method is better without comparing their effectiveness. This study investigates the application of both methods to two datasets: a general-topic dataset (Reuters news data) and a domain-specific dataset (Mitigation dataset). Our findings indicate that: 1. ChatGPT generated new data consistently enhanced model’s classification results for both datasets. 2. Generating new data generally outperforms rewriting existing data, though crafting the prompts carefully is crucial to extract the most valuable information from ChatGPT, particularly for domain-specific data. 3. The augmentation data size affects the effectiveness of DA; however, we observed a plateau after incorporating 10 samples. 4. Combining the rewritten sample with new generated sample can potentially further improve the model’s performance.
97 MATHEMATICS AND COMPUTING↗
Hydropower capacity factors trending down in the United States
Abstract The United States hydropower fleet has faced increasing environmental and regulatory pressures over the last half century, potentially constraining total generation. Here we show that annual capacity factor has declined at four fifths of United States hydropower plants since 1980, with two thirds of decreasing trends significant at p < 0.05. Results are based on an analysis of annual energy generation totals and nameplate capacities for 610 plants (>5 megawatt), representing 87% of total conventional hydropower capacity in the United States. On aggregate, changes in capacity factor imply a fleetwide, cumulative generation decrease of 23% since 1980 before factoring in capacity upgrades—akin to retiring a Hoover Dam once every two to three years. Changes in water availability explain energy decline in only 21% of plants, highlighting the importance of non-climatic drivers of generation, including deterioration of plant equipment as well as changes to dam operations in support of nonpower objectives.
13 HYDRO ENERGY↗