Regional Weather Variable Predictions by Machine Learning with Near-Surface Observational and Atmospheric Numerical Data
Authors:
Yihe Zhang,
Bryce Turney,
Purushottam Sigdel,
Xu Yuan,
Eric Rappin,
Adrian Lago,
Sytske Kimball,
Li Chen,
Paul Darby,
Lu Peng,
Sercan Aygun,
Yazhou Tu,
M. Hassan Najafi,
Nian-Feng Tzeng
Abstract:
Accurate and timely regional weather prediction is vital for sectors dependent on weather-related decisions. Traditional prediction methods, based on atmospheric equations, often struggle with coarse temporal resolutions and inaccuracies. This paper presents a novel machine learning (ML) model, called MiMa (short for Micro-Macro), that integrates both near-surface observational data from Kentucky…
▽ More
Accurate and timely regional weather prediction is vital for sectors dependent on weather-related decisions. Traditional prediction methods, based on atmospheric equations, often struggle with coarse temporal resolutions and inaccuracies. This paper presents a novel machine learning (ML) model, called MiMa (short for Micro-Macro), that integrates both near-surface observational data from Kentucky Mesonet stations (collected every five minutes, known as Micro data) and hourly atmospheric numerical outputs (termed as Macro data) for fine-resolution weather forecasting. The MiMa model employs an encoder-decoder transformer structure, with two encoders for processing multivariate data from both datasets and a decoder for forecasting weather variables over short time horizons. Each instance of the MiMa model, called a modelet, predicts the values of a specific weather parameter at an individual Mesonet station. The approach is extended with Re-MiMa modelets, which are designed to predict weather variables at ungauged locations by training on multivariate data from a few representative stations in a region, tagged with their elevations. Re-MiMa (short for Regional-MiMa) can provide highly accurate predictions across an entire region, even in areas without observational stations. Experimental results show that MiMa significantly outperforms current models, with Re-MiMa offering precise short-term forecasts for ungauged locations, marking a significant advancement in weather forecasting accuracy and applicability.
△ Less
Submitted 10 February, 2025; v1 submitted 11 December, 2024;
originally announced December 2024.
Wind Tunnel Tests of a Pitot-Static Tube Array to Estimate Wind Velocity
Authors:
Matthew Simmons,
Carlos Montalvo,
Sytske Kimball
Abstract:
This paper examines the use of a pitot-static tube array (PSA) to estimate horizontal wind velocity in all directions. This method uses the readings from the PSA as inputs. No flight tests have been done, but this method could potentially be used to estimate wind velocity from small UAVs. To do this, UAV ground speed and attitude must be available as inputs. Wind tunnel data of the PSA was used to…
▽ More
This paper examines the use of a pitot-static tube array (PSA) to estimate horizontal wind velocity in all directions. This method uses the readings from the PSA as inputs. No flight tests have been done, but this method could potentially be used to estimate wind velocity from small UAVs. To do this, UAV ground speed and attitude must be available as inputs. Wind tunnel data of the PSA was used to create an algorithm that estimates wind velocity even when the wind is not parallel with any of the pitot-static tubes. This paper discuses the wind tunnel data, the algorithm, and the application for small UAVs.
△ Less
Submitted 29 January, 2019;
originally announced January 2019.
