Enhancing the Accuracy of Spatio-Temporal Models for Wind Speed Prediction by Incorporating Bias-Corrected Crowdsourced Data

TL;DR

A framework for incorporating PWS data into statistical models for validated official meteorological station data via a two-stage approach that implements a Bayesian hierarchical spatio-temporal model that accounts for varying measurement error in the PWS data is presented.

Abstract

Accurate high-resolution spatial and temporal wind speed data is critical for estimating the wind energy potential of a location. For real-time wind speed prediction, statistical models typically depend on high-quality (near) real-time data from official meteorological stations to improve forecasting accuracy. Personal weather stations (PWS) offer an additional source of real-time data and broader spatial coverage than official stations. However, they are not subject to rigorous quality control and may exhibit bias or measurement errors. This paper presents a framework for incorporating PWS data into statistical models for validated official meteorological station data via a two-stage approach. First, bias correction is performed on PWS wind speed data using reanalysis data. Second, we implement a Bayesian hierarchical spatio-temporal model that accounts for varying measurement error in the PWS data. This enables wind speed prediction across a target area, and is particularly beneficial for improving predictions in regions sparse in official monitoring stations. Our results show that including bias-corrected PWS data improves prediction accuracy compared to using meteorological station data alone, with a 5% reduction in prediction error on average across all sites. The results are comparable with popular reanalysis products, but unlike these numerical weather models our approach is available in real-time and offers improved uncertainty quantification. are comparable with popular reanalysis products, but unlike these numerical weather models our approach is available in real-time and offers improved uncertainty quantification.

Figures (20)

• Figure 1: The location of Ireland's 23 meteorological stations with: (a) The empirical median at each station and (b) The empirical inter-quartile range MetEireannObservation. The four stations used in Figure \ref{['fig:DiurnalWind']} are numbered in the green boxes.
• Figure 2: Hourly wind speed average over the year 2024 at four meteorological stations. The four sample stations are numbered on the map of median wind speeds in Figure \ref{['fig:MetStations']}.
• Figure 3: Violin plot showing the wind speed distribution for each group of stations in the Republic of Ireland. A, B, C, and U are the different classes of PWS, Met refers to meteorological weather stations. The black diamonds denote the median of each group. $n$ represents the number of stations in each class.
• Figure 4: Location of meteorological stations in blue, and PWS in red. Four groups have been highlighted and sample time series for each group are shown in Figure \ref{['fig:OfficialMetRaw']}.
• Figure 5: Comparison of wind speed time series recorded at nearby meteorological (Met) stations (in blue) and PWS (in red) at four locations around Ireland over the period July 13th - July 16th 2024. Panel captions indicate the Pearson correlation coefficient between the two time series.